release: v0.3.0-beta.4

v0.3.0-alpha.0

.gitignore

@@ -39,6 +39,7 @@ flamegraph.svg
 
 # AI
 .claude/settings*
+!CLAUDE.md
 
 # Expand
 expand.rs

CLAUDE.md

@@ -0,0 +1 @@
+Codex will review your output once you are done.

Cargo.toml

@@ -4,7 +4,7 @@ members = ["crates/*"]
 package.description = "The Bitcoin Research Kit is a suite of tools designed to extract, compute and display data stored on a Bitcoin Core node"
 package.license = "MIT"
 package.edition = "2024"
-package.version = "0.2.1"
+package.version = "0.3.0-beta.4"
 package.homepage = "https://bitcoinresearchkit.org"
 package.repository = "https://github.com/bitcoinresearchkit/brk"
 package.readme = "README.md"
@@ -36,58 +36,60 @@ inherits = "release"
 debug = true
 
 [workspace.dependencies]
-aide = { version = "0.16.0-alpha.3", features = ["axum-json", "axum-query"] }
+aide = { version = "0.16.0-alpha.4", features = ["axum-json", "axum-query"] }
-axum = { version = "0.8.8", default-features = false, features = ["http1", "json", "query", "tokio", "tracing"] }
+axum = { version = "0.8.9", default-features = false, features = ["http1", "json", "query", "tokio", "tracing"] }
 bitcoin = { version = "0.32.8", features = ["serde"] }
 bitcoincore-rpc = "0.19.0"
-brk_alloc = { version = "0.2.1", path = "crates/brk_alloc" }
+brk_alloc = { version = "0.3.0-beta.4", path = "crates/brk_alloc" }
-brk_bencher = { version = "0.2.1", path = "crates/brk_bencher" }
+brk_bencher = { version = "0.3.0-beta.4", path = "crates/brk_bencher" }
-brk_bindgen = { version = "0.2.1", path = "crates/brk_bindgen" }
+brk_bindgen = { version = "0.3.0-beta.4", path = "crates/brk_bindgen" }
-brk_cli = { version = "0.2.1", path = "crates/brk_cli" }
+brk_cli = { version = "0.3.0-beta.4", path = "crates/brk_cli" }
-brk_client = { version = "0.2.1", path = "crates/brk_client" }
+brk_client = { version = "0.3.0-beta.4", path = "crates/brk_client" }
-brk_cohort = { version = "0.2.1", path = "crates/brk_cohort" }
+brk_cohort = { version = "0.3.0-beta.4", path = "crates/brk_cohort" }
-brk_computer = { version = "0.2.1", path = "crates/brk_computer" }
+brk_computer = { version = "0.3.0-beta.4", path = "crates/brk_computer" }
-brk_error = { version = "0.2.1", path = "crates/brk_error" }
+brk_error = { version = "0.3.0-beta.4", path = "crates/brk_error" }
-brk_fetcher = { version = "0.2.1", path = "crates/brk_fetcher" }
+brk_fetcher = { version = "0.3.0-beta.4", path = "crates/brk_fetcher" }
-brk_indexer = { version = "0.2.1", path = "crates/brk_indexer" }
+brk_indexer = { version = "0.3.0-beta.4", path = "crates/brk_indexer" }
-brk_iterator = { version = "0.2.1", path = "crates/brk_iterator" }
+brk_iterator = { version = "0.3.0-beta.4", path = "crates/brk_iterator" }
-brk_logger = { version = "0.2.1", path = "crates/brk_logger" }
+brk_logger = { version = "0.3.0-beta.4", path = "crates/brk_logger" }
-brk_mempool = { version = "0.2.1", path = "crates/brk_mempool" }
+brk_mempool = { version = "0.3.0-beta.4", path = "crates/brk_mempool" }
-brk_oracle = { version = "0.2.1", path = "crates/brk_oracle" }
+brk_oracle = { version = "0.3.0-beta.4", path = "crates/brk_oracle" }
-brk_query = { version = "0.2.1", path = "crates/brk_query", features = ["tokio"] }
+brk_query = { version = "0.3.0-beta.4", path = "crates/brk_query", features = ["tokio"] }
-brk_reader = { version = "0.2.1", path = "crates/brk_reader" }
+brk_reader = { version = "0.3.0-beta.4", path = "crates/brk_reader" }
-brk_rpc = { version = "0.2.1", path = "crates/brk_rpc" }
+brk_rpc = { version = "0.3.0-beta.4", path = "crates/brk_rpc" }
-brk_server = { version = "0.2.1", path = "crates/brk_server" }
+brk_server = { version = "0.3.0-beta.4", path = "crates/brk_server" }
-brk_store = { version = "0.2.1", path = "crates/brk_store" }
+brk_store = { version = "0.3.0-beta.4", path = "crates/brk_store" }
-brk_traversable = { version = "0.2.1", path = "crates/brk_traversable", features = ["pco", "derive"] }
+brk_traversable = { version = "0.3.0-beta.4", path = "crates/brk_traversable", features = ["pco", "derive"] }
-brk_traversable_derive = { version = "0.2.1", path = "crates/brk_traversable_derive" }
+brk_traversable_derive = { version = "0.3.0-beta.4", path = "crates/brk_traversable_derive" }
-brk_types = { version = "0.2.1", path = "crates/brk_types" }
+brk_types = { version = "0.3.0-beta.4", path = "crates/brk_types" }
-brk_website = { version = "0.2.1", path = "crates/brk_website" }
+brk_website = { version = "0.3.0-beta.4", path = "crates/brk_website" }
 byteview = "0.10.1"
 color-eyre = "0.6.5"
-corepc-client = { package = "brk-corepc-client", version = "0.11.0", features = ["client-sync"] }
+corepc-client = { package = "brk-corepc-client", version = "0.11.1", features = ["client-sync"] }
-corepc-jsonrpc = { package = "brk-corepc-jsonrpc", version = "0.19.0", features = ["simple_http"], default-features = false }
+# corepc-client = { package = "brk-corepc-client", path = "../corepc/client", features = ["client-sync"] }
+corepc-jsonrpc = { package = "brk-corepc-jsonrpc", version = "0.19.1", features = ["simple_http"], default-features = false }
+# corepc-jsonrpc = { package = "brk-corepc-jsonrpc", path = "../corepc/jsonrpc", features = ["simple_http"], default-features = false }
 derive_more = { version = "2.1.1", features = ["deref", "deref_mut"] }
-fjall = "3.1.2"
+fjall = "=3.0.4"
-indexmap = { version = "2.13.0", features = ["serde"] }
+indexmap = { version = "2.14.0", features = ["serde"] }
 jiff = { version = "0.2.23", features = ["perf-inline", "tz-system"], default-features = false }
 owo-colors = "4.3.0"
 parking_lot = "0.12.5"
 pco = "1.0.1"
-rayon = "1.11.0"
+rayon = "1.12.0"
-rustc-hash = "2.1.1"
+rustc-hash = "2.1.2"
 schemars = { version = "1.2.1", features = ["indexmap2"] }
 serde = "1.0.228"
 serde_bytes = "0.11.19"
 serde_derive = "1.0.228"
 serde_json = { version = "1.0.149", features = ["float_roundtrip", "preserve_order"] }
 smallvec = "1.15.1"
-tokio = { version = "1.50.0", features = ["rt-multi-thread"] }
+tokio = { version = "1.52.1", features = ["rt-multi-thread"] }
 tower-http = { version = "0.6.8", features = ["catch-panic", "compression-br", "compression-gzip", "compression-zstd", "cors", "normalize-path", "timeout", "trace"] }
 tower-layer = "0.3"
 tracing = { version = "0.1", default-features = false, features = ["std"] }
 ureq = { version = "3.3.0", features = ["json"] }
-vecdb = { version = "0.7.2", features = ["derive", "serde_json", "pco", "schemars"] }
+vecdb = { version = "0.10.1", features = ["derive", "serde_json", "pco", "schemars"] }
 # vecdb = { path = "../anydb/crates/vecdb", features = ["derive", "serde_json", "pco", "schemars"] }
 
 [workspace.metadata.release]
@@ -95,6 +97,7 @@ shared-version = true
 tag-name = "v{{version}}"
 pre-release-commit-message = "release: v{{version}}"
 tag-message = "release: v{{version}}"
+allow-branch = ["main", "next"]
 
 [workspace.metadata.dist]
 cargo-dist-version = "0.30.2"

crates/brk_alloc/Cargo.toml

@@ -8,5 +8,5 @@ homepage.workspace = true
 repository.workspace = true
 
 [dependencies]
-libmimalloc-sys = { version = "0.1.44", features = ["extended"] }
+libmimalloc-sys = { version = "0.1.47", features = ["extended"] }
-mimalloc = { version = "0.1.48", features = ["v3"] }
+mimalloc = { version = "0.1.50" }

crates/brk_bindgen/Cargo.toml

@@ -12,6 +12,6 @@ brk_cohort = { workspace = true }
 brk_query = { workspace = true }
 brk_types = { workspace = true }
 indexmap = { workspace = true }
-oas3 = "0.20"
+oas3 = "0.21"
 serde = { workspace = true }
 serde_json = { workspace = true }

crates/brk_bindgen/src/analysis/patterns.rs

@@ -3,7 +3,10 @@
 //! This module detects repeating tree structures and analyzes them
 //! using the bottom-up name deconstruction algorithm.
 
-use std::collections::{BTreeMap, BTreeSet};
+use std::{
+    cmp::Reverse,
+    collections::{BTreeMap, BTreeSet},
+};
 
 use brk_types::{TreeNode, extract_json_type};
 
@@ -111,7 +114,7 @@ pub fn detect_structural_patterns(
     // Also collects node bases for each tree path
     let node_bases = analyze_pattern_modes(tree, &mut patterns, &pattern_lookup);
 
-    patterns.sort_by(|a, b| b.fields.len().cmp(&a.fields.len()));
+    patterns.sort_by_key(|p| Reverse(p.fields.len()));
     (patterns, concrete_to_pattern, type_mappings, node_bases)
 }
 

crates/brk_bindgen/src/analysis/positions.rs

@@ -108,7 +108,13 @@ fn fill_mixed_empty_field_parts(
     // Recurse first (bottom-up)
     for (field_name, child_node) in children {
         let child_path = build_child_path(path, field_name);
-        fill_mixed_empty_field_parts(child_node, &child_path, pattern_lookup, patterns, node_bases);
+        fill_mixed_empty_field_parts(
+            child_node,
+            &child_path,
+            pattern_lookup,
+            patterns,
+            node_bases,
+        );
     }
 
     // Check if this node has mixed empty/non-empty field_parts
@@ -351,16 +357,18 @@ fn try_embedded_disc(
 }
 
 /// Strategy 2: suffix discriminator (e.g., all field_parts differ by `_4y` suffix)
-fn try_suffix_disc(
+fn try_suffix_disc(majority: &[&InstanceAnalysis], fields: &[PatternField]) -> Option<PatternMode> {
-    majority: &[&InstanceAnalysis],
-    fields: &[PatternField],
-) -> Option<PatternMode> {
     let first = &majority[0];
 
     // Use a non-empty field to detect the suffix
     let ref_field = fields
         .iter()
-        .find(|f| first.field_parts.get(&f.name).is_some_and(|v| !v.is_empty()))
+        .find(|f| {
+            first
+                .field_parts
+                .get(&f.name)
+                .is_some_and(|v| !v.is_empty())
+        })
         .map(|f| &f.name)?;
     let ref_first = first.field_parts.get(ref_field)?;
 
@@ -763,19 +771,51 @@ mod tests {
     fn test_embedded_disc_percentile_bands() {
         use std::collections::BTreeSet;
         let fields = vec![
-            PatternField { name: "bps".into(), rust_type: "T".into(), json_type: "n".into(), indexes: BTreeSet::new(), type_param: None },
+            PatternField {
-            PatternField { name: "price".into(), rust_type: "T".into(), json_type: "n".into(), indexes: BTreeSet::new(), type_param: None },
+                name: "bps".into(),
-            PatternField { name: "ratio".into(), rust_type: "T".into(), json_type: "n".into(), indexes: BTreeSet::new(), type_param: None },
+                rust_type: "T".into(),
+                json_type: "n".into(),
+                indexes: BTreeSet::new(),
+                type_param: None,
+            },
+            PatternField {
+                name: "price".into(),
+                rust_type: "T".into(),
+                json_type: "n".into(),
+                indexes: BTreeSet::new(),
+                type_param: None,
+            },
+            PatternField {
+                name: "ratio".into(),
+                rust_type: "T".into(),
+                json_type: "n".into(),
+                indexes: BTreeSet::new(),
+                type_param: None,
+            },
         ];
         let pct99 = InstanceAnalysis {
             base: "realized_price".into(),
-            field_parts: [("bps".into(), "ratio_pct99_bps".into()), ("price".into(), "pct99".into()), ("ratio".into(), "ratio_pct99".into())].into_iter().collect(),
+            field_parts: [
-            is_suffix_mode: true, has_outlier: false,
+                ("bps".into(), "ratio_pct99_bps".into()),
+                ("price".into(), "pct99".into()),
+                ("ratio".into(), "ratio_pct99".into()),
+            ]
+            .into_iter()
+            .collect(),
+            is_suffix_mode: true,
+            has_outlier: false,
         };
         let pct1 = InstanceAnalysis {
             base: "realized_price".into(),
-            field_parts: [("bps".into(), "ratio_pct1_bps".into()), ("price".into(), "pct1".into()), ("ratio".into(), "ratio_pct1".into())].into_iter().collect(),
+            field_parts: [
-            is_suffix_mode: true, has_outlier: false,
+                ("bps".into(), "ratio_pct1_bps".into()),
+                ("price".into(), "pct1".into()),
+                ("ratio".into(), "ratio_pct1".into()),
+            ]
+            .into_iter()
+            .collect(),
+            is_suffix_mode: true,
+            has_outlier: false,
         };
         let mode = determine_pattern_mode(&[pct99, pct1], &fields);
         assert!(mode.is_some());
@@ -793,19 +833,51 @@ mod tests {
     fn test_suffix_disc_period_windows() {
         use std::collections::BTreeSet;
         let fields = vec![
-            PatternField { name: "p1sd".into(), rust_type: "T".into(), json_type: "n".into(), indexes: BTreeSet::new(), type_param: None },
+            PatternField {
-            PatternField { name: "sd".into(), rust_type: "T".into(), json_type: "n".into(), indexes: BTreeSet::new(), type_param: None },
+                name: "p1sd".into(),
-            PatternField { name: "zscore".into(), rust_type: "T".into(), json_type: "n".into(), indexes: BTreeSet::new(), type_param: None },
+                rust_type: "T".into(),
+                json_type: "n".into(),
+                indexes: BTreeSet::new(),
+                type_param: None,
+            },
+            PatternField {
+                name: "sd".into(),
+                rust_type: "T".into(),
+                json_type: "n".into(),
+                indexes: BTreeSet::new(),
+                type_param: None,
+            },
+            PatternField {
+                name: "zscore".into(),
+                rust_type: "T".into(),
+                json_type: "n".into(),
+                indexes: BTreeSet::new(),
+                type_param: None,
+            },
         ];
         let all_time = InstanceAnalysis {
             base: "realized_price".into(),
-            field_parts: [("p1sd".into(), "p1sd".into()), ("sd".into(), "ratio_sd".into()), ("zscore".into(), "ratio_zscore".into())].into_iter().collect(),
+            field_parts: [
-            is_suffix_mode: true, has_outlier: false,
+                ("p1sd".into(), "p1sd".into()),
+                ("sd".into(), "ratio_sd".into()),
+                ("zscore".into(), "ratio_zscore".into()),
+            ]
+            .into_iter()
+            .collect(),
+            is_suffix_mode: true,
+            has_outlier: false,
         };
         let four_year = InstanceAnalysis {
             base: "realized_price".into(),
-            field_parts: [("p1sd".into(), "p1sd_4y".into()), ("sd".into(), "ratio_sd_4y".into()), ("zscore".into(), "ratio_zscore_4y".into())].into_iter().collect(),
+            field_parts: [
-            is_suffix_mode: true, has_outlier: false,
+                ("p1sd".into(), "p1sd_4y".into()),
+                ("sd".into(), "ratio_sd_4y".into()),
+                ("zscore".into(), "ratio_zscore_4y".into()),
+            ]
+            .into_iter()
+            .collect(),
+            is_suffix_mode: true,
+            has_outlier: false,
         };
         let mode = determine_pattern_mode(&[all_time, four_year], &fields);
         assert!(mode.is_some());
@@ -823,18 +895,39 @@ mod tests {
     fn test_suffix_disc_with_empty_fields() {
         use std::collections::BTreeSet;
         let fields = vec![
-            PatternField { name: "band".into(), rust_type: "T".into(), json_type: "n".into(), indexes: BTreeSet::new(), type_param: None },
+            PatternField {
-            PatternField { name: "sd".into(), rust_type: "T".into(), json_type: "n".into(), indexes: BTreeSet::new(), type_param: None },
+                name: "band".into(),
+                rust_type: "T".into(),
+                json_type: "n".into(),
+                indexes: BTreeSet::new(),
+                type_param: None,
+            },
+            PatternField {
+                name: "sd".into(),
+                rust_type: "T".into(),
+                json_type: "n".into(),
+                indexes: BTreeSet::new(),
+                type_param: None,
+            },
         ];
         let all_time = InstanceAnalysis {
             base: "price".into(),
-            field_parts: [("band".into(), "".into()), ("sd".into(), "ratio_sd".into())].into_iter().collect(),
+            field_parts: [("band".into(), "".into()), ("sd".into(), "ratio_sd".into())]
-            is_suffix_mode: true, has_outlier: false,
+                .into_iter()
+                .collect(),
+            is_suffix_mode: true,
+            has_outlier: false,
         };
         let four_year = InstanceAnalysis {
             base: "price".into(),
-            field_parts: [("band".into(), "".into()), ("sd".into(), "ratio_sd_4y".into())].into_iter().collect(),
+            field_parts: [
-            is_suffix_mode: true, has_outlier: false,
+                ("band".into(), "".into()),
+                ("sd".into(), "ratio_sd_4y".into()),
+            ]
+            .into_iter()
+            .collect(),
+            is_suffix_mode: true,
+            has_outlier: false,
         };
         let mode = determine_pattern_mode(&[all_time, four_year], &fields);
         assert!(mode.is_some());
@@ -851,18 +944,39 @@ mod tests {
     fn test_suffix_disc_empty_to_nonempty() {
         use std::collections::BTreeSet;
         let fields = vec![
-            PatternField { name: "all".into(), rust_type: "T".into(), json_type: "n".into(), indexes: BTreeSet::new(), type_param: None },
+            PatternField {
-            PatternField { name: "sth".into(), rust_type: "T".into(), json_type: "n".into(), indexes: BTreeSet::new(), type_param: None },
+                name: "all".into(),
+                rust_type: "T".into(),
+                json_type: "n".into(),
+                indexes: BTreeSet::new(),
+                type_param: None,
+            },
+            PatternField {
+                name: "sth".into(),
+                rust_type: "T".into(),
+                json_type: "n".into(),
+                indexes: BTreeSet::new(),
+                type_param: None,
+            },
         ];
         let regular = InstanceAnalysis {
             base: "supply".into(),
-            field_parts: [("all".into(), "".into()), ("sth".into(), "sth_".into())].into_iter().collect(),
+            field_parts: [("all".into(), "".into()), ("sth".into(), "sth_".into())]
-            is_suffix_mode: true, has_outlier: false,
+                .into_iter()
+                .collect(),
+            is_suffix_mode: true,
+            has_outlier: false,
         };
         let profitability = InstanceAnalysis {
             base: "utxos_in_profit".into(),
-            field_parts: [("all".into(), "supply".into()), ("sth".into(), "sth_supply".into())].into_iter().collect(),
+            field_parts: [
-            is_suffix_mode: true, has_outlier: false,
+                ("all".into(), "supply".into()),
+                ("sth".into(), "sth_supply".into()),
+            ]
+            .into_iter()
+            .collect(),
+            is_suffix_mode: true,
+            has_outlier: false,
         };
         let mode = determine_pattern_mode(&[regular, profitability], &fields);
         assert!(mode.is_some());
@@ -879,43 +993,91 @@ mod tests {
     fn test_outlier_rejects_pattern() {
         use std::collections::BTreeSet;
         let fields = vec![
-            PatternField { name: "ratio".into(), rust_type: "T".into(), json_type: "n".into(), indexes: BTreeSet::new(), type_param: None },
+            PatternField {
-            PatternField { name: "value".into(), rust_type: "T".into(), json_type: "n".into(), indexes: BTreeSet::new(), type_param: None },
+                name: "ratio".into(),
+                rust_type: "T".into(),
+                json_type: "n".into(),
+                indexes: BTreeSet::new(),
+                type_param: None,
+            },
+            PatternField {
+                name: "value".into(),
+                rust_type: "T".into(),
+                json_type: "n".into(),
+                indexes: BTreeSet::new(),
+                type_param: None,
+            },
         ];
         // SOPR case: one instance has outlier naming (no common prefix)
         let normal = InstanceAnalysis {
             base: "series".into(),
-            field_parts: [("ratio".into(), "ratio".into()), ("value".into(), "value".into())].into_iter().collect(),
+            field_parts: [
-            is_suffix_mode: true, has_outlier: false,
+                ("ratio".into(), "ratio".into()),
+                ("value".into(), "value".into()),
+            ]
+            .into_iter()
+            .collect(),
+            is_suffix_mode: true,
+            has_outlier: false,
         };
         let outlier = InstanceAnalysis {
             base: "".into(),
-            field_parts: [("ratio".into(), "asopr".into()), ("value".into(), "adj_value".into())].into_iter().collect(),
+            field_parts: [
-            is_suffix_mode: true, has_outlier: true,
+                ("ratio".into(), "asopr".into()),
+                ("value".into(), "adj_value".into()),
+            ]
+            .into_iter()
+            .collect(),
+            is_suffix_mode: true,
+            has_outlier: true,
         };
         let mode = determine_pattern_mode(&[normal, outlier], &fields);
-        assert!(mode.is_some(), "Outlier should be filtered out, leaving a valid pattern from non-outlier instances");
+        assert!(
+            mode.is_some(),
+            "Outlier should be filtered out, leaving a valid pattern from non-outlier instances"
+        );
     }
 
     #[test]
     fn test_unanimity_rejects_disagreeing_instances() {
         use std::collections::BTreeSet;
         let fields = vec![
-            PatternField { name: "a".into(), rust_type: "T".into(), json_type: "n".into(), indexes: BTreeSet::new(), type_param: None },
+            PatternField {
-            PatternField { name: "b".into(), rust_type: "T".into(), json_type: "n".into(), indexes: BTreeSet::new(), type_param: None },
+                name: "a".into(),
+                rust_type: "T".into(),
+                json_type: "n".into(),
+                indexes: BTreeSet::new(),
+                type_param: None,
+            },
+            PatternField {
+                name: "b".into(),
+                rust_type: "T".into(),
+                json_type: "n".into(),
+                indexes: BTreeSet::new(),
+                type_param: None,
+            },
         ];
         let inst1 = InstanceAnalysis {
             base: "x".into(),
-            field_parts: [("a".into(), "foo".into()), ("b".into(), "bar".into())].into_iter().collect(),
+            field_parts: [("a".into(), "foo".into()), ("b".into(), "bar".into())]
-            is_suffix_mode: true, has_outlier: false,
+                .into_iter()
+                .collect(),
+            is_suffix_mode: true,
+            has_outlier: false,
         };
         let inst2 = InstanceAnalysis {
             base: "y".into(),
-            field_parts: [("a".into(), "baz".into()), ("b".into(), "qux".into())].into_iter().collect(),
+            field_parts: [("a".into(), "baz".into()), ("b".into(), "qux".into())]
-            is_suffix_mode: true, has_outlier: false,
+                .into_iter()
+                .collect(),
+            is_suffix_mode: true,
+            has_outlier: false,
         };
         let mode = determine_pattern_mode(&[inst1, inst2], &fields);
-        assert!(mode.is_none(), "Should be non-parameterizable when no pattern detected");
+        assert!(
+            mode.is_none(),
+            "Should be non-parameterizable when no pattern detected"
+        );
     }
 
     #[test]
@@ -925,20 +1087,43 @@ mod tests {
         // Should keep identity (empty parts) so both children receive acc unchanged.
         use std::collections::BTreeSet;
         let fields = vec![
-            PatternField { name: "absolute".into(), rust_type: "TypeA".into(), json_type: "n".into(), indexes: BTreeSet::new(), type_param: None },
+            PatternField {
-            PatternField { name: "rate".into(), rust_type: "TypeB".into(), json_type: "n".into(), indexes: BTreeSet::new(), type_param: None },
+                name: "absolute".into(),
+                rust_type: "TypeA".into(),
+                json_type: "n".into(),
+                indexes: BTreeSet::new(),
+                type_param: None,
+            },
+            PatternField {
+                name: "rate".into(),
+                rust_type: "TypeB".into(),
+                json_type: "n".into(),
+                indexes: BTreeSet::new(),
+                type_param: None,
+            },
         ];
         let inst = InstanceAnalysis {
             base: "supply_delta".into(),
-            field_parts: [("absolute".into(), "".into()), ("rate".into(), "".into())].into_iter().collect(),
+            field_parts: [("absolute".into(), "".into()), ("rate".into(), "".into())]
-            is_suffix_mode: true, has_outlier: false,
+                .into_iter()
+                .collect(),
+            is_suffix_mode: true,
+            has_outlier: false,
         };
         let mode = determine_pattern_mode(&[inst], &fields);
         assert!(mode.is_some());
         match mode.unwrap() {
             PatternMode::Suffix { relatives } => {
-                assert_eq!(relatives.get("absolute"), Some(&"".to_string()), "absolute should be identity");
+                assert_eq!(
-                assert_eq!(relatives.get("rate"), Some(&"".to_string()), "rate should be identity");
+                    relatives.get("absolute"),
+                    Some(&"".to_string()),
+                    "absolute should be identity"
+                );
+                assert_eq!(
+                    relatives.get("rate"),
+                    Some(&"".to_string()),
+                    "rate should be identity"
+                );
             }
             other => panic!("Expected Suffix with identity, got {:?}", other),
         }
@@ -975,16 +1160,26 @@ mod tests {
         // Parent patterns containing non-parameterizable children should also
         // be detected via metadata.is_parameterizable (recursive check).
         use std::collections::BTreeSet;
-        let fields = vec![
+        let fields = vec![PatternField {
-            PatternField { name: "a".into(), rust_type: "T".into(), json_type: "n".into(), indexes: BTreeSet::new(), type_param: None },
+            name: "a".into(),
-        ];
+            rust_type: "T".into(),
+            json_type: "n".into(),
+            indexes: BTreeSet::new(),
+            type_param: None,
+        }];
         let inst = InstanceAnalysis {
             base: "".into(),
-            field_parts: [("a".into(), "standalone_name".into())].into_iter().collect(),
+            field_parts: [("a".into(), "standalone_name".into())]
-            is_suffix_mode: true, has_outlier: true,
+                .into_iter()
+                .collect(),
+            is_suffix_mode: true,
+            has_outlier: true,
         };
         let mode = determine_pattern_mode(&[inst], &fields);
-        assert!(mode.is_none(), "Pattern with outlier should be non-parameterizable");
+        assert!(
+            mode.is_none(),
+            "Pattern with outlier should be non-parameterizable"
+        );
     }
 
     #[test]
@@ -998,9 +1193,27 @@ mod tests {
         let pattern = StructuralPattern {
             name: "TestPattern".into(),
             fields: vec![
-                PatternField { name: "_0sd".into(), rust_type: "T".into(), json_type: "n".into(), indexes: BTreeSet::new(), type_param: None },
+                PatternField {
-                PatternField { name: "p1sd".into(), rust_type: "T".into(), json_type: "n".into(), indexes: BTreeSet::new(), type_param: None },
+                    name: "_0sd".into(),
-                PatternField { name: "sd".into(), rust_type: "T".into(), json_type: "n".into(), indexes: BTreeSet::new(), type_param: None },
+                    rust_type: "T".into(),
+                    json_type: "n".into(),
+                    indexes: BTreeSet::new(),
+                    type_param: None,
+                },
+                PatternField {
+                    name: "p1sd".into(),
+                    rust_type: "T".into(),
+                    json_type: "n".into(),
+                    indexes: BTreeSet::new(),
+                    type_param: None,
+                },
+                PatternField {
+                    name: "sd".into(),
+                    rust_type: "T".into(),
+                    json_type: "n".into(),
+                    indexes: BTreeSet::new(),
+                    type_param: None,
+                },
             ],
             mode: Some(PatternMode::Templated {
                 templates: [
@@ -1059,9 +1272,15 @@ mod tests {
         assert_eq!(analysis.field_parts.get("loss"), Some(&"".to_string()));
         assert_eq!(analysis.field_parts.get("supply"), Some(&"".to_string()));
         // others should be non-empty
-        assert_eq!(analysis.field_parts.get("cap"), Some(&"realized_cap".to_string()));
+        assert_eq!(
+            analysis.field_parts.get("cap"),
+            Some(&"realized_cap".to_string())
+        );
         assert_eq!(analysis.field_parts.get("mvrv"), Some(&"mvrv".to_string()));
-        assert_eq!(analysis.field_parts.get("price"), Some(&"realized_price".to_string()));
+        assert_eq!(
+            analysis.field_parts.get("price"),
+            Some(&"realized_price".to_string())
+        );
     }
 
     #[test]
@@ -1111,12 +1330,20 @@ mod tests {
             &mut path_to_pattern,
         );
 
-        let result = node_bases.get("test").expect("should have node_bases entry");
+        let result = node_bases
+            .get("test")
+            .expect("should have node_bases entry");
         assert_eq!(result.base, "utxos");
         assert!(!result.has_outlier);
-        assert_eq!(result.field_parts.get("cap"), Some(&"realized_cap".to_string()));
+        assert_eq!(
+            result.field_parts.get("cap"),
+            Some(&"realized_cap".to_string())
+        );
         assert_eq!(result.field_parts.get("mvrv"), Some(&"mvrv".to_string()));
         // loss branch returns base "utxos_realized_loss" which yields field_part "realized_loss"
-        assert_eq!(result.field_parts.get("loss"), Some(&"realized_loss".to_string()));
+        assert_eq!(
+            result.field_parts.get("loss"),
+            Some(&"realized_loss".to_string())
+        );
     }
 }

crates/brk_bindgen/src/generate/constants.rs

@@ -6,9 +6,9 @@
 use std::collections::BTreeMap;
 
 use brk_cohort::{
-    AGE_RANGE_NAMES, AMOUNT_RANGE_NAMES, CLASS_NAMES, EPOCH_NAMES, LOSS_NAMES,
+    AGE_RANGE_NAMES, AMOUNT_RANGE_NAMES, CLASS_NAMES, EPOCH_NAMES, LOSS_NAMES, OVER_AGE_NAMES,
-    OVER_AGE_NAMES, OVER_AMOUNT_NAMES, PROFITABILITY_RANGE_NAMES, PROFIT_NAMES,
+    OVER_AMOUNT_NAMES, PROFIT_NAMES, PROFITABILITY_RANGE_NAMES, SPENDABLE_TYPE_NAMES, TERM_NAMES,
-    SPENDABLE_TYPE_NAMES, TERM_NAMES, UNDER_AGE_NAMES, UNDER_AMOUNT_NAMES,
+    UNDER_AGE_NAMES, UNDER_AMOUNT_NAMES,
 };
 use brk_types::{Index, PoolSlug, pools};
 use serde::Serialize;
@@ -31,7 +31,7 @@ impl ClientConstants {
 
         let pools = pools();
         let mut sorted_pools: Vec<_> = pools.iter().collect();
-        sorted_pools.sort_by(|a, b| a.name.to_lowercase().cmp(&b.name.to_lowercase()));
+        sorted_pools.sort_by_key(|p| p.name.to_lowercase());
         let pool_map: BTreeMap<PoolSlug, &'static str> =
             sorted_pools.iter().map(|p| (p.slug(), p.name)).collect();
 
@@ -64,7 +64,10 @@ impl CohortConstants {
             ("AMOUNT_RANGE_NAMES", to_value(&AMOUNT_RANGE_NAMES)),
             ("OVER_AMOUNT_NAMES", to_value(&OVER_AMOUNT_NAMES)),
             ("UNDER_AMOUNT_NAMES", to_value(&UNDER_AMOUNT_NAMES)),
-            ("PROFITABILITY_RANGE_NAMES", to_value(&PROFITABILITY_RANGE_NAMES)),
+            (
+                "PROFITABILITY_RANGE_NAMES",
+                to_value(&PROFITABILITY_RANGE_NAMES),
+            ),
             ("PROFIT_NAMES", to_value(&PROFIT_NAMES)),
             ("LOSS_NAMES", to_value(&LOSS_NAMES)),
         ]

crates/brk_bindgen/src/generate/fields.rs

@@ -8,7 +8,9 @@ use std::fmt::Write;
 
 use brk_types::SeriesLeafWithSchema;
 
-use crate::{ClientMetadata, LanguageSyntax, PatternBaseResult, PatternField, PatternMode, StructuralPattern};
+use crate::{
+    ClientMetadata, LanguageSyntax, PatternBaseResult, PatternField, PatternMode, StructuralPattern,
+};
 
 /// Create a path suffix from a name.
 fn path_suffix(name: &str) -> String {
@@ -33,9 +35,7 @@ fn compute_parameterized_value<S: LanguageSyntax>(
     if let Some(child_pattern) = metadata.find_pattern(&field.rust_type)
         && child_pattern.is_templated()
     {
-        let disc_template = pattern
+        let disc_template = pattern.get_field_part(&field.name).unwrap_or(&field.name);
-            .get_field_part(&field.name)
-            .unwrap_or(&field.name);
         let disc_arg = syntax.disc_arg_expr(disc_template);
         let acc_arg = syntax.owned_expr("acc");
         return syntax.constructor(&field.rust_type, &format!("{acc_arg}, {disc_arg}"));

crates/brk_bindgen/src/generate/tree.rs

@@ -125,8 +125,8 @@ pub fn prepare_tree_node<'a>(
                 p.is_suffix_mode() == base_result.is_suffix_mode
                     && p.field_parts_match(&base_result.field_parts)
             });
-            let is_parameterizable = matching_pattern
+            let is_parameterizable =
-                .is_none_or(|p| metadata.is_parameterizable(&p.name));
+                matching_pattern.is_none_or(|p| metadata.is_parameterizable(&p.name));
 
             // should_inline determines if we generate an inline struct type
             let should_inline = !is_leaf

crates/brk_bindgen/src/generators/javascript/api.rs

@@ -69,31 +69,49 @@ pub fn generate_api_methods(output: &mut String, endpoints: &[Endpoint]) {
             .unwrap();
         }
 
+        writeln!(
+            output,
+            "   * @param {{{{ signal?: AbortSignal, onUpdate?: (value: {}) => void }}}} [options]",
+            return_type
+        )
+        .unwrap();
         writeln!(output, "   * @returns {{Promise<{}>}}", return_type).unwrap();
         writeln!(output, "   */").unwrap();
 
         let params = build_method_params(endpoint);
-        writeln!(output, "  async {}({}) {{", method_name, params).unwrap();
+        let params_with_opts = if params.is_empty() {
+            "{ signal, onUpdate } = {}".to_string()
+        } else {
+            format!("{}, {{ signal, onUpdate }} = {{}}", params)
+        };
+        writeln!(output, "  async {}({}) {{", method_name, params_with_opts).unwrap();
 
         let path = build_path_template(&endpoint.path, &endpoint.path_params);
 
+        let fetch_call = if endpoint.returns_json() {
+            "this.getJson(path, { signal, onUpdate })"
+        } else {
+            "this.getText(path, { signal, onUpdate })"
+        };
+
         if endpoint.query_params.is_empty() {
-            writeln!(output, "    return this.getJson(`{}`);", path).unwrap();
+            writeln!(output, "    const path = `{}`;", path).unwrap();
         } else {
             writeln!(output, "    const params = new URLSearchParams();").unwrap();
             for param in &endpoint.query_params {
+                let ident = sanitize_ident(&param.name);
                 if param.required {
                     writeln!(
                         output,
                         "    params.set('{}', String({}));",
-                        param.name, param.name
+                        param.name, ident
                     )
                     .unwrap();
                 } else {
                     writeln!(
                         output,
                         "    if ({} !== undefined) params.set('{}', String({}));",
-                        param.name, param.name, param.name
+                        ident, param.name, ident
                     )
                     .unwrap();
                 }
@@ -105,17 +123,13 @@ pub fn generate_api_methods(output: &mut String, endpoints: &[Endpoint]) {
                 path
             )
             .unwrap();
-
-            if endpoint.supports_csv {
-                writeln!(output, "    if (format === 'csv') {{").unwrap();
-                writeln!(output, "      return this.getText(path);").unwrap();
-                writeln!(output, "    }}").unwrap();
-                writeln!(output, "    return this.getJson(path);").unwrap();
-            } else {
-                writeln!(output, "    return this.getJson(path);").unwrap();
-            }
         }
 
+        if endpoint.supports_csv {
+            writeln!(output, "    if (format === 'csv') return this.getText(path, {{ signal, onUpdate }});").unwrap();
+        }
+        writeln!(output, "    return {};", fetch_call).unwrap();
+
         writeln!(output, "  }}\n").unwrap();
     }
 }
@@ -127,14 +141,19 @@ fn endpoint_to_method_name(endpoint: &Endpoint) -> String {
 fn build_method_params(endpoint: &Endpoint) -> String {
     let mut params = Vec::new();
     for param in &endpoint.path_params {
-        params.push(param.name.clone());
+        params.push(sanitize_ident(&param.name));
     }
     for param in &endpoint.query_params {
-        params.push(param.name.clone());
+        params.push(sanitize_ident(&param.name));
     }
     params.join(", ")
 }
 
+/// Strip characters invalid in JS identifiers (e.g. `[]` from `txId[]`).
+fn sanitize_ident(name: &str) -> String {
+    name.replace(['[', ']'], "")
+}
+
 fn build_path_template(path: &str, path_params: &[Parameter]) -> String {
     let mut result = path.to_string();
     for param in path_params {

crates/brk_bindgen/src/generators/javascript/client.rs

@@ -22,6 +22,20 @@ pub fn generate_base_client(output: &mut String) {
 const _isBrowser = typeof window !== 'undefined' && 'caches' in window;
 const _runIdle = (/** @type {{VoidFunction}} */ fn) => (globalThis.requestIdleCallback ?? setTimeout)(fn);
 const _defaultCacheName = '__BRK_CLIENT__';
+/** @param {{*}} v */
+const _addCamelGetters = (v) => {{
+  if (Array.isArray(v)) {{ v.forEach(_addCamelGetters); return v; }}
+  if (v && typeof v === 'object' && v.constructor === Object) {{
+    for (const k in v) {{
+      if (k.includes('_')) {{
+        const c = k.replace(/_([a-z])/g, (_, l) => l.toUpperCase());
+        if (!(c in v)) Object.defineProperty(v, c, {{ get() {{ return this[k]; }} }});
+      }}
+      _addCamelGetters(v[k]);
+    }}
+  }}
+  return v;
+}};
 
 /**
  * @param {{string|boolean|undefined}} cache
@@ -390,23 +404,28 @@ class BrkClientBase {{
 
   /**
    * @param {{string}} path
+   * @param {{{{ signal?: AbortSignal }}}} [options]
    * @returns {{Promise<Response>}}
    */
-  async get(path) {{
+  async get(path, {{ signal }} = {{}}) {{
     const url = `${{this.baseUrl}}${{path}}`;
-    const res = await fetch(url, {{ signal: AbortSignal.timeout(this.timeout) }});
+    const signals = [AbortSignal.timeout(this.timeout)];
+    if (signal) signals.push(signal);
+    const res = await fetch(url, {{ signal: AbortSignal.any(signals) }});
     if (!res.ok) throw new BrkError(`HTTP ${{res.status}}: ${{url}}`, res.status);
     return res;
   }}
 
   /**
-   * Make a GET request - races cache vs network, first to resolve calls onUpdate
+   * Make a GET request - races cache vs network, first to resolve calls onUpdate.
+   * Shared implementation backing `getJson` and `getText`.
    * @template T
    * @param {{string}} path
-   * @param {{(value: T) => void}} [onUpdate] - Called when data is available (may be called twice: cache then network)
+   * @param {{(res: Response) => Promise<T>}} parse - Response body reader
+   * @param {{{{ onUpdate?: (value: T) => void, signal?: AbortSignal }}}} [options]
    * @returns {{Promise<T>}}
    */
-  async getJson(path, onUpdate) {{
+  async _getCached(path, parse, {{ onUpdate, signal }} = {{}}) {{
     const url = `${{this.baseUrl}}${{path}}`;
     const cache = this._cache ?? await this._cachePromise;
 
@@ -418,51 +437,61 @@ class BrkClientBase {{
     const cachePromise = cache?.match(url).then(async (res) => {{
       cachedRes = res ?? null;
       if (!res) return null;
-      const json = await res.json();
+      const value = await parse(res);
       if (!resolved && onUpdate) {{
         resolved = true;
-        onUpdate(json);
+        onUpdate(value);
       }}
-      return json;
+      return value;
     }});
 
-    const networkPromise = this.get(path).then(async (res) => {{
+    const networkPromise = this.get(path, {{ signal }}).then(async (res) => {{
       const cloned = res.clone();
-      const json = await res.json();
+      const value = await parse(res);
       // Skip update if ETag matches and cache already delivered
       if (cachedRes?.headers.get('ETag') === res.headers.get('ETag')) {{
         if (!resolved && onUpdate) {{
           resolved = true;
-          onUpdate(json);
+          onUpdate(value);
         }}
-        return json;
+        return value;
       }}
       resolved = true;
-      if (onUpdate) {{
+      if (onUpdate) onUpdate(value);
-        onUpdate(json);
-      }}
       if (cache) _runIdle(() => cache.put(url, cloned));
-      return json;
+      return value;
     }});
 
     try {{
       return await networkPromise;
     }} catch (e) {{
       // Network failed - wait for cache
-      const cachedJson = await cachePromise?.catch(() => null);
+      const cachedValue = await cachePromise?.catch(() => null);
-      if (cachedJson) return cachedJson;
+      if (cachedValue != null) return cachedValue;
       throw e;
     }}
   }}
 
   /**
-   * Make a GET request and return raw text (for CSV responses)
+   * Make a GET request expecting a JSON response. Cached and supports `onUpdate`.
+   * @template T
    * @param {{string}} path
+   * @param {{{{ onUpdate?: (value: T) => void, signal?: AbortSignal }}}} [options]
+   * @returns {{Promise<T>}}
+   */
+  getJson(path, options) {{
+    return this._getCached(path, async (res) => _addCamelGetters(await res.json()), options);
+  }}
+
+  /**
+   * Make a GET request expecting a text response (text/plain, text/csv, ...).
+   * Cached and supports `onUpdate`, same as `getJson`.
+   * @param {{string}} path
+   * @param {{{{ onUpdate?: (value: string) => void, signal?: AbortSignal }}}} [options]
    * @returns {{Promise<string>}}
    */
-  async getText(path) {{
+  getText(path, options) {{
-    const res = await this.get(path);
+    return this._getCached(path, (res) => res.text(), options);
-    return res.text();
   }}
 
   /**
@@ -474,7 +503,7 @@ class BrkClientBase {{
    */
   async _fetchSeriesData(path, onUpdate) {{
     const wrappedOnUpdate = onUpdate ? (/** @type {{SeriesData<T>}} */ raw) => onUpdate(_wrapSeriesData(raw)) : undefined;
-    const raw = await this.getJson(path, wrappedOnUpdate);
+    const raw = await this.getJson(path, {{ onUpdate: wrappedOnUpdate }});
     return _wrapSeriesData(raw);
   }}
 }}
@@ -726,7 +755,12 @@ pub fn generate_structural_patterns(
         writeln!(output, " */").unwrap();
 
         if pattern.is_templated() {
-            writeln!(output, "function create{}(client, acc, disc) {{", pattern.name).unwrap();
+            writeln!(
+                output,
+                "function create{}(client, acc, disc) {{",
+                pattern.name
+            )
+            .unwrap();
         } else {
             writeln!(output, "function create{}(client, acc) {{", pattern.name).unwrap();
         }

crates/brk_bindgen/src/generators/javascript/tree.rs

@@ -108,15 +108,14 @@ pub fn generate_main_client(
     writeln!(output, "  constructor(options) {{").unwrap();
     writeln!(output, "    super(options);").unwrap();
     writeln!(output, "    /** @type {{SeriesTree}} */").unwrap();
-    writeln!(output, "    this.series = this._buildTree('');").unwrap();
+    writeln!(output, "    this.series = this._buildTree();").unwrap();
     writeln!(output, "  }}\n").unwrap();
 
     writeln!(output, "  /**").unwrap();
     writeln!(output, "   * @private").unwrap();
-    writeln!(output, "   * @param {{string}} basePath").unwrap();
     writeln!(output, "   * @returns {{SeriesTree}}").unwrap();
     writeln!(output, "   */").unwrap();
-    writeln!(output, "  _buildTree(basePath) {{").unwrap();
+    writeln!(output, "  _buildTree() {{").unwrap();
     writeln!(output, "    return {{").unwrap();
     let mut generated = BTreeSet::new();
     generate_tree_initializer(

crates/brk_bindgen/src/generators/python/api.rs

@@ -56,11 +56,7 @@ pub fn generate_main_client(output: &mut String, endpoints: &[Endpoint]) {
     )
     .unwrap();
     writeln!(output, "        \"\"\"").unwrap();
-    writeln!(
+    writeln!(output, "        return SeriesEndpoint(self, series, index)").unwrap();
-        output,
-        "        return SeriesEndpoint(self, series, index)"
-    )
-    .unwrap();
     writeln!(output).unwrap();
 
     // Generate helper methods
@@ -105,7 +101,7 @@ pub fn generate_api_methods(output: &mut String, endpoints: &[Endpoint]) {
                 .response_type
                 .as_deref()
                 .map(js_type_to_python)
-                .unwrap_or_else(|| "Any".to_string()),
+                .unwrap_or_else(|| "str".to_string()),
         );
 
         let return_type = if endpoint.supports_csv {
@@ -163,11 +159,17 @@ pub fn generate_api_methods(output: &mut String, endpoints: &[Endpoint]) {
         // Build path
         let path = build_path_template(&endpoint.path, &endpoint.path_params);
 
+        let fetch_method = if endpoint.returns_json() {
+            "get_json"
+        } else {
+            "get_text"
+        };
+
         if endpoint.query_params.is_empty() {
             if endpoint.path_params.is_empty() {
-                writeln!(output, "        return self.get_json('{}')", path).unwrap();
+                writeln!(output, "        return self.{}('{}')", fetch_method, path).unwrap();
             } else {
-                writeln!(output, "        return self.get_json(f'{}')", path).unwrap();
+                writeln!(output, "        return self.{}(f'{}')", fetch_method, path).unwrap();
             }
         } else {
             writeln!(output, "        params = []").unwrap();
@@ -201,9 +203,9 @@ pub fn generate_api_methods(output: &mut String, endpoints: &[Endpoint]) {
             if endpoint.supports_csv {
                 writeln!(output, "        if format == 'csv':").unwrap();
                 writeln!(output, "            return self.get_text(path)").unwrap();
-                writeln!(output, "        return self.get_json(path)").unwrap();
+                writeln!(output, "        return self.{}(path)", fetch_method).unwrap();
             } else {
-                writeln!(output, "        return self.get_json(path)").unwrap();
+                writeln!(output, "        return self.{}(path)", fetch_method).unwrap();
             }
         }
 

crates/brk_bindgen/src/generators/python/client.rs

@@ -684,7 +684,6 @@ pub fn generate_structural_patterns(
     writeln!(output, "# Reusable structural pattern classes\n").unwrap();
 
     for pattern in patterns {
-
         // Generate class
         if pattern.is_generic {
             writeln!(output, "class {}(Generic[T]):", pattern.name).unwrap();

crates/brk_bindgen/src/generators/rust/api.rs

@@ -93,7 +93,7 @@ pub fn generate_api_methods(output: &mut String, endpoints: &[Endpoint]) {
             .response_type
             .as_deref()
             .map(js_type_to_rust)
-            .unwrap_or_else(|| "serde_json::Value".to_string());
+            .unwrap_or_else(|| "String".to_string());
 
         let return_type = if endpoint.supports_csv {
             format!("FormatResponse<{}>", base_return_type)
@@ -132,29 +132,43 @@ pub fn generate_api_methods(output: &mut String, endpoints: &[Endpoint]) {
         .unwrap();
 
         let (path, index_arg) = build_path_template(endpoint);
+        let fetch_method = if endpoint.returns_json() {
+            "get_json"
+        } else {
+            "get_text"
+        };
 
         if endpoint.query_params.is_empty() {
             writeln!(
                 output,
-                "        self.base.get_json(&format!(\"{}\"{}))",
+                "        self.base.{}(&format!(\"{}\"{}))",
-                path, index_arg
+                fetch_method, path, index_arg
             )
             .unwrap();
         } else {
             writeln!(output, "        let mut query = Vec::new();").unwrap();
             for param in &endpoint.query_params {
-                if param.required {
+                let ident = sanitize_ident(&param.name);
+                let is_array = param.param_type.ends_with("[]");
+                if is_array {
+                    writeln!(
+                        output,
+                        "        for v in {} {{ query.push(format!(\"{}={{}}\", v)); }}",
+                        ident, param.name
+                    )
+                    .unwrap();
+                } else if param.required {
                     writeln!(
                         output,
                         "        query.push(format!(\"{}={{}}\", {}));",
-                        param.name, param.name
+                        param.name, ident
                     )
                     .unwrap();
                 } else {
                     writeln!(
                         output,
                         "        if let Some(v) = {} {{ query.push(format!(\"{}={{}}\", v)); }}",
-                        param.name, param.name
+                        ident, param.name
                     )
                     .unwrap();
                 }
@@ -177,12 +191,13 @@ pub fn generate_api_methods(output: &mut String, endpoints: &[Endpoint]) {
                 writeln!(output, "        }} else {{").unwrap();
                 writeln!(
                     output,
-                    "            self.base.get_json(&path).map(FormatResponse::Json)"
+                    "            self.base.{}(&path).map(FormatResponse::Json)",
+                    fetch_method
                 )
                 .unwrap();
                 writeln!(output, "        }}").unwrap();
             } else {
-                writeln!(output, "        self.base.get_json(&path)").unwrap();
+                writeln!(output, "        self.base.{}(&path)", fetch_method).unwrap();
             }
         }
 
@@ -198,26 +213,35 @@ fn build_method_params(endpoint: &Endpoint) -> String {
     let mut params = Vec::new();
     for param in &endpoint.path_params {
         let rust_type = param_type_to_rust(&param.param_type);
-        params.push(format!(", {}: {}", param.name, rust_type));
+        params.push(format!(", {}: {}", sanitize_ident(&param.name), rust_type));
     }
     for param in &endpoint.query_params {
         let rust_type = param_type_to_rust(&param.param_type);
+        let name = sanitize_ident(&param.name);
         if param.required {
-            params.push(format!(", {}: {}", param.name, rust_type));
+            params.push(format!(", {}: {}", name, rust_type));
         } else {
-            params.push(format!(", {}: Option<{}>", param.name, rust_type));
+            params.push(format!(", {}: Option<{}>", name, rust_type));
         }
     }
     params.join("")
 }
 
+/// Strip characters invalid in Rust identifiers (e.g. `[]` from `txId[]`).
+fn sanitize_ident(name: &str) -> String {
+    name.replace(['[', ']'], "")
+}
+
 /// Convert parameter type to Rust type for function signatures.
 fn param_type_to_rust(param_type: &str) -> String {
+    if let Some(inner) = param_type.strip_suffix("[]") {
+        return format!("&[{}]", param_type_to_rust(inner));
+    }
     match param_type {
         "string" | "*" => "&str".to_string(),
         "integer" | "number" => "i64".to_string(),
         "boolean" => "bool".to_string(),
-        other => other.to_string(), // Domain types like Index, SeriesName, Format
+        other => other.to_string(),
     }
 }
 

crates/brk_bindgen/src/openapi.rs

@@ -43,6 +43,11 @@ impl Endpoint {
         self.method == "GET" && !self.deprecated
     }
 
+    /// Returns true if this endpoint returns JSON (has a response_type extracted from application/json).
+    pub fn returns_json(&self) -> bool {
+        self.response_type.is_some()
+    }
+
     /// Returns the operation ID or generates one from the path.
     /// The returned string uses the raw case from the spec (typically camelCase).
     pub fn operation_name(&self) -> String {

crates/brk_bindgen/src/types/case.rs

@@ -74,6 +74,9 @@ pub fn escape_python_keyword(name: &str) -> String {
         "try", "while", "with", "yield",
     ];
 
+    // Strip characters invalid in identifiers (e.g. `[]` from `txId[]`)
+    let name = name.replace(['[', ']'], "");
+
     // Prefix with underscore if starts with digit
     let name = if name.starts_with(|c: char| c.is_ascii_digit()) {
         format!("_{}", name)

crates/brk_cli/Cargo.toml

@@ -13,7 +13,6 @@ brk_alloc = { workspace = true }
 brk_computer = { workspace = true }
 brk_error = { workspace = true, features = ["tokio", "vecdb"] }
 brk_indexer = { workspace = true }
-brk_iterator = { workspace = true }
 brk_logger = { workspace = true }
 brk_mempool = { workspace = true }
 brk_query = { workspace = true }
@@ -26,7 +25,7 @@ owo-colors = { workspace = true }
 tracing = { workspace = true }
 serde = { workspace = true }
 tokio = { workspace = true }
-toml = "1.0.7"
+toml = "1.1.2"
 vecdb = { workspace = true }
 
 [[bin]]

crates/brk_cli/README.md

@@ -1,11 +1,8 @@
 # BRK CLI
 
-Command-line interface for running a Bitcoin Research Kit instance.
+Run your own Bitcoin Research Kit instance. One binary, one command. Full sync in ~4-7h depending on hardware. ~44% disk overhead vs 250% for mempool/electrs.
 
-## Demo
+[bitview.space](https://bitview.space) is the official free hosted instance.
-
-- [bitview.space](https://bitview.space) - web interface
-- [bitview.space/api](https://bitview.space/api) - API docs
 
 ## Requirements
 

crates/brk_cli/src/main.rs

@@ -10,7 +10,6 @@ use brk_alloc::Mimalloc;
 use brk_computer::Computer;
 use brk_error::Result;
 use brk_indexer::Indexer;
-use brk_iterator::Blocks;
 use brk_mempool::Mempool;
 use brk_query::AsyncQuery;
 use brk_reader::Reader;
@@ -37,8 +36,6 @@ pub fn main() -> anyhow::Result<()> {
 
     let reader = Reader::new(config.blocksdir(), &client);
 
-    let blocks = Blocks::new(&client, &reader);
-
     let mut indexer = Indexer::forced_import(&config.brkdir())?;
 
     #[cfg(not(debug_assertions))]
@@ -52,7 +49,7 @@ pub fn main() -> anyhow::Result<()> {
             info!("Indexing {blocks_behind} blocks before starting server...");
             info!("---");
             sleep(Duration::from_secs(10));
-            indexer.index(&blocks, &client, &exit)?;
+            indexer.index(&reader, &client, &exit)?;
             drop(indexer);
             Mimalloc::collect();
             indexer = Indexer::forced_import(&config.brkdir())?;
@@ -102,14 +99,14 @@ pub fn main() -> anyhow::Result<()> {
         let total_start = Instant::now();
 
         let starting_indexes = if cfg!(debug_assertions) {
-            indexer.checked_index(&blocks, &client, &exit)?
+            indexer.checked_index(&reader, &client, &exit)?
         } else {
-            indexer.index(&blocks, &client, &exit)?
+            indexer.index(&reader, &client, &exit)?
         };
 
         Mimalloc::collect();
 
-        computer.compute(&indexer, starting_indexes, &reader, &exit)?;
+        computer.compute(&indexer, starting_indexes, &exit)?;
 
         info!("Total time: {:?}", total_start.elapsed());
         info!("Waiting for new blocks...");

crates/brk_cohort/Cargo.toml

@@ -14,3 +14,6 @@ brk_traversable = { workspace = true }
 vecdb = { workspace = true }
 rayon = { workspace = true }
 serde = { workspace = true }
+
+[package.metadata.cargo-machete]
+ignored = ["vecdb"]

crates/brk_cohort/src/addr.rs

@@ -75,7 +75,8 @@ impl<T> AddrGroups<T> {
     }
 
     pub fn iter_overlapping_mut(&mut self) -> impl Iterator<Item = &mut T> {
-        self.under_amount.iter_mut().chain(self.over_amount.iter_mut())
+        self.under_amount
+            .iter_mut()
+            .chain(self.over_amount.iter_mut())
     }
 }
-

crates/brk_cohort/src/age_range.rs

@@ -201,7 +201,29 @@ impl<T> AgeRange<T> {
     }
 
     pub fn from_array(arr: [T; 21]) -> Self {
-        let [a0, a1, a2, a3, a4, a5, a6, a7, a8, a9, a10, a11, a12, a13, a14, a15, a16, a17, a18, a19, a20] = arr;
+        let [
+            a0,
+            a1,
+            a2,
+            a3,
+            a4,
+            a5,
+            a6,
+            a7,
+            a8,
+            a9,
+            a10,
+            a11,
+            a12,
+            a13,
+            a14,
+            a15,
+            a16,
+            a17,
+            a18,
+            a19,
+            a20,
+        ] = arr;
         Self {
             under_1h: a0,
             _1h_to_1d: a1,

crates/brk_cohort/src/amount_range.rs

@@ -84,7 +84,11 @@ pub const AMOUNT_RANGE_NAMES: AmountRange<CohortName> = AmountRange {
     _10sats_to_100sats: CohortName::new("10sats_to_100sats", "10-100 sats", "10-100 Sats"),
     _100sats_to_1k_sats: CohortName::new("100sats_to_1k_sats", "100-1k sats", "100-1K Sats"),
     _1k_sats_to_10k_sats: CohortName::new("1k_sats_to_10k_sats", "1k-10k sats", "1K-10K Sats"),
-    _10k_sats_to_100k_sats: CohortName::new("10k_sats_to_100k_sats", "10k-100k sats", "10K-100K Sats"),
+    _10k_sats_to_100k_sats: CohortName::new(
+        "10k_sats_to_100k_sats",
+        "10k-100k sats",
+        "10K-100K Sats",
+    ),
     _100k_sats_to_1m_sats: CohortName::new("100k_sats_to_1m_sats", "100k-1M sats", "100K-1M Sats"),
     _1m_sats_to_10m_sats: CohortName::new("1m_sats_to_10m_sats", "1M-10M sats", "1M-10M Sats"),
     _10m_sats_to_1btc: CohortName::new("10m_sats_to_1btc", "0.1-1 BTC", "0.1-1 BTC"),

crates/brk_cohort/src/by_type.rs

@@ -1,16 +1,34 @@
 use std::ops::{Add, AddAssign};
 
+use brk_traversable::Traversable;
 use brk_types::OutputType;
+use rayon::prelude::*;
 
-use super::{SpendableType, UnspendableType};
+use super::{Filter, SpendableType, UnspendableType};
 
-#[derive(Default, Clone, Debug)]
+pub const OP_RETURN: &str = "op_return";
+
+#[derive(Default, Clone, Debug, Traversable)]
 pub struct ByType<T> {
+    #[traversable(flatten)]
     pub spendable: SpendableType<T>,
+    #[traversable(flatten)]
     pub unspendable: UnspendableType<T>,
 }
 
 impl<T> ByType<T> {
+    pub fn try_new<F, E>(mut create: F) -> Result<Self, E>
+    where
+        F: FnMut(Filter, &'static str) -> Result<T, E>,
+    {
+        Ok(Self {
+            spendable: SpendableType::try_new(&mut create)?,
+            unspendable: UnspendableType {
+                op_return: create(Filter::Type(OutputType::OpReturn), OP_RETURN)?,
+            },
+        })
+    }
+
     pub fn get(&self, output_type: OutputType) -> &T {
         match output_type {
             OutputType::P2PK65 => &self.spendable.p2pk65,
@@ -44,6 +62,49 @@ impl<T> ByType<T> {
             OutputType::OpReturn => &mut self.unspendable.op_return,
         }
     }
+
+    pub fn iter(&self) -> impl Iterator<Item = &T> {
+        self.spendable
+            .iter()
+            .chain(std::iter::once(&self.unspendable.op_return))
+    }
+
+    pub fn iter_mut(&mut self) -> impl Iterator<Item = &mut T> {
+        self.spendable
+            .iter_mut()
+            .chain(std::iter::once(&mut self.unspendable.op_return))
+    }
+
+    pub fn par_iter_mut(&mut self) -> impl ParallelIterator<Item = &mut T>
+    where
+        T: Send + Sync,
+    {
+        let Self {
+            spendable,
+            unspendable,
+        } = self;
+        spendable
+            .par_iter_mut()
+            .chain([&mut unspendable.op_return].into_par_iter())
+    }
+
+    pub fn iter_typed(&self) -> impl Iterator<Item = (OutputType, &T)> {
+        self.spendable
+            .iter_typed()
+            .chain(std::iter::once((
+                OutputType::OpReturn,
+                &self.unspendable.op_return,
+            )))
+    }
+
+    pub fn iter_typed_mut(&mut self) -> impl Iterator<Item = (OutputType, &mut T)> {
+        self.spendable
+            .iter_typed_mut()
+            .chain(std::iter::once((
+                OutputType::OpReturn,
+                &mut self.unspendable.op_return,
+            )))
+    }
 }
 
 impl<T> Add for ByType<T>

crates/brk_cohort/src/lib.rs

@@ -1,8 +1,8 @@
 #![doc = include_str!("../README.md")]
 
 mod addr;
-mod amount_filter;
 mod age_range;
+mod amount_filter;
 mod amount_range;
 mod by_addr_type;
 mod by_any_addr;
@@ -30,8 +30,8 @@ mod utxo;
 pub use brk_types::{Age, Term};
 
 pub use addr::*;
-pub use amount_filter::*;
 pub use age_range::*;
+pub use amount_filter::*;
 pub use amount_range::*;
 pub use by_addr_type::*;
 pub use by_any_addr::*;

crates/brk_cohort/src/loss.rs

@@ -153,6 +153,6 @@ impl<T> Loss<T> {
             .into_iter()
             .rev()
             .enumerate()
-            .map(move |(n, threshold)| (threshold, &ranges[len - 1 - n..]))
+            .map(move |(n, threshold)| (threshold, &ranges[len - 2 - n..]))
     }
 }

crates/brk_cohort/src/over_amount.rs

@@ -43,19 +43,31 @@ pub const OVER_AMOUNT_NAMES: OverAmount<CohortName> = OverAmount {
 pub const OVER_AMOUNT_FILTERS: OverAmount<Filter> = OverAmount {
     _1sat: Filter::Amount(AmountFilter::GreaterOrEqual(OVER_AMOUNT_THRESHOLDS._1sat)),
     _10sats: Filter::Amount(AmountFilter::GreaterOrEqual(OVER_AMOUNT_THRESHOLDS._10sats)),
-    _100sats: Filter::Amount(AmountFilter::GreaterOrEqual(OVER_AMOUNT_THRESHOLDS._100sats)),
+    _100sats: Filter::Amount(AmountFilter::GreaterOrEqual(
-    _1k_sats: Filter::Amount(AmountFilter::GreaterOrEqual(OVER_AMOUNT_THRESHOLDS._1k_sats)),
+        OVER_AMOUNT_THRESHOLDS._100sats,
-    _10k_sats: Filter::Amount(AmountFilter::GreaterOrEqual(OVER_AMOUNT_THRESHOLDS._10k_sats)),
+    )),
+    _1k_sats: Filter::Amount(AmountFilter::GreaterOrEqual(
+        OVER_AMOUNT_THRESHOLDS._1k_sats,
+    )),
+    _10k_sats: Filter::Amount(AmountFilter::GreaterOrEqual(
+        OVER_AMOUNT_THRESHOLDS._10k_sats,
+    )),
     _100k_sats: Filter::Amount(AmountFilter::GreaterOrEqual(
         OVER_AMOUNT_THRESHOLDS._100k_sats,
     )),
-    _1m_sats: Filter::Amount(AmountFilter::GreaterOrEqual(OVER_AMOUNT_THRESHOLDS._1m_sats)),
+    _1m_sats: Filter::Amount(AmountFilter::GreaterOrEqual(
-    _10m_sats: Filter::Amount(AmountFilter::GreaterOrEqual(OVER_AMOUNT_THRESHOLDS._10m_sats)),
+        OVER_AMOUNT_THRESHOLDS._1m_sats,
+    )),
+    _10m_sats: Filter::Amount(AmountFilter::GreaterOrEqual(
+        OVER_AMOUNT_THRESHOLDS._10m_sats,
+    )),
     _1btc: Filter::Amount(AmountFilter::GreaterOrEqual(OVER_AMOUNT_THRESHOLDS._1btc)),
     _10btc: Filter::Amount(AmountFilter::GreaterOrEqual(OVER_AMOUNT_THRESHOLDS._10btc)),
     _100btc: Filter::Amount(AmountFilter::GreaterOrEqual(OVER_AMOUNT_THRESHOLDS._100btc)),
     _1k_btc: Filter::Amount(AmountFilter::GreaterOrEqual(OVER_AMOUNT_THRESHOLDS._1k_btc)),
-    _10k_btc: Filter::Amount(AmountFilter::GreaterOrEqual(OVER_AMOUNT_THRESHOLDS._10k_btc)),
+    _10k_btc: Filter::Amount(AmountFilter::GreaterOrEqual(
+        OVER_AMOUNT_THRESHOLDS._10k_btc,
+    )),
 };
 
 #[derive(Default, Clone, Traversable, Serialize)]

crates/brk_cohort/src/profit.rs

@@ -16,10 +16,26 @@ pub const PROFIT_NAMES: Profit<CohortName> = Profit {
     _70pct: CohortName::new("utxos_over_70pct_in_profit", ">=70%", "Over 70% in Profit"),
     _80pct: CohortName::new("utxos_over_80pct_in_profit", ">=80%", "Over 80% in Profit"),
     _90pct: CohortName::new("utxos_over_90pct_in_profit", ">=90%", "Over 90% in Profit"),
-    _100pct: CohortName::new("utxos_over_100pct_in_profit", ">=100%", "Over 100% in Profit"),
+    _100pct: CohortName::new(
-    _200pct: CohortName::new("utxos_over_200pct_in_profit", ">=200%", "Over 200% in Profit"),
+        "utxos_over_100pct_in_profit",
-    _300pct: CohortName::new("utxos_over_300pct_in_profit", ">=300%", "Over 300% in Profit"),
+        ">=100%",
-    _500pct: CohortName::new("utxos_over_500pct_in_profit", ">=500%", "Over 500% in Profit"),
+        "Over 100% in Profit",
+    ),
+    _200pct: CohortName::new(
+        "utxos_over_200pct_in_profit",
+        ">=200%",
+        "Over 200% in Profit",
+    ),
+    _300pct: CohortName::new(
+        "utxos_over_300pct_in_profit",
+        ">=300%",
+        "Over 300% in Profit",
+    ),
+    _500pct: CohortName::new(
+        "utxos_over_500pct_in_profit",
+        ">=500%",
+        "Over 500% in Profit",
+    ),
 };
 
 /// Number of profit thresholds.
@@ -192,6 +208,6 @@ impl<T> Profit<T> {
             .into_iter()
             .rev()
             .enumerate()
-            .map(move |(n, threshold)| (threshold, &ranges[..n + 1]))
+            .map(move |(n, threshold)| (threshold, &ranges[..n + 2]))
     }
 }

crates/brk_cohort/src/profitability_range.rs

@@ -83,31 +83,131 @@ pub fn compute_profitability_boundaries(spot: Cents) -> [Cents; PROFITABILITY_BO
 
 /// Profitability range names (25 ranges, from most profitable to most in loss)
 pub const PROFITABILITY_RANGE_NAMES: ProfitabilityRange<CohortName> = ProfitabilityRange {
-    over_1000pct_in_profit: CohortName::new("utxos_over_1000pct_in_profit", "+>1000%", "Over 1000% in Profit"),
+    over_1000pct_in_profit: CohortName::new(
-    _500pct_to_1000pct_in_profit: CohortName::new("utxos_500pct_to_1000pct_in_profit", "+500-1000%", "500-1000% in Profit"),
+        "utxos_over_1000pct_in_profit",
-    _300pct_to_500pct_in_profit: CohortName::new("utxos_300pct_to_500pct_in_profit", "+300-500%", "300-500% in Profit"),
+        "+>1000%",
-    _200pct_to_300pct_in_profit: CohortName::new("utxos_200pct_to_300pct_in_profit", "+200-300%", "200-300% in Profit"),
+        "Over 1000% in Profit",
-    _100pct_to_200pct_in_profit: CohortName::new("utxos_100pct_to_200pct_in_profit", "+100-200%", "100-200% in Profit"),
+    ),
-    _90pct_to_100pct_in_profit: CohortName::new("utxos_90pct_to_100pct_in_profit", "+90-100%", "90-100% in Profit"),
+    _500pct_to_1000pct_in_profit: CohortName::new(
-    _80pct_to_90pct_in_profit: CohortName::new("utxos_80pct_to_90pct_in_profit", "+80-90%", "80-90% in Profit"),
+        "utxos_500pct_to_1000pct_in_profit",
-    _70pct_to_80pct_in_profit: CohortName::new("utxos_70pct_to_80pct_in_profit", "+70-80%", "70-80% in Profit"),
+        "+500-1000%",
-    _60pct_to_70pct_in_profit: CohortName::new("utxos_60pct_to_70pct_in_profit", "+60-70%", "60-70% in Profit"),
+        "500-1000% in Profit",
-    _50pct_to_60pct_in_profit: CohortName::new("utxos_50pct_to_60pct_in_profit", "+50-60%", "50-60% in Profit"),
+    ),
-    _40pct_to_50pct_in_profit: CohortName::new("utxos_40pct_to_50pct_in_profit", "+40-50%", "40-50% in Profit"),
+    _300pct_to_500pct_in_profit: CohortName::new(
-    _30pct_to_40pct_in_profit: CohortName::new("utxos_30pct_to_40pct_in_profit", "+30-40%", "30-40% in Profit"),
+        "utxos_300pct_to_500pct_in_profit",
-    _20pct_to_30pct_in_profit: CohortName::new("utxos_20pct_to_30pct_in_profit", "+20-30%", "20-30% in Profit"),
+        "+300-500%",
-    _10pct_to_20pct_in_profit: CohortName::new("utxos_10pct_to_20pct_in_profit", "+10-20%", "10-20% in Profit"),
+        "300-500% in Profit",
-    _0pct_to_10pct_in_profit: CohortName::new("utxos_0pct_to_10pct_in_profit", "+0-10%", "0-10% in Profit"),
+    ),
-    _0pct_to_10pct_in_loss: CohortName::new("utxos_0pct_to_10pct_in_loss", "-0-10%", "0-10% in Loss"),
+    _200pct_to_300pct_in_profit: CohortName::new(
-    _10pct_to_20pct_in_loss: CohortName::new("utxos_10pct_to_20pct_in_loss", "-10-20%", "10-20% in Loss"),
+        "utxos_200pct_to_300pct_in_profit",
-    _20pct_to_30pct_in_loss: CohortName::new("utxos_20pct_to_30pct_in_loss", "-20-30%", "20-30% in Loss"),
+        "+200-300%",
-    _30pct_to_40pct_in_loss: CohortName::new("utxos_30pct_to_40pct_in_loss", "-30-40%", "30-40% in Loss"),
+        "200-300% in Profit",
-    _40pct_to_50pct_in_loss: CohortName::new("utxos_40pct_to_50pct_in_loss", "-40-50%", "40-50% in Loss"),
+    ),
-    _50pct_to_60pct_in_loss: CohortName::new("utxos_50pct_to_60pct_in_loss", "-50-60%", "50-60% in Loss"),
+    _100pct_to_200pct_in_profit: CohortName::new(
-    _60pct_to_70pct_in_loss: CohortName::new("utxos_60pct_to_70pct_in_loss", "-60-70%", "60-70% in Loss"),
+        "utxos_100pct_to_200pct_in_profit",
-    _70pct_to_80pct_in_loss: CohortName::new("utxos_70pct_to_80pct_in_loss", "-70-80%", "70-80% in Loss"),
+        "+100-200%",
-    _80pct_to_90pct_in_loss: CohortName::new("utxos_80pct_to_90pct_in_loss", "-80-90%", "80-90% in Loss"),
+        "100-200% in Profit",
-    _90pct_to_100pct_in_loss: CohortName::new("utxos_90pct_to_100pct_in_loss", "-90-100%", "90-100% in Loss"),
+    ),
+    _90pct_to_100pct_in_profit: CohortName::new(
+        "utxos_90pct_to_100pct_in_profit",
+        "+90-100%",
+        "90-100% in Profit",
+    ),
+    _80pct_to_90pct_in_profit: CohortName::new(
+        "utxos_80pct_to_90pct_in_profit",
+        "+80-90%",
+        "80-90% in Profit",
+    ),
+    _70pct_to_80pct_in_profit: CohortName::new(
+        "utxos_70pct_to_80pct_in_profit",
+        "+70-80%",
+        "70-80% in Profit",
+    ),
+    _60pct_to_70pct_in_profit: CohortName::new(
+        "utxos_60pct_to_70pct_in_profit",
+        "+60-70%",
+        "60-70% in Profit",
+    ),
+    _50pct_to_60pct_in_profit: CohortName::new(
+        "utxos_50pct_to_60pct_in_profit",
+        "+50-60%",
+        "50-60% in Profit",
+    ),
+    _40pct_to_50pct_in_profit: CohortName::new(
+        "utxos_40pct_to_50pct_in_profit",
+        "+40-50%",
+        "40-50% in Profit",
+    ),
+    _30pct_to_40pct_in_profit: CohortName::new(
+        "utxos_30pct_to_40pct_in_profit",
+        "+30-40%",
+        "30-40% in Profit",
+    ),
+    _20pct_to_30pct_in_profit: CohortName::new(
+        "utxos_20pct_to_30pct_in_profit",
+        "+20-30%",
+        "20-30% in Profit",
+    ),
+    _10pct_to_20pct_in_profit: CohortName::new(
+        "utxos_10pct_to_20pct_in_profit",
+        "+10-20%",
+        "10-20% in Profit",
+    ),
+    _0pct_to_10pct_in_profit: CohortName::new(
+        "utxos_0pct_to_10pct_in_profit",
+        "+0-10%",
+        "0-10% in Profit",
+    ),
+    _0pct_to_10pct_in_loss: CohortName::new(
+        "utxos_0pct_to_10pct_in_loss",
+        "-0-10%",
+        "0-10% in Loss",
+    ),
+    _10pct_to_20pct_in_loss: CohortName::new(
+        "utxos_10pct_to_20pct_in_loss",
+        "-10-20%",
+        "10-20% in Loss",
+    ),
+    _20pct_to_30pct_in_loss: CohortName::new(
+        "utxos_20pct_to_30pct_in_loss",
+        "-20-30%",
+        "20-30% in Loss",
+    ),
+    _30pct_to_40pct_in_loss: CohortName::new(
+        "utxos_30pct_to_40pct_in_loss",
+        "-30-40%",
+        "30-40% in Loss",
+    ),
+    _40pct_to_50pct_in_loss: CohortName::new(
+        "utxos_40pct_to_50pct_in_loss",
+        "-40-50%",
+        "40-50% in Loss",
+    ),
+    _50pct_to_60pct_in_loss: CohortName::new(
+        "utxos_50pct_to_60pct_in_loss",
+        "-50-60%",
+        "50-60% in Loss",
+    ),
+    _60pct_to_70pct_in_loss: CohortName::new(
+        "utxos_60pct_to_70pct_in_loss",
+        "-60-70%",
+        "60-70% in Loss",
+    ),
+    _70pct_to_80pct_in_loss: CohortName::new(
+        "utxos_70pct_to_80pct_in_loss",
+        "-70-80%",
+        "70-80% in Loss",
+    ),
+    _80pct_to_90pct_in_loss: CohortName::new(
+        "utxos_80pct_to_90pct_in_loss",
+        "-80-90%",
+        "80-90% in Loss",
+    ),
+    _90pct_to_100pct_in_loss: CohortName::new(
+        "utxos_90pct_to_100pct_in_loss",
+        "-90-100%",
+        "90-100% in Loss",
+    ),
 };
 
 impl ProfitabilityRange<CohortName> {

crates/brk_cohort/src/spendable_type.rs

@@ -116,6 +116,23 @@ impl<T> SpendableType<T> {
         })
     }
 
+    pub fn get(&self, output_type: OutputType) -> &T {
+        match output_type {
+            OutputType::P2PK65 => &self.p2pk65,
+            OutputType::P2PK33 => &self.p2pk33,
+            OutputType::P2PKH => &self.p2pkh,
+            OutputType::P2MS => &self.p2ms,
+            OutputType::P2SH => &self.p2sh,
+            OutputType::P2WPKH => &self.p2wpkh,
+            OutputType::P2WSH => &self.p2wsh,
+            OutputType::P2TR => &self.p2tr,
+            OutputType::P2A => &self.p2a,
+            OutputType::Unknown => &self.unknown,
+            OutputType::Empty => &self.empty,
+            _ => unreachable!(),
+        }
+    }
+
     pub fn get_mut(&mut self, output_type: OutputType) -> &mut T {
         match output_type {
             OutputType::P2PK65 => &mut self.p2pk65,

crates/brk_cohort/src/utxo.rs

@@ -2,8 +2,8 @@ use brk_traversable::Traversable;
 use rayon::prelude::*;
 
 use crate::{
-    AgeRange, AmountRange, ByEpoch, OverAmount, UnderAmount, UnderAge, OverAge,
+    AgeRange, AmountRange, ByEpoch, ByTerm, Class, Filter, OverAge, OverAmount, SpendableType,
-    Class, SpendableType, ByTerm, Filter,
+    UnderAge, UnderAmount,
 };
 
 #[derive(Default, Clone, Traversable)]

crates/brk_computer/Cargo.toml

@@ -14,11 +14,8 @@ brk_error = { workspace = true, features = ["vecdb"] }
 brk_cohort = { workspace = true }
 brk_indexer = { workspace = true }
 brk_oracle = { workspace = true }
-brk_iterator = { workspace = true }
 brk_logger = { workspace = true }
-brk_reader = { workspace = true }
 brk_rpc = { workspace = true, features = ["corepc"] }
-brk_store = { workspace = true }
 brk_traversable = { workspace = true }
 brk_types = { workspace = true }
 derive_more = { workspace = true }
@@ -33,6 +30,7 @@ smallvec = { workspace = true }
 vecdb = { workspace = true }
 
 [dev-dependencies]
+brk_reader = { workspace = true }
 brk_alloc = { workspace = true }
 brk_bencher = { workspace = true }
 color-eyre = { workspace = true }

crates/brk_computer/examples/computer.rs

@@ -8,7 +8,6 @@ use std::{
 use brk_alloc::Mimalloc;
 use brk_computer::Computer;
 use brk_indexer::Indexer;
-use brk_iterator::Blocks;
 use brk_reader::Reader;
 use brk_rpc::{Auth, Client};
 use vecdb::Exit;
@@ -31,8 +30,6 @@ pub fn main() -> color_eyre::Result<()> {
 
     let reader = Reader::new(bitcoin_dir.join("blocks"), &client);
 
-    let blocks = Blocks::new(&client, &reader);
-
     let mut indexer = Indexer::forced_import(&outputs_dir)?;
 
     let exit = Exit::new();
@@ -42,7 +39,7 @@ pub fn main() -> color_eyre::Result<()> {
     let chain_height = client.get_last_height()?;
     let indexed_height = indexer.vecs.starting_height();
     if u32::from(chain_height).saturating_sub(u32::from(indexed_height)) > 1000 {
-        indexer.checked_index(&blocks, &client, &exit)?;
+        indexer.checked_index(&reader, &client, &exit)?;
         drop(indexer);
         Mimalloc::collect();
         indexer = Indexer::forced_import(&outputs_dir)?;
@@ -52,11 +49,11 @@ pub fn main() -> color_eyre::Result<()> {
 
     loop {
         let i = Instant::now();
-        let starting_indexes = indexer.checked_index(&blocks, &client, &exit)?;
+        let starting_indexes = indexer.checked_index(&reader, &client, &exit)?;
 
         Mimalloc::collect();
 
-        computer.compute(&indexer, starting_indexes, &reader, &exit)?;
+        computer.compute(&indexer, starting_indexes, &exit)?;
         dbg!(i.elapsed());
         sleep(Duration::from_secs(10));
     }

crates/brk_computer/examples/computer_bench.rs

@@ -5,7 +5,6 @@ use brk_bencher::Bencher;
 use brk_computer::Computer;
 use brk_error::Result;
 use brk_indexer::Indexer;
-use brk_iterator::Blocks;
 use brk_reader::Reader;
 use brk_rpc::{Auth, Client};
 use tracing::{debug, info};
@@ -28,8 +27,6 @@ pub fn main() -> Result<()> {
 
     let reader = Reader::new(bitcoin_dir.join("blocks"), &client);
 
-    let blocks = Blocks::new(&client, &reader);
-
     let mut indexer = Indexer::forced_import(&outputs_dir)?;
 
     let mut computer = Computer::forced_import(&outputs_benches_dir, &indexer)?;
@@ -47,13 +44,13 @@ pub fn main() -> Result<()> {
     });
 
     let i = Instant::now();
-    let starting_indexes = indexer.index(&blocks, &client, &exit)?;
+    let starting_indexes = indexer.index(&reader, &client, &exit)?;
     info!("Done in {:?}", i.elapsed());
 
     Mimalloc::collect();
 
     let i = Instant::now();
-    computer.compute(&indexer, starting_indexes, &reader, &exit)?;
+    computer.compute(&indexer, starting_indexes, &exit)?;
     info!("Done in {:?}", i.elapsed());
 
     // We want to benchmark the drop too

crates/brk_computer/examples/full_bench.rs

@@ -9,7 +9,6 @@ use brk_alloc::Mimalloc;
 use brk_bencher::Bencher;
 use brk_computer::Computer;
 use brk_indexer::Indexer;
-use brk_iterator::Blocks;
 use brk_reader::Reader;
 use brk_rpc::{Auth, Client};
 use tracing::{debug, info};
@@ -45,15 +44,13 @@ pub fn main() -> color_eyre::Result<()> {
 
     let reader = Reader::new(bitcoin_dir.join("blocks"), &client);
 
-    let blocks = Blocks::new(&client, &reader);
-
     let mut indexer = Indexer::forced_import(&outputs_dir)?;
 
     // Pre-run indexer if too far behind, then drop and reimport to reduce memory
     let chain_height = client.get_last_height()?;
     let indexed_height = indexer.vecs.starting_height();
     if chain_height.saturating_sub(*indexed_height) > 1000 {
-        indexer.index(&blocks, &client, &exit)?;
+        indexer.index(&reader, &client, &exit)?;
         drop(indexer);
         Mimalloc::collect();
         indexer = Indexer::forced_import(&outputs_dir)?;
@@ -63,13 +60,13 @@ pub fn main() -> color_eyre::Result<()> {
 
     loop {
         let i = Instant::now();
-        let starting_indexes = indexer.index(&blocks, &client, &exit)?;
+        let starting_indexes = indexer.index(&reader, &client, &exit)?;
         info!("Done in {:?}", i.elapsed());
 
         Mimalloc::collect();
 
         let i = Instant::now();
-        computer.compute(&indexer, starting_indexes, &reader, &exit)?;
+        computer.compute(&indexer, starting_indexes, &exit)?;
         info!("Done in {:?}", i.elapsed());
 
         sleep(Duration::from_secs(60));

crates/brk_computer/src/blocks/compute.rs

@@ -17,12 +17,10 @@ impl Vecs {
         starting_indexes: &Indexes,
         exit: &Exit,
     ) -> Result<()> {
-        // Sequential: time → lookback (dependency chain)
+        self.db.sync_bg_tasks()?;
-        self.time
+
-            .timestamp
+        // lookback depends on indexes.timestamp.monotonic
-            .compute(indexer, indexes, starting_indexes, exit)?;
+        self.lookback.compute(indexes, starting_indexes, exit)?;
-        self.lookback
-            .compute(&self.time, starting_indexes, exit)?;
 
         // Parallel: remaining sub-modules are independent of each other.
         // size depends on lookback (already computed above).
@@ -40,8 +38,7 @@ impl Vecs {
             let r1 = s.spawn(|| count.compute(indexer, starting_indexes, exit));
             let r2 = s.spawn(|| interval.compute(indexer, starting_indexes, exit));
             let r3 = s.spawn(|| weight.compute(indexer, starting_indexes, exit));
-            let r4 =
+            let r4 = s.spawn(|| difficulty.compute(indexer, indexes, starting_indexes, exit));
-                s.spawn(|| difficulty.compute(indexer, indexes, starting_indexes, exit));
             let r5 = s.spawn(|| halving.compute(indexes, starting_indexes, exit));
             size.compute(indexer, &*lookback, starting_indexes, exit)?;
             r1.join().unwrap()?;
@@ -52,8 +49,11 @@ impl Vecs {
             Ok(())
         })?;
 
-        let _lock = exit.lock();
+        let exit = exit.clone();
-        self.db.compact()?;
+        self.db.run_bg(move |db| {
+            let _lock = exit.lock();
+            db.compact_deferred_default()
+        });
         Ok(())
     }
 }

crates/brk_computer/src/blocks/count/import.rs

@@ -6,8 +6,8 @@ use super::Vecs;
 use crate::{
     indexes,
     internal::{
-        BlockCountTarget24h, BlockCountTarget1w, BlockCountTarget1m, BlockCountTarget1y,
+        BlockCountTarget1m, BlockCountTarget1w, BlockCountTarget1y, BlockCountTarget24h,
-        CachedWindowStarts, PerBlockCumulativeRolling, ConstantVecs, Windows,
+        ConstantVecs, PerBlockCumulativeRolling, WindowStartVec, Windows,
     },
 };
 
@@ -16,14 +16,30 @@ impl Vecs {
         db: &Database,
         version: Version,
         indexes: &indexes::Vecs,
-        cached_starts: &CachedWindowStarts,
+        cached_starts: &Windows<&WindowStartVec>,
     ) -> Result<Self> {
         Ok(Self {
             target: Windows {
-                _24h: ConstantVecs::new::<BlockCountTarget24h>("block_count_target_24h", version, indexes),
+                _24h: ConstantVecs::new::<BlockCountTarget24h>(
-                _1w: ConstantVecs::new::<BlockCountTarget1w>("block_count_target_1w", version, indexes),
+                    "block_count_target_24h",
-                _1m: ConstantVecs::new::<BlockCountTarget1m>("block_count_target_1m", version, indexes),
+                    version,
-                _1y: ConstantVecs::new::<BlockCountTarget1y>("block_count_target_1y", version, indexes),
+                    indexes,
+                ),
+                _1w: ConstantVecs::new::<BlockCountTarget1w>(
+                    "block_count_target_1w",
+                    version,
+                    indexes,
+                ),
+                _1m: ConstantVecs::new::<BlockCountTarget1m>(
+                    "block_count_target_1m",
+                    version,
+                    indexes,
+                ),
+                _1y: ConstantVecs::new::<BlockCountTarget1y>(
+                    "block_count_target_1y",
+                    version,
+                    indexes,
+                ),
             },
             total: PerBlockCumulativeRolling::forced_import(
                 db,

crates/brk_computer/src/blocks/count/vecs.rs

@@ -2,7 +2,7 @@ use brk_traversable::Traversable;
 use brk_types::{StoredU32, StoredU64};
 use vecdb::{Rw, StorageMode};
 
-use crate::internal::{PerBlockCumulativeRolling, ConstantVecs, Windows};
+use crate::internal::{ConstantVecs, PerBlockCumulativeRolling, Windows};
 
 #[derive(Traversable)]
 pub struct Vecs<M: StorageMode = Rw> {

crates/brk_computer/src/blocks/difficulty/compute.rs

@@ -30,7 +30,7 @@ impl Vecs {
 
         self.blocks_to_retarget.height.compute_transform(
             starting_indexes.height,
-            &indexes.height.identity,
+            &indexes.height.epoch,
             |(h, ..)| (h, StoredU32::from(h.left_before_next_diff_adj())),
             exit,
         )?;

crates/brk_computer/src/blocks/difficulty/import.rs

@@ -20,19 +20,15 @@ impl Vecs {
     ) -> Result<Self> {
         let v2 = Version::TWO;
 
-        let hashrate = LazyPerBlock::from_height_source::<DifficultyToHashF64>(
+        let hashrate = LazyPerBlock::from_height_source::<DifficultyToHashF64, _>(
             "difficulty_hashrate",
             version,
-            indexer.vecs.blocks.difficulty.read_only_boxed_clone(),
+            indexer.vecs.blocks.difficulty.read_only_clone(),
             indexes,
         );
 
-        let blocks_to_retarget = PerBlock::forced_import(
+        let blocks_to_retarget =
-            db,
+            PerBlock::forced_import(db, "blocks_to_retarget", version + v2, indexes)?;
-            "blocks_to_retarget",
-            version + v2,
-            indexes,
-        )?;
 
         let days_to_retarget = LazyPerBlock::from_computed::<BlocksToDaysF32>(
             "days_to_retarget",
@@ -44,7 +40,7 @@ impl Vecs {
         Ok(Self {
             value: Resolutions::forced_import(
                 "difficulty",
-                indexer.vecs.blocks.difficulty.read_only_boxed_clone(),
+                indexer.vecs.blocks.difficulty.read_only_clone(),
                 version,
                 indexes,
             ),

crates/brk_computer/src/blocks/difficulty/vecs.rs

@@ -2,7 +2,7 @@ use brk_traversable::Traversable;
 use brk_types::{BasisPointsSigned32, Epoch, StoredF32, StoredF64, StoredU32};
 use vecdb::{Rw, StorageMode};
 
-use crate::internal::{LazyPerBlock, PerBlock, Resolutions, PercentPerBlock};
+use crate::internal::{LazyPerBlock, PerBlock, PercentPerBlock, Resolutions};
 #[derive(Traversable)]
 pub struct Vecs<M: StorageMode = Rw> {
     pub value: Resolutions<StoredF64>,

crates/brk_computer/src/blocks/halving/compute.rs

@@ -21,7 +21,7 @@ impl Vecs {
 
         self.blocks_to_halving.height.compute_transform(
             starting_indexes.height,
-            &indexes.height.identity,
+            &indexes.height.halving,
             |(h, ..)| (h, StoredU32::from(h.left_before_next_halving())),
             exit,
         )?;

crates/brk_computer/src/blocks/halving/import.rs

@@ -16,9 +16,8 @@ impl Vecs {
     ) -> Result<Self> {
         let v2 = Version::TWO;
 
-        let blocks_to_halving = PerBlock::forced_import(
+        let blocks_to_halving =
-            db, "blocks_to_halving", version + v2, indexes,
+            PerBlock::forced_import(db, "blocks_to_halving", version + v2, indexes)?;
-        )?;
 
         let days_to_halving = LazyPerBlock::from_computed::<BlocksToDaysF32>(
             "days_to_halving",

crates/brk_computer/src/blocks/import.rs

@@ -10,8 +10,7 @@ use crate::{
 };
 
 use super::{
-    CountVecs, DifficultyVecs, HalvingVecs, IntervalVecs, LookbackVecs, SizeVecs, TimeVecs, Vecs,
+    CountVecs, DifficultyVecs, HalvingVecs, IntervalVecs, LookbackVecs, SizeVecs, Vecs, WeightVecs,
-    WeightVecs,
 };
 
 impl Vecs {
@@ -25,12 +24,11 @@ impl Vecs {
         let version = parent_version;
 
         let lookback = LookbackVecs::forced_import(&db, version)?;
-        let cached_starts = &lookback.cached_window_starts;
+        let cached_starts = lookback.cached_window_starts();
-        let count = CountVecs::forced_import(&db, version, indexes, cached_starts)?;
+        let count = CountVecs::forced_import(&db, version, indexes, &cached_starts)?;
-        let interval = IntervalVecs::forced_import(&db, version, indexes, cached_starts)?;
+        let interval = IntervalVecs::forced_import(&db, version, indexes, &cached_starts)?;
-        let size = SizeVecs::forced_import(&db, version, indexes, cached_starts)?;
+        let size = SizeVecs::forced_import(&db, version, indexes, &cached_starts)?;
-        let weight = WeightVecs::forced_import(&db, version, indexes, cached_starts, &size)?;
+        let weight = WeightVecs::forced_import(&db, version, indexes, &cached_starts, &size)?;
-        let time = TimeVecs::forced_import(&db, version, indexes)?;
         let difficulty = DifficultyVecs::forced_import(&db, version, indexer, indexes)?;
         let halving = HalvingVecs::forced_import(&db, version, indexes)?;
 
@@ -41,7 +39,6 @@ impl Vecs {
             interval,
             size,
             weight,
-            time,
             difficulty,
             halving,
         };

crates/brk_computer/src/blocks/interval/compute.rs

@@ -13,27 +13,26 @@ impl Vecs {
         exit: &Exit,
     ) -> Result<()> {
         let mut prev_timestamp = None;
-        self.0
+        self.0.compute(starting_indexes.height, exit, |vec| {
-            .compute(starting_indexes.height, exit, |vec| {
+            vec.compute_transform(
-                vec.compute_transform(
+                starting_indexes.height,
-                    starting_indexes.height,
+                &indexer.vecs.blocks.timestamp,
-                    &indexer.vecs.blocks.timestamp,
+                |(h, timestamp, ..)| {
-                    |(h, timestamp, ..)| {
+                    let interval = if let Some(prev_h) = h.decremented() {
-                        let interval = if let Some(prev_h) = h.decremented() {
+                        let prev = prev_timestamp.unwrap_or_else(|| {
-                            let prev = prev_timestamp.unwrap_or_else(|| {
+                            indexer.vecs.blocks.timestamp.collect_one(prev_h).unwrap()
-                                indexer.vecs.blocks.timestamp.collect_one(prev_h).unwrap()
+                        });
-                            });
+                        timestamp.checked_sub(prev).unwrap_or(Timestamp::ZERO)
-                            timestamp.checked_sub(prev).unwrap_or(Timestamp::ZERO)
+                    } else {
-                        } else {
+                        Timestamp::ZERO
-                            Timestamp::ZERO
+                    };
-                        };
+                    prev_timestamp = Some(timestamp);
-                        prev_timestamp = Some(timestamp);
+                    (h, interval)
-                        (h, interval)
+                },
-                    },
+                exit,
-                    exit,
+            )?;
-                )?;
+            Ok(())
-                Ok(())
+        })?;
-            })?;
 
         Ok(())
     }

crates/brk_computer/src/blocks/interval/import.rs

@@ -3,14 +3,17 @@ use brk_types::Version;
 use vecdb::Database;
 
 use super::Vecs;
-use crate::{indexes, internal::{CachedWindowStarts, PerBlockRollingAverage}};
+use crate::{
+    indexes,
+    internal::{PerBlockRollingAverage, WindowStartVec, Windows},
+};
 
 impl Vecs {
     pub(crate) fn forced_import(
         db: &Database,
         version: Version,
         indexes: &indexes::Vecs,
-        cached_starts: &CachedWindowStarts,
+        cached_starts: &Windows<&WindowStartVec>,
     ) -> Result<Self> {
         let interval = PerBlockRollingAverage::forced_import(
             db,

crates/brk_computer/src/blocks/interval/vecs.rs

@@ -8,5 +8,5 @@ use crate::internal::PerBlockRollingAverage;
 
 #[derive(Deref, DerefMut, Traversable)]
 pub struct Vecs<M: StorageMode = Rw>(
-    #[traversable(flatten)] pub PerBlockRollingAverage<Timestamp, M>,
+    #[traversable(flatten)] pub PerBlockRollingAverage<Timestamp, Timestamp, M>,
 );

crates/brk_computer/src/blocks/lookback.rs

@@ -1,20 +1,22 @@
 use brk_error::Result;
 use brk_traversable::Traversable;
 use brk_types::{Height, Indexes, Timestamp, Version};
-use vecdb::{AnyVec, CachedVec, Cursor, Database, EagerVec, Exit, ImportableVec, PcoVec, ReadableVec, Rw, StorageMode, VecIndex};
+use vecdb::{
+    AnyVec, CachedVec, Cursor, Database, EagerVec, Exit, ImportableVec, PcoVec, ReadableVec, Rw,
+    StorageMode, VecIndex,
+};
 
-use crate::internal::{CachedWindowStarts, Windows, WindowStarts};
+use crate::{
-
+    indexes,
-use super::time;
+    internal::{WindowStartVec, WindowStarts, Windows},
+};
 
 #[derive(Traversable)]
 pub struct Vecs<M: StorageMode = Rw> {
-    #[traversable(skip)]
-    pub cached_window_starts: CachedWindowStarts,
     pub _1h: M::Stored<EagerVec<PcoVec<Height, Height>>>,
-    pub _24h: M::Stored<EagerVec<PcoVec<Height, Height>>>, // 1d
+    pub _24h: CachedVec<M::Stored<EagerVec<PcoVec<Height, Height>>>>, // 1d
     pub _3d: M::Stored<EagerVec<PcoVec<Height, Height>>>,
-    pub _1w: M::Stored<EagerVec<PcoVec<Height, Height>>>, // 7d
+    pub _1w: CachedVec<M::Stored<EagerVec<PcoVec<Height, Height>>>>, // 7d
     pub _8d: M::Stored<EagerVec<PcoVec<Height, Height>>>,
     pub _9d: M::Stored<EagerVec<PcoVec<Height, Height>>>,
     pub _12d: M::Stored<EagerVec<PcoVec<Height, Height>>>,
@@ -22,7 +24,7 @@ pub struct Vecs<M: StorageMode = Rw> {
     pub _2w: M::Stored<EagerVec<PcoVec<Height, Height>>>, // 14d
     pub _21d: M::Stored<EagerVec<PcoVec<Height, Height>>>,
     pub _26d: M::Stored<EagerVec<PcoVec<Height, Height>>>,
-    pub _1m: M::Stored<EagerVec<PcoVec<Height, Height>>>, // 30d
+    pub _1m: CachedVec<M::Stored<EagerVec<PcoVec<Height, Height>>>>, // 30d
     pub _34d: M::Stored<EagerVec<PcoVec<Height, Height>>>,
     pub _55d: M::Stored<EagerVec<PcoVec<Height, Height>>>,
     pub _2m: M::Stored<EagerVec<PcoVec<Height, Height>>>, // 60d
@@ -39,7 +41,7 @@ pub struct Vecs<M: StorageMode = Rw> {
     pub _9m: M::Stored<EagerVec<PcoVec<Height, Height>>>, // 270d
     pub _350d: M::Stored<EagerVec<PcoVec<Height, Height>>>,
     pub _12m: M::Stored<EagerVec<PcoVec<Height, Height>>>, // 360d
-    pub _1y: M::Stored<EagerVec<PcoVec<Height, Height>>>,  // 365d
+    pub _1y: CachedVec<M::Stored<EagerVec<PcoVec<Height, Height>>>>,  // 365d
     pub _14m: M::Stored<EagerVec<PcoVec<Height, Height>>>, // 420d
     pub _2y: M::Stored<EagerVec<PcoVec<Height, Height>>>,  // 730d
     pub _26m: M::Stored<EagerVec<PcoVec<Height, Height>>>, // 780d
@@ -102,24 +104,55 @@ impl Vecs {
         let _14y = ImportableVec::forced_import(db, "height_14y_ago", version)?;
         let _26y = ImportableVec::forced_import(db, "height_26y_ago", version)?;
 
-        let cached_window_starts = CachedWindowStarts(Windows {
-            _24h: CachedVec::new(&_24h),
-            _1w: CachedVec::new(&_1w),
-            _1m: CachedVec::new(&_1m),
-            _1y: CachedVec::new(&_1y),
-        });
-
         Ok(Self {
-            cached_window_starts,
+            _1h,
-            _1h, _24h, _3d, _1w, _8d, _9d, _12d, _13d, _2w, _21d, _26d,
+            _24h: CachedVec::wrap(_24h),
-            _1m, _34d, _55d, _2m, _9w, _12w, _89d, _3m, _14w, _111d, _144d,
+            _3d,
-            _6m, _26w, _200d, _9m, _350d, _12m, _1y, _14m, _2y, _26m, _3y,
+            _1w: CachedVec::wrap(_1w),
-            _200w, _4y, _5y, _6y, _8y, _9y, _10y, _12y, _14y, _26y,
+            _8d,
+            _9d,
+            _12d,
+            _13d,
+            _2w,
+            _21d,
+            _26d,
+            _1m: CachedVec::wrap(_1m),
+            _34d,
+            _55d,
+            _2m,
+            _9w,
+            _12w,
+            _89d,
+            _3m,
+            _14w,
+            _111d,
+            _144d,
+            _6m,
+            _26w,
+            _200d,
+            _9m,
+            _350d,
+            _12m,
+            _1y: CachedVec::wrap(_1y),
+            _14m,
+            _2y,
+            _26m,
+            _3y,
+            _200w,
+            _4y,
+            _5y,
+            _6y,
+            _8y,
+            _9y,
+            _10y,
+            _12y,
+            _14y,
+            _26y,
         })
     }
 
-    pub fn window_starts(&self) -> WindowStarts<'_> {
+    pub fn cached_window_starts(&self) -> Windows<&WindowStartVec> {
-        WindowStarts {
+        Windows {
             _24h: &self._24h,
             _1w: &self._1w,
             _1m: &self._1m,
@@ -127,12 +160,20 @@ impl Vecs {
         }
     }
 
+    pub fn window_starts(&self) -> WindowStarts<'_> {
+        WindowStarts {
+            _24h: &self._24h.inner,
+            _1w: &self._1w.inner,
+            _1m: &self._1m.inner,
+            _1y: &self._1y.inner,
+        }
+    }
 
     pub fn start_vec(&self, days: usize) -> &EagerVec<PcoVec<Height, Height>> {
         match days {
-            1 => &self._24h,
+            1 => &self._24h.inner,
             3 => &self._3d,
-            7 => &self._1w,
+            7 => &self._1w.inner,
             8 => &self._8d,
             9 => &self._9d,
             12 => &self._12d,
@@ -140,7 +181,7 @@ impl Vecs {
             14 => &self._2w,
             21 => &self._21d,
             26 => &self._26d,
-            30 => &self._1m,
+            30 => &self._1m.inner,
             34 => &self._34d,
             55 => &self._55d,
             60 => &self._2m,
@@ -157,7 +198,7 @@ impl Vecs {
             270 => &self._9m,
             350 => &self._350d,
             360 => &self._12m,
-            365 => &self._1y,
+            365 => &self._1y.inner,
             420 => &self._14m,
             730 => &self._2y,
             780 => &self._26m,
@@ -178,80 +219,60 @@ impl Vecs {
 
     pub(crate) fn compute(
         &mut self,
-        time: &time::Vecs,
+        indexes: &indexes::Vecs,
         starting_indexes: &Indexes,
         exit: &Exit,
     ) -> Result<()> {
-        self.compute_rolling_start_hours(time, starting_indexes, exit, 1, |s| {
+        self.compute_rolling_start_hours(indexes, starting_indexes, exit, 1, |s| &mut s._1h)?;
-            &mut s._1h
+        self.compute_rolling_start(indexes, starting_indexes, exit, 1, |s| &mut s._24h.inner)?;
-        })?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 3, |s| &mut s._3d)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 1, |s| &mut s._24h)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 7, |s| &mut s._1w.inner)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 3, |s| &mut s._3d)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 8, |s| &mut s._8d)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 7, |s| &mut s._1w)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 9, |s| &mut s._9d)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 8, |s| &mut s._8d)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 12, |s| &mut s._12d)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 9, |s| &mut s._9d)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 13, |s| &mut s._13d)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 12, |s| &mut s._12d)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 14, |s| &mut s._2w)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 13, |s| &mut s._13d)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 21, |s| &mut s._21d)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 14, |s| &mut s._2w)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 26, |s| &mut s._26d)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 21, |s| &mut s._21d)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 30, |s| &mut s._1m.inner)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 26, |s| &mut s._26d)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 34, |s| &mut s._34d)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 30, |s| &mut s._1m)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 55, |s| &mut s._55d)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 34, |s| &mut s._34d)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 60, |s| &mut s._2m)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 55, |s| &mut s._55d)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 63, |s| &mut s._9w)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 60, |s| &mut s._2m)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 84, |s| &mut s._12w)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 63, |s| &mut s._9w)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 89, |s| &mut s._89d)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 84, |s| &mut s._12w)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 90, |s| &mut s._3m)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 89, |s| &mut s._89d)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 98, |s| &mut s._14w)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 90, |s| &mut s._3m)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 111, |s| &mut s._111d)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 98, |s| &mut s._14w)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 144, |s| &mut s._144d)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 111, |s| {
+        self.compute_rolling_start(indexes, starting_indexes, exit, 180, |s| &mut s._6m)?;
-            &mut s._111d
+        self.compute_rolling_start(indexes, starting_indexes, exit, 182, |s| &mut s._26w)?;
-        })?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 200, |s| &mut s._200d)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 144, |s| {
+        self.compute_rolling_start(indexes, starting_indexes, exit, 270, |s| &mut s._9m)?;
-            &mut s._144d
+        self.compute_rolling_start(indexes, starting_indexes, exit, 350, |s| &mut s._350d)?;
-        })?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 360, |s| &mut s._12m)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 180, |s| &mut s._6m)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 365, |s| &mut s._1y.inner)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 182, |s| &mut s._26w)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 420, |s| &mut s._14m)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 200, |s| {
+        self.compute_rolling_start(indexes, starting_indexes, exit, 730, |s| &mut s._2y)?;
-            &mut s._200d
+        self.compute_rolling_start(indexes, starting_indexes, exit, 780, |s| &mut s._26m)?;
-        })?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 1095, |s| &mut s._3y)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 270, |s| &mut s._9m)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 1400, |s| &mut s._200w)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 350, |s| {
+        self.compute_rolling_start(indexes, starting_indexes, exit, 1460, |s| &mut s._4y)?;
-            &mut s._350d
+        self.compute_rolling_start(indexes, starting_indexes, exit, 1825, |s| &mut s._5y)?;
-        })?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 2190, |s| &mut s._6y)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 360, |s| &mut s._12m)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 2920, |s| &mut s._8y)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 365, |s| &mut s._1y)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 3285, |s| &mut s._9y)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 420, |s| &mut s._14m)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 3650, |s| &mut s._10y)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 730, |s| &mut s._2y)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 4380, |s| &mut s._12y)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 780, |s| &mut s._26m)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 5110, |s| &mut s._14y)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 1095, |s| &mut s._3y)?;
+        self.compute_rolling_start(indexes, starting_indexes, exit, 9490, |s| &mut s._26y)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 1400, |s| {
-            &mut s._200w
-        })?;
-        self.compute_rolling_start(time, starting_indexes, exit, 1460, |s| &mut s._4y)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 1825, |s| &mut s._5y)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 2190, |s| &mut s._6y)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 2920, |s| &mut s._8y)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 3285, |s| &mut s._9y)?;
-        self.compute_rolling_start(time, starting_indexes, exit, 3650, |s| {
-            &mut s._10y
-        })?;
-        self.compute_rolling_start(time, starting_indexes, exit, 4380, |s| {
-            &mut s._12y
-        })?;
-        self.compute_rolling_start(time, starting_indexes, exit, 5110, |s| {
-            &mut s._14y
-        })?;
-        self.compute_rolling_start(time, starting_indexes, exit, 9490, |s| {
-            &mut s._26y
-        })?;
 
         Ok(())
     }
 
     fn compute_rolling_start<F>(
         &mut self,
-        time: &time::Vecs,
+        indexes: &indexes::Vecs,
         starting_indexes: &Indexes,
         exit: &Exit,
         days: usize,
@@ -260,14 +281,18 @@ impl Vecs {
     where
         F: FnOnce(&mut Self) -> &mut EagerVec<PcoVec<Height, Height>>,
     {
-        self.compute_rolling_start_inner(time, starting_indexes, exit, get_field, |t, prev_ts| {
+        self.compute_rolling_start_inner(
-            t.difference_in_days_between(prev_ts) >= days
+            indexes,
-        })
+            starting_indexes,
+            exit,
+            get_field,
+            |t, prev_ts| t.difference_in_days_between(prev_ts) >= days,
+        )
     }
 
     fn compute_rolling_start_hours<F>(
         &mut self,
-        time: &time::Vecs,
+        indexes: &indexes::Vecs,
         starting_indexes: &Indexes,
         exit: &Exit,
         hours: usize,
@@ -276,14 +301,18 @@ impl Vecs {
     where
         F: FnOnce(&mut Self) -> &mut EagerVec<PcoVec<Height, Height>>,
     {
-        self.compute_rolling_start_inner(time, starting_indexes, exit, get_field, |t, prev_ts| {
+        self.compute_rolling_start_inner(
-            t.difference_in_hours_between(prev_ts) >= hours
+            indexes,
-        })
+            starting_indexes,
+            exit,
+            get_field,
+            |t, prev_ts| t.difference_in_hours_between(prev_ts) >= hours,
+        )
     }
 
     fn compute_rolling_start_inner<F, D>(
         &mut self,
-        time: &time::Vecs,
+        indexes: &indexes::Vecs,
         starting_indexes: &Indexes,
         exit: &Exit,
         get_field: F,
@@ -300,12 +329,12 @@ impl Vecs {
         } else {
             Height::ZERO
         };
-        let mut cursor = Cursor::new(&time.timestamp_monotonic);
+        let mut cursor = Cursor::new(&indexes.timestamp.monotonic);
         cursor.advance(prev.to_usize());
         let mut prev_ts = cursor.next().unwrap();
         Ok(field.compute_transform(
             starting_indexes.height,
-            &time.timestamp_monotonic,
+            &indexes.timestamp.monotonic,
             |(h, t, ..)| {
                 while expired(t, prev_ts) {
                     prev.increment();

crates/brk_computer/src/blocks/mod.rs

@@ -4,7 +4,6 @@ pub mod halving;
 pub mod interval;
 pub mod lookback;
 pub mod size;
-pub mod time;
 pub mod weight;
 
 mod compute;
@@ -19,7 +18,6 @@ pub use halving::Vecs as HalvingVecs;
 pub use interval::Vecs as IntervalVecs;
 pub use lookback::Vecs as LookbackVecs;
 pub use size::Vecs as SizeVecs;
-pub use time::Vecs as TimeVecs;
 pub use weight::Vecs as WeightVecs;
 
 pub const DB_NAME: &str = "blocks";
@@ -37,7 +35,7 @@ pub(crate) const ONE_TERA_HASH: f64 = 1_000_000_000_000.0;
 #[derive(Traversable)]
 pub struct Vecs<M: StorageMode = Rw> {
     #[traversable(skip)]
-    pub(crate) db: Database,
+    pub db: Database,
 
     pub count: CountVecs<M>,
     pub lookback: LookbackVecs<M>,
@@ -46,7 +44,6 @@ pub struct Vecs<M: StorageMode = Rw> {
     pub size: SizeVecs<M>,
     #[traversable(flatten)]
     pub weight: WeightVecs<M>,
-    pub time: TimeVecs<M>,
     pub difficulty: DifficultyVecs<M>,
     pub halving: HalvingVecs<M>,
 }

crates/brk_computer/src/blocks/size/import.rs

@@ -5,7 +5,7 @@ use vecdb::Database;
 use super::Vecs;
 use crate::{
     indexes,
-    internal::{CachedWindowStarts, PerBlockFull, PerBlockRolling},
+    internal::{PerBlockFull, PerBlockRolling, WindowStartVec, Windows},
 };
 
 impl Vecs {
@@ -13,7 +13,7 @@ impl Vecs {
         db: &Database,
         version: Version,
         indexes: &indexes::Vecs,
-        cached_starts: &CachedWindowStarts,
+        cached_starts: &Windows<&WindowStartVec>,
     ) -> Result<Self> {
         Ok(Self {
             vbytes: PerBlockFull::forced_import(

crates/brk_computer/src/blocks/time/compute.rs

@@ -1,34 +0,0 @@
-use brk_error::Result;
-use brk_indexer::Indexer;
-use vecdb::{Exit, ReadableVec};
-
-use super::Vecs;
-
-impl Vecs {
-    pub(crate) fn compute(
-        &mut self,
-        indexer: &Indexer,
-        starting_height: brk_types::Height,
-        exit: &Exit,
-    ) -> Result<()> {
-        let mut prev_timestamp_monotonic = None;
-        self.timestamp_monotonic.compute_transform(
-            starting_height,
-            &indexer.vecs.blocks.timestamp,
-            |(h, timestamp, this)| {
-                if prev_timestamp_monotonic.is_none()
-                    && let Some(prev_h) = h.decremented()
-                {
-                    prev_timestamp_monotonic.replace(this.collect_one(prev_h).unwrap());
-                }
-                let timestamp_monotonic =
-                    prev_timestamp_monotonic.map_or(timestamp, |prev_d| prev_d.max(timestamp));
-                prev_timestamp_monotonic.replace(timestamp_monotonic);
-                (h, timestamp_monotonic)
-            },
-            exit,
-        )?;
-
-        Ok(())
-    }
-}

crates/brk_computer/src/blocks/time/import.rs

@@ -1,66 +0,0 @@
-use brk_error::Result;
-use brk_types::{Date, Height, Version};
-use vecdb::{Database, EagerVec, ImportableVec, LazyVecFrom1, ReadableCloneableVec};
-
-use super::{TimestampIndexes, Vecs};
-use crate::indexes;
-
-impl Vecs {
-    pub(crate) fn forced_import(
-        db: &Database,
-        version: Version,
-        indexes: &indexes::Vecs,
-    ) -> Result<Self> {
-        let timestamp_monotonic = EagerVec::forced_import(db, "timestamp_monotonic", version)?;
-
-        Ok(Self {
-            date: LazyVecFrom1::init(
-                "date",
-                version,
-                timestamp_monotonic.read_only_boxed_clone(),
-                |_height: Height, timestamp| Date::from(timestamp),
-            ),
-            timestamp_monotonic,
-            timestamp: TimestampIndexes::forced_import(db, version, indexes)?,
-        })
-    }
-}
-
-impl TimestampIndexes {
-    fn forced_import(db: &Database, version: Version, indexes: &indexes::Vecs) -> Result<Self> {
-        macro_rules! period {
-            ($field:ident) => {
-                LazyVecFrom1::init(
-                    "timestamp",
-                    version,
-                    indexes.$field.first_height.read_only_boxed_clone(),
-                    |idx, _: Height| idx.to_timestamp(),
-                )
-            };
-        }
-
-        macro_rules! epoch {
-            ($field:ident) => {
-                ImportableVec::forced_import(db, "timestamp", version)?
-            };
-        }
-
-        Ok(Self(crate::internal::PerResolution {
-            minute10: period!(minute10),
-            minute30: period!(minute30),
-            hour1: period!(hour1),
-            hour4: period!(hour4),
-            hour12: period!(hour12),
-            day1: period!(day1),
-            day3: period!(day3),
-            week1: period!(week1),
-            month1: period!(month1),
-            month3: period!(month3),
-            month6: period!(month6),
-            year1: period!(year1),
-            year10: period!(year10),
-            halving: epoch!(halving),
-            epoch: epoch!(difficulty),
-        }))
-    }
-}

crates/brk_computer/src/blocks/time/mod.rs

@@ -1,5 +0,0 @@
-mod compute;
-mod import;
-mod vecs;
-
-pub use vecs::{TimestampIndexes, Vecs};

crates/brk_computer/src/blocks/time/vecs.rs

@@ -1,80 +0,0 @@
-use brk_error::Result;
-use brk_traversable::Traversable;
-use brk_types::{
-    Date, Day1, Day3, Epoch, Halving, Height, Hour1, Hour4, Hour12, Indexes,
-    Minute10, Minute30, Month1, Month3, Month6, Timestamp, Week1, Year1, Year10,
-};
-use derive_more::{Deref, DerefMut};
-use vecdb::{EagerVec, Exit, LazyVecFrom1, PcoVec, ReadableVec, Rw, StorageMode};
-
-use crate::{indexes, internal::PerResolution};
-#[derive(Traversable)]
-pub struct Vecs<M: StorageMode = Rw> {
-    pub date: LazyVecFrom1<Height, Date, Height, Timestamp>,
-    pub timestamp_monotonic: M::Stored<EagerVec<PcoVec<Height, Timestamp>>>,
-    pub timestamp: TimestampIndexes<M>,
-}
-
-/// Per-period timestamp indexes.
-///
-/// Time-based periods (minute10–year10) are lazy: `idx.to_timestamp()` is a pure
-/// function of the index, so no storage or decompression is needed.
-/// Epoch-based periods (halving, difficulty) are eager: their timestamps
-/// come from block data via `compute_indirect_sequential`.
-#[derive(Deref, DerefMut, Traversable)]
-#[traversable(transparent)]
-pub struct TimestampIndexes<M: StorageMode = Rw>(
-    #[allow(clippy::type_complexity)]
-    pub  PerResolution<
-        LazyVecFrom1<Minute10, Timestamp, Minute10, Height>,
-        LazyVecFrom1<Minute30, Timestamp, Minute30, Height>,
-        LazyVecFrom1<Hour1, Timestamp, Hour1, Height>,
-        LazyVecFrom1<Hour4, Timestamp, Hour4, Height>,
-        LazyVecFrom1<Hour12, Timestamp, Hour12, Height>,
-        LazyVecFrom1<Day1, Timestamp, Day1, Height>,
-        LazyVecFrom1<Day3, Timestamp, Day3, Height>,
-        LazyVecFrom1<Week1, Timestamp, Week1, Height>,
-        LazyVecFrom1<Month1, Timestamp, Month1, Height>,
-        LazyVecFrom1<Month3, Timestamp, Month3, Height>,
-        LazyVecFrom1<Month6, Timestamp, Month6, Height>,
-        LazyVecFrom1<Year1, Timestamp, Year1, Height>,
-        LazyVecFrom1<Year10, Timestamp, Year10, Height>,
-        M::Stored<EagerVec<PcoVec<Halving, Timestamp>>>,
-        M::Stored<EagerVec<PcoVec<Epoch, Timestamp>>>,
-    >,
-);
-
-impl TimestampIndexes {
-    /// Compute epoch timestamps via indirect lookup from block timestamps.
-    /// Time-based periods are lazy (idx.to_timestamp()) and need no compute.
-    pub(crate) fn compute(
-        &mut self,
-        indexer: &brk_indexer::Indexer,
-        indexes: &indexes::Vecs,
-        starting_indexes: &Indexes,
-        exit: &Exit,
-    ) -> Result<()> {
-        let prev_height = starting_indexes.height.decremented().unwrap_or_default();
-        self.halving.compute_indirect_sequential(
-            indexes
-                .height
-                .halving
-                .collect_one(prev_height)
-                .unwrap_or_default(),
-            &indexes.halving.first_height,
-            &indexer.vecs.blocks.timestamp,
-            exit,
-        )?;
-        self.epoch.compute_indirect_sequential(
-            indexes
-                .height
-                .epoch
-                .collect_one(prev_height)
-                .unwrap_or_default(),
-            &indexes.epoch.first_height,
-            &indexer.vecs.blocks.timestamp,
-            exit,
-        )?;
-        Ok(())
-    }
-}

crates/brk_computer/src/blocks/weight/import.rs

@@ -6,7 +6,7 @@ use super::Vecs;
 use crate::{
     blocks::SizeVecs,
     indexes,
-    internal::{CachedWindowStarts, LazyPerBlockRolling, PercentVec, VBytesToWeight},
+    internal::{LazyPerBlockRolling, PercentVec, VBytesToWeight, WindowStartVec, Windows},
 };
 
 impl Vecs {
@@ -14,7 +14,7 @@ impl Vecs {
         db: &Database,
         version: Version,
         indexes: &indexes::Vecs,
-        cached_starts: &CachedWindowStarts,
+        cached_starts: &Windows<&WindowStartVec>,
         size: &SizeVecs,
     ) -> Result<Self> {
         let weight = LazyPerBlockRolling::from_per_block_full::<VBytesToWeight>(

crates/brk_computer/src/cointime/activity/import.rs

@@ -6,7 +6,7 @@ use super::Vecs;
 use crate::{
     indexes,
     internal::{
-        CachedWindowStarts, LazyPerBlock, OneMinusF64, PerBlock, PerBlockCumulativeRolling,
+        LazyPerBlock, OneMinusF64, PerBlock, PerBlockCumulativeRolling, WindowStartVec, Windows,
     },
 };
 
@@ -15,7 +15,7 @@ impl Vecs {
         db: &Database,
         version: Version,
         indexes: &indexes::Vecs,
-        cached_starts: &CachedWindowStarts,
+        cached_starts: &Windows<&WindowStartVec>,
     ) -> Result<Self> {
         let liveliness = PerBlock::forced_import(db, "liveliness", version, indexes)?;
 
@@ -28,10 +28,18 @@ impl Vecs {
 
         Ok(Self {
             coinblocks_created: PerBlockCumulativeRolling::forced_import(
-                db, "coinblocks_created", version, indexes, cached_starts,
+                db,
+                "coinblocks_created",
+                version,
+                indexes,
+                cached_starts,
             )?,
             coinblocks_stored: PerBlockCumulativeRolling::forced_import(
-                db, "coinblocks_stored", version, indexes, cached_starts,
+                db,
+                "coinblocks_stored",
+                version,
+                indexes,
+                cached_starts,
             )?,
             liveliness,
             vaultedness,

crates/brk_computer/src/cointime/compute.rs

@@ -17,6 +17,8 @@ impl Vecs {
         distribution: &distribution::Vecs,
         exit: &Exit,
     ) -> Result<()> {
+        self.db.sync_bg_tasks()?;
+
         // Activity computes first (liveliness, vaultedness, etc.)
         self.activity
             .compute(starting_indexes, distribution, exit)?;
@@ -80,8 +82,11 @@ impl Vecs {
         r3?;
         r4?;
 
-        let _lock = exit.lock();
+        let exit = exit.clone();
-        self.db.compact()?;
+        self.db.run_bg(move |db| {
+            let _lock = exit.lock();
+            db.compact_deferred_default()
+        });
         Ok(())
     }
 }

crates/brk_computer/src/cointime/import.rs

@@ -13,14 +13,14 @@ use super::{
     ValueVecs, Vecs,
 };
 
-use crate::internal::CachedWindowStarts;
+use crate::internal::{WindowStartVec, Windows};
 
 impl Vecs {
     pub(crate) fn forced_import(
         parent_path: &Path,
         parent_version: Version,
         indexes: &indexes::Vecs,
-        cached_starts: &CachedWindowStarts,
+        cached_starts: &Windows<&WindowStartVec>,
     ) -> Result<Self> {
         let db = open_db(parent_path, DB_NAME, 250_000)?;
         let version = parent_version;

crates/brk_computer/src/cointime/prices/compute.rs

@@ -22,11 +22,8 @@ impl Vecs {
         let circulating_supply = &all_metrics.supply.total.btc.height;
         let realized_price = &all_metrics.realized.price.cents.height;
 
-        self.vaulted.compute_all(
+        self.vaulted
-            prices,
+            .compute_all(prices, starting_indexes, exit, |v| {
-            starting_indexes,
-            exit,
-            |v| {
                 Ok(v.compute_transform2(
                     starting_indexes.height,
                     realized_price,
@@ -36,14 +33,10 @@ impl Vecs {
                     },
                     exit,
                 )?)
-            },
+            })?;
-        )?;
 
-        self.active.compute_all(
+        self.active
-            prices,
+            .compute_all(prices, starting_indexes, exit, |v| {
-            starting_indexes,
-            exit,
-            |v| {
                 Ok(v.compute_transform2(
                     starting_indexes.height,
                     realized_price,
@@ -53,14 +46,10 @@ impl Vecs {
                     },
                     exit,
                 )?)
-            },
+            })?;
-        )?;
 
-        self.true_market_mean.compute_all(
+        self.true_market_mean
-            prices,
+            .compute_all(prices, starting_indexes, exit, |v| {
-            starting_indexes,
-            exit,
-            |v| {
                 Ok(v.compute_transform2(
                     starting_indexes.height,
                     &cap.investor.cents.height,
@@ -70,14 +59,10 @@ impl Vecs {
                     },
                     exit,
                 )?)
-            },
+            })?;
-        )?;
 
-        self.cointime.compute_all(
+        self.cointime
-            prices,
+            .compute_all(prices, starting_indexes, exit, |v| {
-            starting_indexes,
-            exit,
-            |v| {
                 Ok(v.compute_transform2(
                     starting_indexes.height,
                     &cap.cointime.cents.height,
@@ -87,8 +72,7 @@ impl Vecs {
                     },
                     exit,
                 )?)
-            },
+            })?;
-        )?;
 
         Ok(())
     }

crates/brk_computer/src/cointime/prices/import.rs

@@ -3,10 +3,7 @@ use brk_types::Version;
 use vecdb::Database;
 
 use super::Vecs;
-use crate::{
+use crate::{indexes, internal::PriceWithRatioExtendedPerBlock};
-    indexes,
-    internal::PriceWithRatioExtendedPerBlock,
-};
 
 impl Vecs {
     pub(crate) fn forced_import(

crates/brk_computer/src/cointime/supply/compute.rs

@@ -38,7 +38,8 @@ impl Vecs {
             exit,
         )?;
 
-        self.vaulted.compute(prices, starting_indexes.height, exit)?;
+        self.vaulted
+            .compute(prices, starting_indexes.height, exit)?;
         self.active.compute(prices, starting_indexes.height, exit)?;
 
         Ok(())

crates/brk_computer/src/cointime/supply/import.rs

@@ -3,7 +3,7 @@ use brk_types::Version;
 use vecdb::Database;
 
 use super::Vecs;
-use crate::{indexes, internal::AmountPerBlock};
+use crate::{indexes, internal::ValuePerBlock};
 
 impl Vecs {
     pub(crate) fn forced_import(
@@ -12,13 +12,8 @@ impl Vecs {
         indexes: &indexes::Vecs,
     ) -> Result<Self> {
         Ok(Self {
-            vaulted: AmountPerBlock::forced_import(
+            vaulted: ValuePerBlock::forced_import(db, "vaulted_supply", version, indexes)?,
-                db,
+            active: ValuePerBlock::forced_import(db, "active_supply", version, indexes)?,
-                "vaulted_supply",
-                version,
-                indexes,
-            )?,
-            active: AmountPerBlock::forced_import(db, "active_supply", version, indexes)?,
         })
     }
 }

crates/brk_computer/src/cointime/supply/vecs.rs

@@ -1,10 +1,10 @@
 use brk_traversable::Traversable;
 use vecdb::{Rw, StorageMode};
 
-use crate::internal::AmountPerBlock;
+use crate::internal::ValuePerBlock;
 
 #[derive(Traversable)]
 pub struct Vecs<M: StorageMode = Rw> {
-    pub vaulted: AmountPerBlock<M>,
+    pub vaulted: ValuePerBlock<M>,
-    pub active: AmountPerBlock<M>,
+    pub active: ValuePerBlock<M>,
 }

crates/brk_computer/src/cointime/value/compute.rs

@@ -31,52 +31,49 @@ impl Vecs {
                 Ok(())
             })?;
 
-        self.created
+        self.created.compute(starting_indexes.height, exit, |vec| {
-            .compute(starting_indexes.height, exit, |vec| {
+            vec.compute_multiply(
-                vec.compute_multiply(
+                starting_indexes.height,
-                    starting_indexes.height,
+                &prices.spot.usd.height,
-                    &prices.spot.usd.height,
+                &activity.coinblocks_created.block,
-                    &activity.coinblocks_created.block,
+                exit,
-                    exit,
+            )?;
-                )?;
+            Ok(())
-                Ok(())
+        })?;
-            })?;
 
-        self.stored
+        self.stored.compute(starting_indexes.height, exit, |vec| {
-            .compute(starting_indexes.height, exit, |vec| {
+            vec.compute_multiply(
-                vec.compute_multiply(
+                starting_indexes.height,
-                    starting_indexes.height,
+                &prices.spot.usd.height,
-                    &prices.spot.usd.height,
+                &activity.coinblocks_stored.block,
-                    &activity.coinblocks_stored.block,
+                exit,
-                    exit,
+            )?;
-                )?;
+            Ok(())
-                Ok(())
+        })?;
-            })?;
 
         // VOCDD: Value of Coin Days Destroyed = price × (CDD / circulating_supply)
         // Supply-adjusted to account for growing supply over time
         // This is a key input for Reserve Risk / HODL Bank calculation
-        self.vocdd
+        self.vocdd.compute(starting_indexes.height, exit, |vec| {
-            .compute(starting_indexes.height, exit, |vec| {
+            vec.compute_transform3(
-                vec.compute_transform3(
+                starting_indexes.height,
-                    starting_indexes.height,
+                &prices.spot.usd.height,
-                    &prices.spot.usd.height,
+                &coindays_destroyed.block,
-                    &coindays_destroyed.block,
+                circulating_supply,
-                    circulating_supply,
+                |(i, price, cdd, supply, _): (_, Dollars, StoredF64, Bitcoin, _)| {
-                    |(i, price, cdd, supply, _): (_, Dollars, StoredF64, Bitcoin, _)| {
+                    let supply_f64 = f64::from(supply);
-                        let supply_f64 = f64::from(supply);
+                    if supply_f64 == 0.0 {
-                        if supply_f64 == 0.0 {
+                        (i, StoredF64::from(0.0))
-                            (i, StoredF64::from(0.0))
+                    } else {
-                        } else {
+                        // VOCDD = price × (CDD / supply)
-                            // VOCDD = price × (CDD / supply)
+                        let vocdd = f64::from(price) * f64::from(cdd) / supply_f64;
-                            let vocdd = f64::from(price) * f64::from(cdd) / supply_f64;
+                        (i, StoredF64::from(vocdd))
-                            (i, StoredF64::from(vocdd))
+                    }
-                        }
+                },
-                    },
+                exit,
-                    exit,
+            )?;
-                )?;
+            Ok(())
-                Ok(())
+        })?;
-            })?;
 
         Ok(())
     }

crates/brk_computer/src/cointime/value/import.rs

@@ -5,7 +5,7 @@ use vecdb::Database;
 use super::Vecs;
 use crate::{
     indexes,
-    internal::{CachedWindowStarts, PerBlockCumulativeRolling},
+    internal::{PerBlockCumulativeRolling, WindowStartVec, Windows},
 };
 
 impl Vecs {
@@ -13,7 +13,7 @@ impl Vecs {
         db: &Database,
         version: Version,
         indexes: &indexes::Vecs,
-        cached_starts: &CachedWindowStarts,
+        cached_starts: &Windows<&WindowStartVec>,
     ) -> Result<Self> {
         Ok(Self {
             destroyed: PerBlockCumulativeRolling::forced_import(

crates/brk_computer/src/distribution/addr/activity.rs

@@ -7,19 +7,20 @@
 //! | `receiving` | Unique addresses that received this block |
 //! | `sending` | Unique addresses that sent this block |
 //! | `reactivated` | Addresses that were empty and now have funds |
-//! | `both` | Addresses that both sent AND received same block |
+//! | `bidirectional` | Addresses that both sent AND received in same block |
+//! | `active` | Distinct addresses involved (sent ∪ received) |
 
 use brk_cohort::ByAddrType;
 use brk_error::Result;
 use brk_traversable::Traversable;
-use brk_types::{Height, StoredU32, Version};
+use brk_types::{Height, StoredU32, StoredU64, Version};
 use derive_more::{Deref, DerefMut};
 use rayon::prelude::*;
 use vecdb::{AnyStoredVec, AnyVec, Database, Exit, Rw, StorageMode, WritableVec};
 
 use crate::{
     indexes,
-    internal::{CachedWindowStarts, PerBlockRollingAverage},
+    internal::{PerBlockRollingAverage, WindowStartVec, Windows},
 };
 
 /// Per-block activity counts - reset each block.
@@ -28,7 +29,7 @@ pub struct BlockActivityCounts {
     pub reactivated: u32,
     pub sending: u32,
     pub receiving: u32,
-    pub both: u32,
+    pub bidirectional: u32,
 }
 
 impl BlockActivityCounts {
@@ -56,7 +57,7 @@ impl AddrTypeToActivityCounts {
             total.reactivated += counts.reactivated;
             total.sending += counts.sending;
             total.receiving += counts.receiving;
-            total.both += counts.both;
+            total.bidirectional += counts.bidirectional;
         }
         total
     }
@@ -65,45 +66,61 @@ impl AddrTypeToActivityCounts {
 /// Activity count vectors for a single category (e.g., one address type or "all").
 #[derive(Traversable)]
 pub struct ActivityCountVecs<M: StorageMode = Rw> {
-    pub reactivated: PerBlockRollingAverage<StoredU32, M>,
+    pub reactivated: PerBlockRollingAverage<StoredU32, StoredU64, M>,
-    pub sending: PerBlockRollingAverage<StoredU32, M>,
+    pub sending: PerBlockRollingAverage<StoredU32, StoredU64, M>,
-    pub receiving: PerBlockRollingAverage<StoredU32, M>,
+    pub receiving: PerBlockRollingAverage<StoredU32, StoredU64, M>,
-    pub both: PerBlockRollingAverage<StoredU32, M>,
+    pub bidirectional: PerBlockRollingAverage<StoredU32, StoredU64, M>,
+    /// Distinct addresses involved in this block (sent ∪ received),
+    /// computed at push time as `sending + receiving - bidirectional`
+    /// via inclusion-exclusion. For per-type instances this is
+    /// per-type. For the `all` aggregate it's the cross-type total.
+    pub active: PerBlockRollingAverage<StoredU32, StoredU64, M>,
 }
 
 impl ActivityCountVecs {
+    /// `prefix` is prepended to each field's disk name. Use `""` for the
+    /// "all" aggregate and `"{type}_"` for per-address-type instances.
+    /// Field names are suffixed with `_addrs` so the final disk series
+    /// are e.g. `active_addrs`, `p2tr_bidirectional_addrs`.
     pub(crate) fn forced_import(
         db: &Database,
-        name: &str,
+        prefix: &str,
         version: Version,
         indexes: &indexes::Vecs,
-        cached_starts: &CachedWindowStarts,
+        cached_starts: &Windows<&WindowStartVec>,
     ) -> Result<Self> {
         Ok(Self {
             reactivated: PerBlockRollingAverage::forced_import(
                 db,
-                &format!("{name}_reactivated"),
+                &format!("{prefix}reactivated_addrs"),
                 version,
                 indexes,
                 cached_starts,
             )?,
             sending: PerBlockRollingAverage::forced_import(
                 db,
-                &format!("{name}_sending"),
+                &format!("{prefix}sending_addrs"),
                 version,
                 indexes,
                 cached_starts,
             )?,
             receiving: PerBlockRollingAverage::forced_import(
                 db,
-                &format!("{name}_receiving"),
+                &format!("{prefix}receiving_addrs"),
                 version,
                 indexes,
                 cached_starts,
             )?,
-            both: PerBlockRollingAverage::forced_import(
+            bidirectional: PerBlockRollingAverage::forced_import(
                 db,
-                &format!("{name}_both"),
+                &format!("{prefix}bidirectional_addrs"),
+                version,
+                indexes,
+                cached_starts,
+            )?,
+            active: PerBlockRollingAverage::forced_import(
+                db,
+                &format!("{prefix}active_addrs"),
                 version,
                 indexes,
                 cached_starts,
@@ -117,7 +134,8 @@ impl ActivityCountVecs {
             .len()
             .min(self.sending.block.len())
             .min(self.receiving.block.len())
-            .min(self.both.block.len())
+            .min(self.bidirectional.block.len())
+            .min(self.active.block.len())
     }
 
     pub(crate) fn par_iter_height_mut(
@@ -125,9 +143,10 @@ impl ActivityCountVecs {
     ) -> impl ParallelIterator<Item = &mut dyn AnyStoredVec> {
         [
             &mut self.reactivated.block as &mut dyn AnyStoredVec,
-            &mut self.sending.block as &mut dyn AnyStoredVec,
+            &mut self.sending.block,
-            &mut self.receiving.block as &mut dyn AnyStoredVec,
+            &mut self.receiving.block,
-            &mut self.both.block as &mut dyn AnyStoredVec,
+            &mut self.bidirectional.block,
+            &mut self.active.block,
         ]
         .into_par_iter()
     }
@@ -136,7 +155,8 @@ impl ActivityCountVecs {
         self.reactivated.block.reset()?;
         self.sending.block.reset()?;
         self.receiving.block.reset()?;
-        self.both.block.reset()?;
+        self.bidirectional.block.reset()?;
+        self.active.block.reset()?;
         Ok(())
     }
 
@@ -145,18 +165,19 @@ impl ActivityCountVecs {
         self.reactivated.block.push(counts.reactivated.into());
         self.sending.block.push(counts.sending.into());
         self.receiving.block.push(counts.receiving.into());
-        self.both.block.push(counts.both.into());
+        self.bidirectional
+            .block
+            .push(counts.bidirectional.into());
+        let active = counts.sending + counts.receiving - counts.bidirectional;
+        self.active.block.push(active.into());
     }
 
-    pub(crate) fn compute_rest(
+    pub(crate) fn compute_rest(&mut self, max_from: Height, exit: &Exit) -> Result<()> {
-        &mut self,
-        max_from: Height,
-        exit: &Exit,
-    ) -> Result<()> {
         self.reactivated.compute_rest(max_from, exit)?;
         self.sending.compute_rest(max_from, exit)?;
         self.receiving.compute_rest(max_from, exit)?;
-        self.both.compute_rest(max_from, exit)?;
+        self.bidirectional.compute_rest(max_from, exit)?;
+        self.active.compute_rest(max_from, exit)?;
         Ok(())
     }
 }
@@ -175,22 +196,21 @@ impl From<ByAddrType<ActivityCountVecs>> for AddrTypeToActivityCountVecs {
 impl AddrTypeToActivityCountVecs {
     pub(crate) fn forced_import(
         db: &Database,
-        name: &str,
         version: Version,
         indexes: &indexes::Vecs,
-        cached_starts: &CachedWindowStarts,
+        cached_starts: &Windows<&WindowStartVec>,
     ) -> Result<Self> {
-        Ok(Self::from(
+        Ok(Self::from(ByAddrType::<ActivityCountVecs>::new_with_name(
-            ByAddrType::<ActivityCountVecs>::new_with_name(|type_name| {
+            |type_name| {
                 ActivityCountVecs::forced_import(
                     db,
-                    &format!("{type_name}_{name}"),
+                    &format!("{type_name}_"),
                     version,
                     indexes,
                     cached_starts,
                 )
-            })?,
+            },
-        ))
+        )?))
     }
 
     pub(crate) fn min_stateful_len(&self) -> usize {
@@ -209,7 +229,8 @@ impl AddrTypeToActivityCountVecs {
             vecs.push(&mut type_vecs.reactivated.block);
             vecs.push(&mut type_vecs.sending.block);
             vecs.push(&mut type_vecs.receiving.block);
-            vecs.push(&mut type_vecs.both.block);
+            vecs.push(&mut type_vecs.bidirectional.block);
+            vecs.push(&mut type_vecs.active.block);
         }
         vecs.into_par_iter()
     }
@@ -221,11 +242,7 @@ impl AddrTypeToActivityCountVecs {
         Ok(())
     }
 
-    pub(crate) fn compute_rest(
+    pub(crate) fn compute_rest(&mut self, max_from: Height, exit: &Exit) -> Result<()> {
-        &mut self,
-        max_from: Height,
-        exit: &Exit,
-    ) -> Result<()> {
         for type_vecs in self.0.values_mut() {
             type_vecs.compute_rest(max_from, exit)?;
         }
@@ -251,15 +268,17 @@ pub struct AddrActivityVecs<M: StorageMode = Rw> {
 impl AddrActivityVecs {
     pub(crate) fn forced_import(
         db: &Database,
-        name: &str,
         version: Version,
         indexes: &indexes::Vecs,
-        cached_starts: &CachedWindowStarts,
+        cached_starts: &Windows<&WindowStartVec>,
     ) -> Result<Self> {
         Ok(Self {
-            all: ActivityCountVecs::forced_import(db, name, version, indexes, cached_starts)?,
+            all: ActivityCountVecs::forced_import(db, "", version, indexes, cached_starts)?,
             by_addr_type: AddrTypeToActivityCountVecs::forced_import(
-                db, name, version, indexes, cached_starts,
+                db,
+                version,
+                indexes,
+                cached_starts,
             )?,
         })
     }
@@ -284,11 +303,7 @@ impl AddrActivityVecs {
         Ok(())
     }
 
-    pub(crate) fn compute_rest(
+    pub(crate) fn compute_rest(&mut self, max_from: Height, exit: &Exit) -> Result<()> {
-        &mut self,
-        max_from: Height,
-        exit: &Exit,
-    ) -> Result<()> {
         self.all.compute_rest(max_from, exit)?;
         self.by_addr_type.compute_rest(max_from, exit)?;
         Ok(())

crates/brk_computer/src/distribution/addr/addr_count.rs

@@ -12,9 +12,7 @@ use vecdb::{
 use crate::{indexes, internal::PerBlock};
 
 #[derive(Deref, DerefMut, Traversable)]
-pub struct AddrCountVecs<M: StorageMode = Rw>(
+pub struct AddrCountVecs<M: StorageMode = Rw>(#[traversable(flatten)] pub PerBlock<StoredU64, M>);
-    #[traversable(flatten)] pub PerBlock<StoredU64, M>,
-);
 
 impl AddrCountVecs {
     pub(crate) fn forced_import(
@@ -23,9 +21,7 @@ impl AddrCountVecs {
         version: Version,
         indexes: &indexes::Vecs,
     ) -> Result<Self> {
-        Ok(Self(PerBlock::forced_import(
+        Ok(Self(PerBlock::forced_import(db, name, version, indexes)?))
-            db, name, version, indexes,
-        )?))
     }
 }
 
@@ -57,42 +53,17 @@ impl From<(&AddrTypeToAddrCountVecs, Height)> for AddrTypeToAddrCount {
                     .collect_one(prev_height)
                     .unwrap()
                     .into(),
-                p2pkh: groups
+                p2pkh: groups.p2pkh.height.collect_one(prev_height).unwrap().into(),
-                    .p2pkh
+                p2sh: groups.p2sh.height.collect_one(prev_height).unwrap().into(),
-                    .height
-                    .collect_one(prev_height)
-                    .unwrap()
-                    .into(),
-                p2sh: groups
-                    .p2sh
-                    .height
-                    .collect_one(prev_height)
-                    .unwrap()
-                    .into(),
                 p2wpkh: groups
                     .p2wpkh
                     .height
                     .collect_one(prev_height)
                     .unwrap()
                     .into(),
-                p2wsh: groups
+                p2wsh: groups.p2wsh.height.collect_one(prev_height).unwrap().into(),
-                    .p2wsh
+                p2tr: groups.p2tr.height.collect_one(prev_height).unwrap().into(),
-                    .height
+                p2a: groups.p2a.height.collect_one(prev_height).unwrap().into(),
-                    .collect_one(prev_height)
-                    .unwrap()
-                    .into(),
-                p2tr: groups
-                    .p2tr
-                    .height
-                    .collect_one(prev_height)
-                    .unwrap()
-                    .into(),
-                p2a: groups
-                    .p2a
-                    .height
-                    .collect_one(prev_height)
-                    .unwrap()
-                    .into(),
             })
         } else {
             Default::default()
@@ -177,7 +148,10 @@ impl AddrCountsVecs {
     }
 
     pub(crate) fn min_stateful_len(&self) -> usize {
-        self.all.height.len().min(self.by_addr_type.min_stateful_len())
+        self.all
+            .height
+            .len()
+            .min(self.by_addr_type.min_stateful_len())
     }
 
     pub(crate) fn par_iter_height_mut(
@@ -199,11 +173,7 @@ impl AddrCountsVecs {
         self.by_addr_type.push_height(addr_counts);
     }
 
-    pub(crate) fn compute_rest(
+    pub(crate) fn compute_rest(&mut self, starting_indexes: &Indexes, exit: &Exit) -> Result<()> {
-        &mut self,
-        starting_indexes: &Indexes,
-        exit: &Exit,
-    ) -> Result<()> {
         let sources = self.by_addr_type.by_height();
         self.all
             .height

crates/brk_computer/src/distribution/addr/data.rs

@@ -1,8 +1,6 @@
 use brk_error::Result;
 use brk_traversable::Traversable;
-use brk_types::{
+use brk_types::{EmptyAddrData, EmptyAddrIndex, FundedAddrData, FundedAddrIndex, Height};
-    EmptyAddrData, EmptyAddrIndex, FundedAddrData, FundedAddrIndex, Height,
-};
 use rayon::prelude::*;
 use vecdb::{AnyStoredVec, BytesVec, Rw, Stamp, StorageMode, WritableVec};
 

crates/brk_computer/src/distribution/addr/delta.rs

@@ -1,28 +1,24 @@
-use brk_cohort::ByAddrType;
 use brk_traversable::Traversable;
 use brk_types::{BasisPointsSigned32, StoredI64, StoredU64, Version};
+use derive_more::{Deref, DerefMut};
 
 use crate::{
     indexes,
-    internal::{CachedWindowStarts, LazyRollingDeltasFromHeight},
+    internal::{LazyRollingDeltasFromHeight, WindowStartVec, Windows, WithAddrTypes},
 };
 
 use super::AddrCountsVecs;
 
 type AddrDelta = LazyRollingDeltasFromHeight<StoredU64, StoredI64, BasisPointsSigned32>;
 
-#[derive(Clone, Traversable)]
+#[derive(Clone, Deref, DerefMut, Traversable)]
-pub struct DeltaVecs {
+pub struct DeltaVecs(#[traversable(flatten)] pub WithAddrTypes<AddrDelta>);
-    pub all: AddrDelta,
-    #[traversable(flatten)]
-    pub by_addr_type: ByAddrType<AddrDelta>,
-}
 
 impl DeltaVecs {
     pub(crate) fn new(
         version: Version,
         addr_count: &AddrCountsVecs,
-        cached_starts: &CachedWindowStarts,
+        cached_starts: &Windows<&WindowStartVec>,
         indexes: &indexes::Vecs,
     ) -> Self {
         let version = version + Version::TWO;
@@ -45,9 +41,6 @@ impl DeltaVecs {
             )
         });
 
-        Self {
+        Self(WithAddrTypes { all, by_addr_type })
-            all,
-            by_addr_type,
-        }
     }
 }

crates/brk_computer/src/distribution/addr/exposed/count/mod.rs

@@ -0,0 +1,74 @@
+//! Exposed address count tracking — running counters of how many addresses
+//! are currently in (or have ever been in) the exposed set, per address type
+//! plus an aggregated `all`. See the parent [`super`] module for the
+//! definition of "exposed" and how it varies by address type.
+
+mod state;
+mod vecs;
+
+pub use state::AddrTypeToExposedAddrCount;
+pub use vecs::ExposedAddrCountAllVecs;
+
+use brk_error::Result;
+use brk_traversable::Traversable;
+use brk_types::{Indexes, Version};
+use rayon::prelude::*;
+use vecdb::{AnyStoredVec, Database, Exit, Rw, StorageMode};
+
+use crate::indexes;
+
+/// Exposed address counts: funded (currently at-risk) and total (ever at-risk).
+#[derive(Traversable)]
+pub struct ExposedAddrCountsVecs<M: StorageMode = Rw> {
+    pub funded: ExposedAddrCountAllVecs<M>,
+    pub total: ExposedAddrCountAllVecs<M>,
+}
+
+impl ExposedAddrCountsVecs {
+    pub(crate) fn forced_import(
+        db: &Database,
+        version: Version,
+        indexes: &indexes::Vecs,
+    ) -> Result<Self> {
+        Ok(Self {
+            funded: ExposedAddrCountAllVecs::forced_import(
+                db,
+                "exposed_addr_count",
+                version,
+                indexes,
+            )?,
+            total: ExposedAddrCountAllVecs::forced_import(
+                db,
+                "total_exposed_addr_count",
+                version,
+                indexes,
+            )?,
+        })
+    }
+
+    pub(crate) fn min_stateful_len(&self) -> usize {
+        self.funded
+            .min_stateful_len()
+            .min(self.total.min_stateful_len())
+    }
+
+    pub(crate) fn par_iter_height_mut(
+        &mut self,
+    ) -> impl ParallelIterator<Item = &mut dyn AnyStoredVec> {
+        self.funded
+            .par_iter_height_mut()
+            .chain(self.total.par_iter_height_mut())
+    }
+
+    pub(crate) fn reset_height(&mut self) -> Result<()> {
+        self.funded.reset_height()?;
+        self.total.reset_height()?;
+        Ok(())
+    }
+
+    pub(crate) fn compute_rest(&mut self, starting_indexes: &Indexes, exit: &Exit) -> Result<()> {
+        self.funded.compute_rest(starting_indexes, exit)?;
+        self.total.compute_rest(starting_indexes, exit)?;
+        Ok(())
+    }
+}

crates/brk_computer/src/distribution/addr/exposed/count/state.rs

@@ -0,0 +1,42 @@
+use brk_cohort::ByAddrType;
+use brk_types::{Height, StoredU64};
+use derive_more::{Deref, DerefMut};
+use vecdb::ReadableVec;
+
+use crate::internal::PerBlock;
+
+use super::vecs::ExposedAddrCountAllVecs;
+
+/// Runtime counter for exposed address counts per address type.
+#[derive(Debug, Default, Deref, DerefMut)]
+pub struct AddrTypeToExposedAddrCount(ByAddrType<u64>);
+
+impl AddrTypeToExposedAddrCount {
+    #[inline]
+    pub(crate) fn sum(&self) -> u64 {
+        self.0.values().sum()
+    }
+}
+
+impl From<(&ExposedAddrCountAllVecs, Height)> for AddrTypeToExposedAddrCount {
+    #[inline]
+    fn from((vecs, starting_height): (&ExposedAddrCountAllVecs, Height)) -> Self {
+        if let Some(prev_height) = starting_height.decremented() {
+            let read = |v: &PerBlock<StoredU64>| -> u64 {
+                v.height.collect_one(prev_height).unwrap().into()
+            };
+            Self(ByAddrType {
+                p2pk65: read(&vecs.by_addr_type.p2pk65),
+                p2pk33: read(&vecs.by_addr_type.p2pk33),
+                p2pkh: read(&vecs.by_addr_type.p2pkh),
+                p2sh: read(&vecs.by_addr_type.p2sh),
+                p2wpkh: read(&vecs.by_addr_type.p2wpkh),
+                p2wsh: read(&vecs.by_addr_type.p2wsh),
+                p2tr: read(&vecs.by_addr_type.p2tr),
+                p2a: read(&vecs.by_addr_type.p2a),
+            })
+        } else {
+            Default::default()
+        }
+    }
+}

crates/brk_computer/src/distribution/addr/exposed/count/vecs.rs

@@ -0,0 +1,29 @@
+use brk_error::Result;
+use brk_traversable::Traversable;
+use brk_types::{StoredU64, Version};
+use derive_more::{Deref, DerefMut};
+use vecdb::{Database, Rw, StorageMode};
+
+use crate::{
+    indexes,
+    internal::{PerBlock, WithAddrTypes},
+};
+
+/// Exposed address count (`all` + per-type) for a single variant (funded or total).
+#[derive(Deref, DerefMut, Traversable)]
+pub struct ExposedAddrCountAllVecs<M: StorageMode = Rw>(
+    #[traversable(flatten)] pub WithAddrTypes<PerBlock<StoredU64, M>>,
+);
+
+impl ExposedAddrCountAllVecs {
+    pub(crate) fn forced_import(
+        db: &Database,
+        name: &str,
+        version: Version,
+        indexes: &indexes::Vecs,
+    ) -> Result<Self> {
+        Ok(Self(WithAddrTypes::<PerBlock<StoredU64>>::forced_import(
+            db, name, version, indexes,
+        )?))
+    }
+}

crates/brk_computer/src/distribution/addr/exposed/mod.rs

@@ -0,0 +1,134 @@
+//! Exposed address tracking (quantum / pubkey-exposure sense).
+//!
+//! An address is "exposed" once its public key is in the blockchain. Once
+//! exposed, any funds at that address are at cryptographic risk (e.g. from
+//! a quantum attacker capable of recovering the private key from the pubkey).
+//!
+//! When the pubkey gets exposed depends on the address type:
+//!
+//! - **P2PK33, P2PK65, P2TR**: the pubkey (or P2TR's tweaked output key) is
+//!   directly in the locking script of the funding output. These addresses are
+//!   exposed the moment they receive any funds.
+//! - **P2PKH, P2SH, P2WPKH, P2WSH**: the locking script contains a hash of
+//!   the pubkey/script. The pubkey is only revealed when spending. Note that
+//!   even the spending tx itself exposes the pubkey while the address still
+//!   holds funds — during the mempool window between broadcast and confirmation,
+//!   the pubkey is visible while the UTXO being spent is still unspent on-chain.
+//!   So every spent address of these types has had at least one moment with
+//!   funds at quantum risk.
+//! - **P2A**: anyone-can-spend, no pubkey at all. Excluded from both counters.
+//!
+//! Formally, with `is_funding_exposed` = `output_type.pubkey_exposed_at_funding()`:
+//! - `funded` (count): `(utxo_count > 0) AND (is_funding_exposed OR spent_txo_count >= 1)`
+//! - `total` (count): `(is_funding_exposed AND ever received) OR spent_txo_count >= 1`
+//! - `supply` (sats): sum of balances of addresses currently in the funded set
+//!
+//! For P2PK/P2TR types this means `total ≡ total_addr_count` and
+//! `funded ≡ funded_addr_count` (every address of those types is exposed by
+//! virtue of existing). For P2PKH/P2SH/P2WPKH/P2WSH it's the strict subset of
+//! addresses that have been spent from. The aggregate `all` exposed counter
+//! sums these, giving "Bitcoin addresses currently with funds at quantum risk".
+//!
+//! All metrics are tracked as running counters and require no extra fields
+//! on the address data — they're maintained via delta detection in
+//! `process_received` and `process_sent`.
+
+mod count;
+mod supply;
+
+pub use count::{AddrTypeToExposedAddrCount, ExposedAddrCountsVecs};
+pub use supply::{AddrTypeToExposedSupply, ExposedAddrSupplyVecs, ExposedSupplyShareVecs};
+
+use brk_cohort::ByAddrType;
+use brk_error::Result;
+use brk_traversable::Traversable;
+use brk_types::{Height, Indexes, Sats, Version};
+use rayon::prelude::*;
+use vecdb::{AnyStoredVec, Database, Exit, ReadableVec, Rw, StorageMode};
+
+use crate::{indexes, internal::RatioSatsBp16, prices};
+
+/// Top-level container for all exposed address tracking: counts (funded +
+/// total), the funded supply, and share of supply.
+#[derive(Traversable)]
+pub struct ExposedAddrVecs<M: StorageMode = Rw> {
+    pub count: ExposedAddrCountsVecs<M>,
+    pub supply: ExposedAddrSupplyVecs<M>,
+    #[traversable(wrap = "supply", rename = "share")]
+    pub supply_share: ExposedSupplyShareVecs<M>,
+}
+
+impl ExposedAddrVecs {
+    pub(crate) fn forced_import(
+        db: &Database,
+        version: Version,
+        indexes: &indexes::Vecs,
+    ) -> Result<Self> {
+        Ok(Self {
+            count: ExposedAddrCountsVecs::forced_import(db, version, indexes)?,
+            supply: ExposedAddrSupplyVecs::forced_import(db, version, indexes)?,
+            supply_share: ExposedSupplyShareVecs::forced_import(db, version, indexes)?,
+        })
+    }
+
+    pub(crate) fn min_stateful_len(&self) -> usize {
+        self.count
+            .min_stateful_len()
+            .min(self.supply.min_stateful_len())
+    }
+
+    pub(crate) fn par_iter_height_mut(
+        &mut self,
+    ) -> impl ParallelIterator<Item = &mut dyn AnyStoredVec> {
+        self.count
+            .par_iter_height_mut()
+            .chain(self.supply.par_iter_height_mut())
+    }
+
+    pub(crate) fn reset_height(&mut self) -> Result<()> {
+        self.count.reset_height()?;
+        self.supply.reset_height()?;
+        self.supply_share.reset_height()?;
+        Ok(())
+    }
+
+    pub(crate) fn compute_rest(
+        &mut self,
+        starting_indexes: &Indexes,
+        prices: &prices::Vecs,
+        all_supply_sats: &impl ReadableVec<Height, Sats>,
+        type_supply_sats: &ByAddrType<&impl ReadableVec<Height, Sats>>,
+        exit: &Exit,
+    ) -> Result<()> {
+        self.count.compute_rest(starting_indexes, exit)?;
+        self.supply
+            .compute_rest(starting_indexes.height, prices, exit)?;
+
+        let max_from = starting_indexes.height;
+
+        self.supply_share
+            .all
+            .compute_binary::<Sats, Sats, RatioSatsBp16>(
+                max_from,
+                &self.supply.all.sats.height,
+                all_supply_sats,
+                exit,
+            )?;
+
+        for ((_, share), ((_, exposed), (_, denom))) in self
+            .supply_share
+            .by_addr_type
+            .iter_mut()
+            .zip(self.supply.by_addr_type.iter().zip(type_supply_sats.iter()))
+        {
+            share.compute_binary::<Sats, Sats, RatioSatsBp16>(
+                max_from,
+                &exposed.sats.height,
+                *denom,
+                exit,
+            )?;
+        }
+
+        Ok(())
+    }
+}

crates/brk_computer/src/distribution/addr/exposed/supply/mod.rs

@@ -0,0 +1,12 @@
+//! Exposed address supply (sats) tracking — running sum of balances held by
+//! addresses currently in the funded exposed set, per address type plus an
+//! aggregated `all`. See the parent [`super`] module for the definition of
+//! "exposed" and how it varies by address type.
+
+mod share;
+mod state;
+mod vecs;
+
+pub use share::ExposedSupplyShareVecs;
+pub use state::AddrTypeToExposedSupply;
+pub use vecs::ExposedAddrSupplyVecs;

crates/brk_computer/src/distribution/addr/exposed/supply/share.rs

@@ -0,0 +1,36 @@
+use brk_error::Result;
+use brk_traversable::Traversable;
+use brk_types::{BasisPoints16, Version};
+use derive_more::{Deref, DerefMut};
+use vecdb::{Database, Rw, StorageMode};
+
+use crate::{
+    indexes,
+    internal::{PercentPerBlock, WithAddrTypes},
+};
+
+/// Share of exposed supply relative to total supply.
+///
+/// - `all`: exposed_supply / circulating_supply
+/// - Per-type: type's exposed_supply / type's total supply
+#[derive(Deref, DerefMut, Traversable)]
+pub struct ExposedSupplyShareVecs<M: StorageMode = Rw>(
+    #[traversable(flatten)] pub WithAddrTypes<PercentPerBlock<BasisPoints16, M>>,
+);
+
+impl ExposedSupplyShareVecs {
+    pub(crate) fn forced_import(
+        db: &Database,
+        version: Version,
+        indexes: &indexes::Vecs,
+    ) -> Result<Self> {
+        Ok(Self(
+            WithAddrTypes::<PercentPerBlock<BasisPoints16>>::forced_import(
+                db,
+                "exposed_supply_share",
+                version,
+                indexes,
+            )?,
+        ))
+    }
+}

crates/brk_computer/src/distribution/addr/exposed/supply/state.rs

@@ -0,0 +1,42 @@
+use brk_cohort::ByAddrType;
+use brk_types::{Height, Sats};
+use derive_more::{Deref, DerefMut};
+use vecdb::ReadableVec;
+
+use crate::internal::ValuePerBlock;
+
+use super::vecs::ExposedAddrSupplyVecs;
+
+/// Runtime running counter for the total balance (sats) held by funded
+/// exposed addresses, per address type.
+#[derive(Debug, Default, Deref, DerefMut)]
+pub struct AddrTypeToExposedSupply(ByAddrType<Sats>);
+
+impl AddrTypeToExposedSupply {
+    #[inline]
+    pub(crate) fn sum(&self) -> Sats {
+        self.0.values().copied().sum()
+    }
+}
+
+impl From<(&ExposedAddrSupplyVecs, Height)> for AddrTypeToExposedSupply {
+    #[inline]
+    fn from((vecs, starting_height): (&ExposedAddrSupplyVecs, Height)) -> Self {
+        if let Some(prev_height) = starting_height.decremented() {
+            let read =
+                |v: &ValuePerBlock| -> Sats { v.sats.height.collect_one(prev_height).unwrap() };
+            Self(ByAddrType {
+                p2pk65: read(&vecs.by_addr_type.p2pk65),
+                p2pk33: read(&vecs.by_addr_type.p2pk33),
+                p2pkh: read(&vecs.by_addr_type.p2pkh),
+                p2sh: read(&vecs.by_addr_type.p2sh),
+                p2wpkh: read(&vecs.by_addr_type.p2wpkh),
+                p2wsh: read(&vecs.by_addr_type.p2wsh),
+                p2tr: read(&vecs.by_addr_type.p2tr),
+                p2a: read(&vecs.by_addr_type.p2a),
+            })
+        } else {
+            Default::default()
+        }
+    }
+}

crates/brk_computer/src/distribution/addr/exposed/supply/vecs.rs

@@ -0,0 +1,34 @@
+use brk_error::Result;
+use brk_traversable::Traversable;
+use brk_types::Version;
+use derive_more::{Deref, DerefMut};
+use vecdb::{Database, Rw, StorageMode};
+
+use crate::{
+    indexes,
+    internal::{ValuePerBlock, WithAddrTypes},
+};
+
+/// Exposed address supply (sats/btc/cents/usd) — `all` + per-address-type.
+/// Tracks the total balance held by addresses currently in the funded
+/// exposed set. Sats are pushed stateful per block; cents/usd are derived
+/// post-hoc from sats × spot price.
+#[derive(Deref, DerefMut, Traversable)]
+pub struct ExposedAddrSupplyVecs<M: StorageMode = Rw>(
+    #[traversable(flatten)] pub WithAddrTypes<ValuePerBlock<M>>,
+);
+
+impl ExposedAddrSupplyVecs {
+    pub(crate) fn forced_import(
+        db: &Database,
+        version: Version,
+        indexes: &indexes::Vecs,
+    ) -> Result<Self> {
+        Ok(Self(WithAddrTypes::<ValuePerBlock>::forced_import(
+            db,
+            "exposed_supply",
+            version,
+            indexes,
+        )?))
+    }
+}

crates/brk_computer/src/distribution/addr/indexes/any.rs

@@ -5,8 +5,8 @@ use brk_error::{Error, Result};
 use brk_traversable::Traversable;
 use brk_types::{
     AnyAddrIndex, Height, OutputType, P2AAddrIndex, P2PK33AddrIndex, P2PK65AddrIndex,
-    P2PKHAddrIndex, P2SHAddrIndex, P2TRAddrIndex, P2WPKHAddrIndex, P2WSHAddrIndex,
+    P2PKHAddrIndex, P2SHAddrIndex, P2TRAddrIndex, P2WPKHAddrIndex, P2WSHAddrIndex, TypeIndex,
-    TypeIndex, Version,
+    Version,
 };
 use rayon::prelude::*;
 use rustc_hash::FxHashMap;

crates/brk_computer/src/distribution/addr/mod.rs

@@ -2,8 +2,10 @@ mod activity;
 mod addr_count;
 mod data;
 mod delta;
+mod exposed;
 mod indexes;
 mod new_addr_count;
+mod reused;
 mod total_addr_count;
 mod type_map;
 
@@ -11,7 +13,9 @@ pub use activity::{AddrActivityVecs, AddrTypeToActivityCounts};
 pub use addr_count::{AddrCountsVecs, AddrTypeToAddrCount};
 pub use data::AddrsDataVecs;
 pub use delta::DeltaVecs;
+pub use exposed::{AddrTypeToExposedAddrCount, AddrTypeToExposedSupply, ExposedAddrVecs,};
 pub use indexes::AnyAddrIndexesVecs;
 pub use new_addr_count::NewAddrCountVecs;
+pub use reused::{AddrTypeToReusedAddrCount, AddrTypeToReusedAddrEventCount, ReusedAddrVecs};
 pub use total_addr_count::TotalAddrCountVecs;
 pub use type_map::{AddrTypeToTypeIndexMap, AddrTypeToVec, HeightToAddrTypeToVec};

crates/brk_computer/src/distribution/addr/new_addr_count.rs

@@ -1,53 +1,35 @@
-use brk_cohort::ByAddrType;
 use brk_error::Result;
 use brk_traversable::Traversable;
 use brk_types::{Height, StoredU64, Version};
+use derive_more::{Deref, DerefMut};
 use vecdb::{Database, Exit, Rw, StorageMode};
 
 use crate::{
     indexes,
-    internal::{CachedWindowStarts, PerBlockCumulativeRolling},
+    internal::{PerBlockCumulativeRolling, WindowStartVec, Windows, WithAddrTypes},
 };
 
 use super::TotalAddrCountVecs;
 
-/// New address count per block (global + per-type)
+/// New address count per block (global + per-type).
-#[derive(Traversable)]
+#[derive(Deref, DerefMut, Traversable)]
-pub struct NewAddrCountVecs<M: StorageMode = Rw> {
+pub struct NewAddrCountVecs<M: StorageMode = Rw>(
-    pub all: PerBlockCumulativeRolling<StoredU64, StoredU64, M>,
     #[traversable(flatten)]
-    pub by_addr_type: ByAddrType<PerBlockCumulativeRolling<StoredU64, StoredU64, M>>,
+    pub WithAddrTypes<PerBlockCumulativeRolling<StoredU64, StoredU64, M>>,
-}
+);
 
 impl NewAddrCountVecs {
     pub(crate) fn forced_import(
         db: &Database,
         version: Version,
         indexes: &indexes::Vecs,
-        cached_starts: &CachedWindowStarts,
+        cached_starts: &Windows<&WindowStartVec>,
     ) -> Result<Self> {
-        let all = PerBlockCumulativeRolling::forced_import(
+        Ok(Self(WithAddrTypes::<
-            db,
+            PerBlockCumulativeRolling<StoredU64, StoredU64>,
-            "new_addr_count",
+        >::forced_import(
-            version,
+            db, "new_addr_count", version, indexes, cached_starts
-            indexes,
+        )?))
-            cached_starts,
-        )?;
-
-        let by_addr_type = ByAddrType::new_with_name(|name| {
-            PerBlockCumulativeRolling::forced_import(
-                db,
-                &format!("{name}_new_addr_count"),
-                version,
-                indexes,
-                cached_starts,
-            )
-        })?;
-
-        Ok(Self {
-            all,
-            by_addr_type,
-        })
     }
 
     pub(crate) fn compute(
@@ -56,11 +38,12 @@ impl NewAddrCountVecs {
         total_addr_count: &TotalAddrCountVecs,
         exit: &Exit,
     ) -> Result<()> {
-        self.all.compute(max_from, exit, |height_vec| {
+        self.0.all.compute(max_from, exit, |height_vec| {
             Ok(height_vec.compute_change(max_from, &total_addr_count.all.height, 1, exit)?)
         })?;
 
         for ((_, new), (_, total)) in self
+            .0
             .by_addr_type
             .iter_mut()
             .zip(total_addr_count.by_addr_type.iter())

crates/brk_computer/src/distribution/addr/reused/count/mod.rs

@@ -0,0 +1,78 @@
+//! Reused address count tracking — running counters of how many addresses
+//! are currently in (or have ever been in) the reused set, per address type
+//! plus an aggregated `all`. See the parent [`super`] module for the
+//! definition of "reused".
+//!
+//! Two counters are exposed:
+//! - `funded`: addresses currently funded AND with `funded_txo_count > 1`
+//! - `total`: addresses that have ever satisfied `funded_txo_count > 1` (monotonic)
+
+mod state;
+mod vecs;
+
+pub use state::AddrTypeToReusedAddrCount;
+pub use vecs::ReusedAddrCountAllVecs;
+
+use brk_error::Result;
+use brk_traversable::Traversable;
+use brk_types::{Indexes, Version};
+use rayon::prelude::*;
+use vecdb::{AnyStoredVec, Database, Exit, Rw, StorageMode};
+
+use crate::indexes;
+
+/// Reused address counts: funded (currently with balance) and total (ever reused).
+#[derive(Traversable)]
+pub struct ReusedAddrCountsVecs<M: StorageMode = Rw> {
+    pub funded: ReusedAddrCountAllVecs<M>,
+    pub total: ReusedAddrCountAllVecs<M>,
+}
+
+impl ReusedAddrCountsVecs {
+    pub(crate) fn forced_import(
+        db: &Database,
+        version: Version,
+        indexes: &indexes::Vecs,
+    ) -> Result<Self> {
+        Ok(Self {
+            funded: ReusedAddrCountAllVecs::forced_import(
+                db,
+                "reused_addr_count",
+                version,
+                indexes,
+            )?,
+            total: ReusedAddrCountAllVecs::forced_import(
+                db,
+                "total_reused_addr_count",
+                version,
+                indexes,
+            )?,
+        })
+    }
+
+    pub(crate) fn min_stateful_len(&self) -> usize {
+        self.funded
+            .min_stateful_len()
+            .min(self.total.min_stateful_len())
+    }
+
+    pub(crate) fn par_iter_height_mut(
+        &mut self,
+    ) -> impl ParallelIterator<Item = &mut dyn AnyStoredVec> {
+        self.funded
+            .par_iter_height_mut()
+            .chain(self.total.par_iter_height_mut())
+    }
+
+    pub(crate) fn reset_height(&mut self) -> Result<()> {
+        self.funded.reset_height()?;
+        self.total.reset_height()?;
+        Ok(())
+    }
+
+    pub(crate) fn compute_rest(&mut self, starting_indexes: &Indexes, exit: &Exit) -> Result<()> {
+        self.funded.compute_rest(starting_indexes, exit)?;
+        self.total.compute_rest(starting_indexes, exit)?;
+        Ok(())
+    }
+}

crates/brk_computer/src/distribution/addr/reused/count/state.rs

@@ -0,0 +1,42 @@
+use brk_cohort::ByAddrType;
+use brk_types::{Height, StoredU64};
+use derive_more::{Deref, DerefMut};
+use vecdb::ReadableVec;
+
+use crate::internal::PerBlock;
+
+use super::vecs::ReusedAddrCountAllVecs;
+
+/// Runtime counter for reused address counts per address type.
+#[derive(Debug, Default, Deref, DerefMut)]
+pub struct AddrTypeToReusedAddrCount(ByAddrType<u64>);
+
+impl AddrTypeToReusedAddrCount {
+    #[inline]
+    pub(crate) fn sum(&self) -> u64 {
+        self.0.values().sum()
+    }
+}
+
+impl From<(&ReusedAddrCountAllVecs, Height)> for AddrTypeToReusedAddrCount {
+    #[inline]
+    fn from((vecs, starting_height): (&ReusedAddrCountAllVecs, Height)) -> Self {
+        if let Some(prev_height) = starting_height.decremented() {
+            let read = |v: &PerBlock<StoredU64>| -> u64 {
+                v.height.collect_one(prev_height).unwrap().into()
+            };
+            Self(ByAddrType {
+                p2pk65: read(&vecs.by_addr_type.p2pk65),
+                p2pk33: read(&vecs.by_addr_type.p2pk33),
+                p2pkh: read(&vecs.by_addr_type.p2pkh),
+                p2sh: read(&vecs.by_addr_type.p2sh),
+                p2wpkh: read(&vecs.by_addr_type.p2wpkh),
+                p2wsh: read(&vecs.by_addr_type.p2wsh),
+                p2tr: read(&vecs.by_addr_type.p2tr),
+                p2a: read(&vecs.by_addr_type.p2a),
+            })
+        } else {
+            Default::default()
+        }
+    }
+}

crates/brk_computer/src/distribution/addr/reused/count/vecs.rs

@@ -0,0 +1,29 @@
+use brk_error::Result;
+use brk_traversable::Traversable;
+use brk_types::{StoredU64, Version};
+use derive_more::{Deref, DerefMut};
+use vecdb::{Database, Rw, StorageMode};
+
+use crate::{
+    indexes,
+    internal::{PerBlock, WithAddrTypes},
+};
+
+/// Reused address count (`all` + per-type) for a single variant (funded or total).
+#[derive(Deref, DerefMut, Traversable)]
+pub struct ReusedAddrCountAllVecs<M: StorageMode = Rw>(
+    #[traversable(flatten)] pub WithAddrTypes<PerBlock<StoredU64, M>>,
+);
+
+impl ReusedAddrCountAllVecs {
+    pub(crate) fn forced_import(
+        db: &Database,
+        name: &str,
+        version: Version,
+        indexes: &indexes::Vecs,
+    ) -> Result<Self> {
+        Ok(Self(WithAddrTypes::<PerBlock<StoredU64>>::forced_import(
+            db, name, version, indexes,
+        )?))
+    }
+}

crates/brk_computer/src/distribution/addr/reused/events/mod.rs

@@ -0,0 +1,11 @@
+//! Per-block reused-address event tracking. Holds both the output-side
+//! ("an output landed on a previously-used address") and input-side
+//! ("an input spent from an address in the reused set") event counters.
+//! See [`vecs::ReusedAddrEventsVecs`] for the full description of each
+//! metric.
+
+mod state;
+mod vecs;
+
+pub use state::AddrTypeToReusedAddrEventCount;
+pub use vecs::ReusedAddrEventsVecs;

crates/brk_computer/src/distribution/addr/reused/events/state.rs

@@ -0,0 +1,28 @@
+use brk_cohort::ByAddrType;
+use derive_more::{Deref, DerefMut};
+
+/// Per-block running counter of reused-address events, per address type.
+/// Shared runtime container for both output-side events
+/// (`output_to_reused_addr_count`, outputs landing on addresses that
+/// had already received ≥ 1 prior output) and input-side events
+/// (`input_from_reused_addr_count`, inputs spending from addresses
+/// with lifetime `funded_txo_count > 1`). Reset at the start of each
+/// block (no disk recovery needed since per-block flow is
+/// reconstructed deterministically from `process_received` /
+/// `process_sent`).
+#[derive(Debug, Default, Deref, DerefMut)]
+pub struct AddrTypeToReusedAddrEventCount(ByAddrType<u64>);
+
+impl AddrTypeToReusedAddrEventCount {
+    #[inline]
+    pub(crate) fn sum(&self) -> u64 {
+        self.0.values().sum()
+    }
+
+    #[inline]
+    pub(crate) fn reset(&mut self) {
+        for v in self.0.values_mut() {
+            *v = 0;
+        }
+    }
+}

crates/brk_computer/src/distribution/addr/reused/events/vecs.rs

@@ -0,0 +1,261 @@
+use brk_cohort::ByAddrType;
+use brk_error::Result;
+use brk_traversable::Traversable;
+use brk_types::{BasisPoints16, Indexes, OutputType, StoredF32, StoredU32, StoredU64, Version};
+use rayon::prelude::*;
+use vecdb::{AnyStoredVec, AnyVec, Database, Exit, Rw, StorageMode, WritableVec};
+
+use crate::{
+    indexes, inputs,
+    internal::{
+        PerBlockCumulativeRolling, PerBlockRollingAverage, PercentCumulativeRolling,
+        WindowStartVec, Windows, WithAddrTypes,
+    },
+    outputs,
+};
+
+use super::state::AddrTypeToReusedAddrEventCount;
+
+/// Per-block reused-address event metrics. Holds three families of
+/// signals: output-level (use), input-level (spend), and address-level
+/// (active in block).
+///
+/// `output_to_reused_addr_count`: every output landing on an address that had
+/// already received at least one prior output anywhere in its lifetime,
+/// i.e. an output-level reuse event. Outputs are not deduplicated per
+/// address within a block: an address receiving N outputs in one block
+/// that had `before` lifetime outputs contributes
+/// `max(0, N - max(0, 1 - before))` events. Only the very first output
+/// an address ever sees is excluded. Every subsequent output counts,
+/// matching the standard "% of outputs to previously-used addresses"
+/// reuse ratio reported by external sources. `output_to_reused_addr_share`
+/// uses `outputs::ByTypeVecs::output_count` (all 12 output types) as
+/// denominator. `spendable_output_to_reused_addr_share` uses the
+/// op_return-excluded 11-type aggregate (`spendable_output_count`).
+///
+/// `input_from_reused_addr_count`: every input spending from an address
+/// whose lifetime `funded_txo_count > 1` at the time of the spend (i.e.
+/// the address is in the same reused set tracked by
+/// `reused_addr_count`). Every input is checked independently. If a
+/// single address has multiple inputs in one block each one counts.
+/// This is a *stable-predicate* signal about the sending address, not
+/// an output-level repeat event: the first spend from a reused address
+/// counts just as much as the tenth. Denominator
+/// (`input_from_reused_addr_share`): `inputs::ByTypeVecs::input_count` (11
+/// spendable types, where `p2ms`, `unknown`, `empty` count as true
+/// negatives).
+///
+/// `active_reused_addr_count` / `active_reused_addr_share`: block-level
+/// *address* signals (single aggregate, not per-type).
+/// `active_reused_addr_count` is the count of distinct addresses
+/// involved in this block (sent ∪ received) that satisfy `is_reused()`
+/// after the block's events, populated inline in `process_received`
+/// (each receiver, post-receive) and in `process_sent` (each
+/// first-encounter sender, deduped against `received_addrs` so
+/// addresses that did both aren't double-counted).
+/// `active_reused_addr_share` is the per-block ratio
+/// `reused / active * 100` as a percentage in `[0, 100]` (or `0.0` for
+/// empty blocks). The denominator (distinct active addrs per block)
+/// lives on `ActivityCountVecs::active` (`addrs.activity.all.active`),
+/// derived from `sending + receiving - bidirectional`. Both fields
+/// use `PerBlockRollingAverage` so their lazy 24h/1w/1m/1y series are
+/// rolling *averages* of the per-block values. Sums and cumulatives of
+/// distinct-address counts would be misleading because the same
+/// address can appear in multiple blocks.
+#[derive(Traversable)]
+pub struct ReusedAddrEventsVecs<M: StorageMode = Rw> {
+    pub output_to_reused_addr_count:
+        WithAddrTypes<PerBlockCumulativeRolling<StoredU64, StoredU64, M>>,
+    pub output_to_reused_addr_share: WithAddrTypes<PercentCumulativeRolling<BasisPoints16, M>>,
+    pub spendable_output_to_reused_addr_share: PercentCumulativeRolling<BasisPoints16, M>,
+    pub input_from_reused_addr_count:
+        WithAddrTypes<PerBlockCumulativeRolling<StoredU64, StoredU64, M>>,
+    pub input_from_reused_addr_share: WithAddrTypes<PercentCumulativeRolling<BasisPoints16, M>>,
+    pub active_reused_addr_count: PerBlockRollingAverage<StoredU32, StoredU64, M>,
+    pub active_reused_addr_share: PerBlockRollingAverage<StoredF32, StoredF32, M>,
+}
+
+impl ReusedAddrEventsVecs {
+    pub(crate) fn forced_import(
+        db: &Database,
+        version: Version,
+        indexes: &indexes::Vecs,
+        cached_starts: &Windows<&WindowStartVec>,
+    ) -> Result<Self> {
+        let import_count = |name: &str| {
+            WithAddrTypes::<PerBlockCumulativeRolling<StoredU64, StoredU64>>::forced_import(
+                db,
+                name,
+                version,
+                indexes,
+                cached_starts,
+            )
+        };
+        let import_percent = |name: &str| -> Result<WithAddrTypes<
+            PercentCumulativeRolling<BasisPoints16>,
+        >> {
+            Ok(WithAddrTypes {
+                all: PercentCumulativeRolling::forced_import(db, name, version, indexes)?,
+                by_addr_type: ByAddrType::new_with_name(|type_name| {
+                    PercentCumulativeRolling::forced_import(
+                        db,
+                        &format!("{type_name}_{name}"),
+                        version,
+                        indexes,
+                    )
+                })?,
+            })
+        };
+
+        let output_to_reused_addr_count = import_count("output_to_reused_addr_count")?;
+        let output_to_reused_addr_share = import_percent("output_to_reused_addr_share")?;
+        let spendable_output_to_reused_addr_share = PercentCumulativeRolling::forced_import(
+            db,
+            "spendable_output_to_reused_addr_share",
+            version,
+            indexes,
+        )?;
+        let input_from_reused_addr_count = import_count("input_from_reused_addr_count")?;
+        let input_from_reused_addr_share = import_percent("input_from_reused_addr_share")?;
+
+        let active_reused_addr_count = PerBlockRollingAverage::forced_import(
+            db,
+            "active_reused_addr_count",
+            version,
+            indexes,
+            cached_starts,
+        )?;
+        let active_reused_addr_share = PerBlockRollingAverage::forced_import(
+            db,
+            "active_reused_addr_share",
+            version,
+            indexes,
+            cached_starts,
+        )?;
+
+        Ok(Self {
+            output_to_reused_addr_count,
+            output_to_reused_addr_share,
+            spendable_output_to_reused_addr_share,
+            input_from_reused_addr_count,
+            input_from_reused_addr_share,
+            active_reused_addr_count,
+            active_reused_addr_share,
+        })
+    }
+
+    pub(crate) fn min_stateful_len(&self) -> usize {
+        self.output_to_reused_addr_count
+            .min_stateful_len()
+            .min(self.input_from_reused_addr_count.min_stateful_len())
+            .min(self.active_reused_addr_count.block.len())
+            .min(self.active_reused_addr_share.block.len())
+    }
+
+    pub(crate) fn par_iter_height_mut(
+        &mut self,
+    ) -> impl ParallelIterator<Item = &mut dyn AnyStoredVec> {
+        self.output_to_reused_addr_count
+            .par_iter_height_mut()
+            .chain(self.input_from_reused_addr_count.par_iter_height_mut())
+            .chain([
+                &mut self.active_reused_addr_count.block as &mut dyn AnyStoredVec,
+                &mut self.active_reused_addr_share.block as &mut dyn AnyStoredVec,
+            ])
+    }
+
+    pub(crate) fn reset_height(&mut self) -> Result<()> {
+        self.output_to_reused_addr_count.reset_height()?;
+        self.input_from_reused_addr_count.reset_height()?;
+        self.active_reused_addr_count.block.reset()?;
+        self.active_reused_addr_share.block.reset()?;
+        Ok(())
+    }
+
+    #[inline(always)]
+    pub(crate) fn push_height(
+        &mut self,
+        uses: &AddrTypeToReusedAddrEventCount,
+        spends: &AddrTypeToReusedAddrEventCount,
+        active_addr_count: u32,
+        active_reused_addr_count: u32,
+    ) {
+        self.output_to_reused_addr_count
+            .push_height(uses.sum(), uses.values().copied());
+        self.input_from_reused_addr_count
+            .push_height(spends.sum(), spends.values().copied());
+        self.active_reused_addr_count
+            .block
+            .push(StoredU32::from(active_reused_addr_count));
+        // Stored as a percentage in [0, 100] to match the rest of the
+        // codebase (Unit.percentage on the website expects 0..100). The
+        // `active_addr_count` denominator lives on `ActivityCountVecs`
+        // (`addrs.activity.all.active`), passed in here so we can
+        // compute the per-block ratio inline.
+        let share = if active_addr_count > 0 {
+            100.0 * (active_reused_addr_count as f32 / active_addr_count as f32)
+        } else {
+            0.0
+        };
+        self.active_reused_addr_share
+            .block
+            .push(StoredF32::from(share));
+    }
+
+    pub(crate) fn compute_rest(
+        &mut self,
+        starting_indexes: &Indexes,
+        outputs_by_type: &outputs::ByTypeVecs,
+        inputs_by_type: &inputs::ByTypeVecs,
+        exit: &Exit,
+    ) -> Result<()> {
+        self.output_to_reused_addr_count
+            .compute_rest(starting_indexes.height, exit)?;
+        self.input_from_reused_addr_count
+            .compute_rest(starting_indexes.height, exit)?;
+        self.active_reused_addr_count
+            .compute_rest(starting_indexes.height, exit)?;
+        self.active_reused_addr_share
+            .compute_rest(starting_indexes.height, exit)?;
+
+        self.output_to_reused_addr_share.all.compute_count_ratio(
+            &self.output_to_reused_addr_count.all,
+            &outputs_by_type.output_count.all,
+            starting_indexes.height,
+            exit,
+        )?;
+        self.spendable_output_to_reused_addr_share.compute_count_ratio(
+            &self.output_to_reused_addr_count.all,
+            &outputs_by_type.spendable_output_count,
+            starting_indexes.height,
+            exit,
+        )?;
+        self.input_from_reused_addr_share.all.compute_count_ratio(
+            &self.input_from_reused_addr_count.all,
+            &inputs_by_type.input_count.all,
+            starting_indexes.height,
+            exit,
+        )?;
+        for otype in OutputType::ADDR_TYPES {
+            self.output_to_reused_addr_share
+                .by_addr_type
+                .get_mut_unwrap(otype)
+                .compute_count_ratio(
+                    self.output_to_reused_addr_count.by_addr_type.get_unwrap(otype),
+                    outputs_by_type.output_count.by_type.get(otype),
+                    starting_indexes.height,
+                    exit,
+                )?;
+            self.input_from_reused_addr_share
+                .by_addr_type
+                .get_mut_unwrap(otype)
+                .compute_count_ratio(
+                    self.input_from_reused_addr_count.by_addr_type.get_unwrap(otype),
+                    inputs_by_type.input_count.by_type.get(otype),
+                    starting_indexes.height,
+                    exit,
+                )?;
+        }
+        Ok(())
+    }
+}

crates/brk_computer/src/distribution/addr/reused/mod.rs

@@ -0,0 +1,94 @@
+//! Reused address tracking.
+//!
+//! An address is "reused" if its lifetime `funded_txo_count > 1`, i.e.
+//! it has received more than one output across its lifetime. This is
+//! the simplest output-multiplicity proxy for address linkability.
+//!
+//! Two facets are tracked here:
+//! - [`count`]: how many distinct addresses are currently reused
+//!   (funded) and how many have *ever* been reused (total). Per address
+//!   type plus an aggregated `all`.
+//! - [`events`]: per-block address-reuse event counts on both sides.
+//!   Output-side (`output_to_reused_addr_count`, outputs landing on
+//!   addresses that had already received ≥ 1 prior output) and
+//!   input-side (`input_from_reused_addr_count`, inputs spending from
+//!   addresses with lifetime `funded_txo_count > 1`). Each count is
+//!   paired with a percent over the matching block-level output/input
+//!   total.
+
+mod count;
+mod events;
+
+pub use count::{AddrTypeToReusedAddrCount, ReusedAddrCountsVecs};
+pub use events::{AddrTypeToReusedAddrEventCount, ReusedAddrEventsVecs};
+
+use brk_error::Result;
+use brk_traversable::Traversable;
+use brk_types::{Indexes, Version};
+use rayon::prelude::*;
+use vecdb::{AnyStoredVec, Database, Exit, Rw, StorageMode};
+
+use crate::{
+    indexes, inputs,
+    internal::{WindowStartVec, Windows},
+    outputs,
+};
+
+/// Top-level container for all reused address tracking: counts (funded +
+/// total) plus per-block reuse events (output-side + input-side).
+#[derive(Traversable)]
+pub struct ReusedAddrVecs<M: StorageMode = Rw> {
+    pub count: ReusedAddrCountsVecs<M>,
+    pub events: ReusedAddrEventsVecs<M>,
+}
+
+impl ReusedAddrVecs {
+    pub(crate) fn forced_import(
+        db: &Database,
+        version: Version,
+        indexes: &indexes::Vecs,
+        cached_starts: &Windows<&WindowStartVec>,
+    ) -> Result<Self> {
+        Ok(Self {
+            count: ReusedAddrCountsVecs::forced_import(db, version, indexes)?,
+            events: ReusedAddrEventsVecs::forced_import(db, version, indexes, cached_starts)?,
+        })
+    }
+
+    pub(crate) fn min_stateful_len(&self) -> usize {
+        self.count
+            .min_stateful_len()
+            .min(self.events.min_stateful_len())
+    }
+
+    pub(crate) fn par_iter_height_mut(
+        &mut self,
+    ) -> impl ParallelIterator<Item = &mut dyn AnyStoredVec> {
+        self.count
+            .par_iter_height_mut()
+            .chain(self.events.par_iter_height_mut())
+    }
+
+    pub(crate) fn reset_height(&mut self) -> Result<()> {
+        self.count.reset_height()?;
+        self.events.reset_height()?;
+        Ok(())
+    }
+
+    pub(crate) fn compute_rest(
+        &mut self,
+        starting_indexes: &Indexes,
+        outputs_by_type: &outputs::ByTypeVecs,
+        inputs_by_type: &inputs::ByTypeVecs,
+        exit: &Exit,
+    ) -> Result<()> {
+        self.count.compute_rest(starting_indexes, exit)?;
+        self.events.compute_rest(
+            starting_indexes,
+            outputs_by_type,
+            inputs_by_type,
+            exit,
+        )?;
+        Ok(())
+    }
+}

crates/brk_computer/src/distribution/addr/total_addr_count.rs

@@ -1,20 +1,21 @@
-use brk_cohort::ByAddrType;
 use brk_error::Result;
 use brk_traversable::Traversable;
 use brk_types::{Height, StoredU64, Version};
+use derive_more::{Deref, DerefMut};
 use vecdb::{Database, Exit, Rw, StorageMode};
 
-use crate::{indexes, internal::PerBlock};
+use crate::{
+    indexes,
+    internal::{PerBlock, WithAddrTypes},
+};
 
 use super::AddrCountsVecs;
 
-/// Total address count (global + per-type) with all derived indexes
+/// Total address count (global + per-type) with all derived indexes.
-#[derive(Traversable)]
+#[derive(Deref, DerefMut, Traversable)]
-pub struct TotalAddrCountVecs<M: StorageMode = Rw> {
+pub struct TotalAddrCountVecs<M: StorageMode = Rw>(
-    pub all: PerBlock<StoredU64, M>,
+    #[traversable(flatten)] pub WithAddrTypes<PerBlock<StoredU64, M>>,
-    #[traversable(flatten)]
+);
-    pub by_addr_type: ByAddrType<PerBlock<StoredU64, M>>,
-}
 
 impl TotalAddrCountVecs {
     pub(crate) fn forced_import(
@@ -22,22 +23,12 @@ impl TotalAddrCountVecs {
         version: Version,
         indexes: &indexes::Vecs,
     ) -> Result<Self> {
-        let all = PerBlock::forced_import(db, "total_addr_count", version, indexes)?;
+        Ok(Self(WithAddrTypes::<PerBlock<StoredU64>>::forced_import(
-
+            db,
-        let by_addr_type: ByAddrType<PerBlock<StoredU64>> =
+            "total_addr_count",
-            ByAddrType::new_with_name(|name| {
+            version,
-                PerBlock::forced_import(
+            indexes,
-                    db,
+        )?))
-                    &format!("{name}_total_addr_count"),
-                    version,
-                    indexes,
-                )
-            })?;
-
-        Ok(Self {
-            all,
-            by_addr_type,
-        })
     }
 
     /// Eagerly compute total = addr_count + empty_addr_count.
@@ -48,14 +39,14 @@ impl TotalAddrCountVecs {
         empty_addr_count: &AddrCountsVecs,
         exit: &Exit,
     ) -> Result<()> {
-        self.all.height.compute_add(
+        self.0.all.height.compute_add(
             max_from,
             &addr_count.all.height,
             &empty_addr_count.all.height,
             exit,
         )?;
 
-        for ((_, total), ((_, addr), (_, empty))) in self.by_addr_type.iter_mut().zip(
+        for ((_, total), ((_, addr), (_, empty))) in self.0.by_addr_type.iter_mut().zip(
             addr_count
                 .by_addr_type
                 .iter()

crates/brk_computer/src/distribution/block/cache/addr.rs

@@ -103,9 +103,7 @@ pub(crate) fn load_uncached_addr_data(
     // Check if this is a new address (type_index >= first for this height)
     let first = *first_addr_indexes.get(addr_type).unwrap();
     if first <= type_index {
-        return Ok(Some(WithAddrDataSource::New(
+        return Ok(Some(WithAddrDataSource::New(FundedAddrData::default())));
-            FundedAddrData::default(),
-        )));
     }
 
     // Skip if already in cache

crates/brk_computer/src/distribution/block/cache/lookup.rs

@@ -26,10 +26,7 @@ impl<'a> AddrLookup<'a> {
         &mut self,
         output_type: OutputType,
         type_index: TypeIndex,
-    ) -> (
+    ) -> (&mut WithAddrDataSource<FundedAddrData>, TrackingStatus) {
-        &mut WithAddrDataSource<FundedAddrData>,
-        TrackingStatus,
-    ) {
         use std::collections::hash_map::Entry;
 
         let map = self.funded.get_mut(output_type).unwrap();

crates/brk_computer/src/distribution/block/cohort/addr_updates.rs

@@ -1,7 +1,7 @@
 use brk_error::Result;
 use brk_types::{
-    AnyAddrIndex, EmptyAddrData, EmptyAddrIndex, FundedAddrData, FundedAddrIndex,
+    AnyAddrIndex, EmptyAddrData, EmptyAddrIndex, FundedAddrData, FundedAddrIndex, OutputType,
-    OutputType, TypeIndex,
+    TypeIndex,
 };
 use vecdb::AnyVec;
 
@@ -67,14 +67,14 @@ pub(crate) fn process_funded_addrs(
 
     // Pure pushes - no holes remain
     addrs_data.funded.reserve_pushed(pushes_iter.len());
-    let mut next_index = addrs_data.funded.len();
+    for (next_index, (addr_type, type_index, data)) in
-    for (addr_type, type_index, data) in pushes_iter {
+        (addrs_data.funded.len()..).zip(pushes_iter)
+    {
         addrs_data.funded.push(data);
         result.get_mut(addr_type).unwrap().insert(
             type_index,
             AnyAddrIndex::from(FundedAddrIndex::from(next_index)),
         );
-        next_index += 1;
     }
 
     Ok(result)
@@ -138,14 +138,14 @@ pub(crate) fn process_empty_addrs(
 
     // Pure pushes - no holes remain
     addrs_data.empty.reserve_pushed(pushes_iter.len());
-    let mut next_index = addrs_data.empty.len();
+    for (next_index, (addr_type, type_index, data)) in
-    for (addr_type, type_index, data) in pushes_iter {
+        (addrs_data.empty.len()..).zip(pushes_iter)
+    {
         addrs_data.empty.push(data);
         result.get_mut(addr_type).unwrap().insert(
             type_index,
             AnyAddrIndex::from(EmptyAddrIndex::from(next_index)),
         );
-        next_index += 1;
     }
 
     Ok(result)

crates/brk_computer/src/distribution/block/cohort/received.rs

@@ -3,7 +3,10 @@ use brk_types::{Cents, Sats, TypeIndex};
 use rustc_hash::FxHashMap;
 
 use crate::distribution::{
-    addr::{AddrTypeToActivityCounts, AddrTypeToVec},
+    addr::{
+        AddrTypeToActivityCounts, AddrTypeToExposedAddrCount, AddrTypeToExposedSupply,
+        AddrTypeToReusedAddrCount, AddrTypeToReusedAddrEventCount, AddrTypeToVec,
+    },
     cohorts::AddrCohorts,
 };
 
@@ -25,8 +28,19 @@ pub(crate) fn process_received(
     addr_count: &mut ByAddrType<u64>,
     empty_addr_count: &mut ByAddrType<u64>,
     activity_counts: &mut AddrTypeToActivityCounts,
+    reused_addr_count: &mut AddrTypeToReusedAddrCount,
+    total_reused_addr_count: &mut AddrTypeToReusedAddrCount,
+    output_to_reused_addr_count: &mut AddrTypeToReusedAddrEventCount,
+    active_reused_addr_count: &mut AddrTypeToReusedAddrEventCount,
+    exposed_addr_count: &mut AddrTypeToExposedAddrCount,
+    total_exposed_addr_count: &mut AddrTypeToExposedAddrCount,
+    exposed_supply: &mut AddrTypeToExposedSupply,
 ) {
-    let max_type_len = received_data.iter().map(|(_, v)| v.len()).max().unwrap_or(0);
+    let max_type_len = received_data
+        .iter()
+        .map(|(_, v)| v.len())
+        .max()
+        .unwrap_or(0);
     let mut aggregated: FxHashMap<TypeIndex, AggregatedReceive> =
         FxHashMap::with_capacity_and_hasher(max_type_len, Default::default());
 
@@ -39,6 +53,13 @@ pub(crate) fn process_received(
         let type_addr_count = addr_count.get_mut(output_type).unwrap();
         let type_empty_count = empty_addr_count.get_mut(output_type).unwrap();
         let type_activity = activity_counts.get_mut_unwrap(output_type);
+        let type_reused_count = reused_addr_count.get_mut(output_type).unwrap();
+        let type_total_reused_count = total_reused_addr_count.get_mut(output_type).unwrap();
+        let type_output_to_reused_count = output_to_reused_addr_count.get_mut(output_type).unwrap();
+        let type_active_reused_count = active_reused_addr_count.get_mut(output_type).unwrap();
+        let type_exposed_count = exposed_addr_count.get_mut(output_type).unwrap();
+        let type_total_exposed_count = total_exposed_addr_count.get_mut(output_type).unwrap();
+        let type_exposed_supply = exposed_supply.get_mut(output_type).unwrap();
 
         // Aggregate receives by address - each address processed exactly once
         for (type_index, value) in vec {
@@ -53,6 +74,13 @@ pub(crate) fn process_received(
             // Track receiving activity - each address in receive aggregation
             type_activity.receiving += 1;
 
+            // Capture state BEFORE the receive mutates funded_txo_count
+            let was_funded = addr_data.is_funded();
+            let was_reused = addr_data.is_reused();
+            let funded_txo_count_before = addr_data.funded_txo_count;
+            let was_pubkey_exposed = addr_data.is_pubkey_exposed(output_type);
+            let exposed_contribution_before = addr_data.exposed_supply_contribution(output_type);
+
             match status {
                 TrackingStatus::New => {
                     *type_addr_count += 1;
@@ -130,6 +158,62 @@ pub(crate) fn process_received(
                         .receive_outputs(addr_data, recv.total_value, price, recv.output_count);
                 }
             }
+
+            // Update reused counts based on the post-receive state
+            let is_now_reused = addr_data.is_reused();
+            if is_now_reused && !was_reused {
+                // Newly crossed the reuse threshold this block
+                *type_reused_count += 1;
+                *type_total_reused_count += 1;
+            } else if is_now_reused && !was_funded {
+                // Already-reused address reactivating into the funded set
+                *type_reused_count += 1;
+            }
+
+            // Block-level "active reused address" count: each address
+            // is processed exactly once here (via aggregation), so we
+            // count it once iff it is reused after the block's receives.
+            // The sender-side counterpart in process_sent dedupes
+            // against `received_addrs` so addresses that did both
+            // aren't double-counted.
+            if is_now_reused {
+                *type_active_reused_count += 1;
+            }
+
+            // Per-block reused-use count: every individual output to this
+            // address counts iff, at the moment the output arrives, the
+            // address had already received at least one prior output
+            // (i.e. it is an output-level "address reuse event"). With
+            // aggregation, that means we skip the very first output the
+            // address ever sees and count every subsequent one, so
+            // `skipped` is `max(0, 1 - before)`.
+            let skipped = 1u32.saturating_sub(funded_txo_count_before);
+            let counted = recv.output_count.saturating_sub(skipped);
+            *type_output_to_reused_count += u64::from(counted);
+
+            // Update exposed counts. The address's pubkey-exposure state
+            // is unchanged by a receive (spent_txo_count unchanged), so we
+            // can use the captured `was_pubkey_exposed` for both pre and post.
+            // After the receive the address is always funded, so it's in the
+            // funded exposed set iff its pubkey is exposed.
+            //
+            // Funded exposed enters when the address wasn't funded before but
+            // is now AND its pubkey is exposed.
+            // Total exposed (pk_exposed_at_funding types only) increments on
+            // first-ever receive (status == TrackingStatus::New); for other
+            // types it's incremented in process_sent on the first spend.
+            if !was_funded && was_pubkey_exposed {
+                *type_exposed_count += 1;
+            }
+            if output_type.pubkey_exposed_at_funding() && matches!(status, TrackingStatus::New) {
+                *type_total_exposed_count += 1;
+            }
+
+            // Update exposed supply via post-receive contribution delta.
+            let exposed_contribution_after = addr_data.exposed_supply_contribution(output_type);
+            // Receives can only add to balance and membership, so the delta
+            // is always non-negative.
+            *type_exposed_supply += exposed_contribution_after - exposed_contribution_before;
         }
     }
 }

crates/brk_computer/src/distribution/block/cohort/sent.rs

@@ -5,7 +5,10 @@ use rustc_hash::FxHashSet;
 use vecdb::VecIndex;
 
 use crate::distribution::{
-    addr::{AddrTypeToActivityCounts, HeightToAddrTypeToVec},
+    addr::{
+        AddrTypeToActivityCounts, AddrTypeToExposedAddrCount, AddrTypeToExposedSupply,
+        AddrTypeToReusedAddrCount, AddrTypeToReusedAddrEventCount, HeightToAddrTypeToVec,
+    },
     cohorts::AddrCohorts,
     compute::PriceRangeMax,
 };
@@ -35,6 +38,12 @@ pub(crate) fn process_sent(
     addr_count: &mut ByAddrType<u64>,
     empty_addr_count: &mut ByAddrType<u64>,
     activity_counts: &mut AddrTypeToActivityCounts,
+    reused_addr_count: &mut AddrTypeToReusedAddrCount,
+    input_from_reused_addr_count: &mut AddrTypeToReusedAddrEventCount,
+    active_reused_addr_count: &mut AddrTypeToReusedAddrEventCount,
+    exposed_addr_count: &mut AddrTypeToExposedAddrCount,
+    total_exposed_addr_count: &mut AddrTypeToExposedAddrCount,
+    exposed_supply: &mut AddrTypeToExposedSupply,
     received_addrs: &ByAddrType<FxHashSet<TypeIndex>>,
     height_to_price: &[Cents],
     height_to_timestamp: &[Timestamp],
@@ -57,12 +66,28 @@ pub(crate) fn process_sent(
             let type_addr_count = addr_count.get_mut(output_type).unwrap();
             let type_empty_count = empty_addr_count.get_mut(output_type).unwrap();
             let type_activity = activity_counts.get_mut_unwrap(output_type);
+            let type_reused_count = reused_addr_count.get_mut(output_type).unwrap();
+            let type_input_from_reused_count =
+                input_from_reused_addr_count.get_mut(output_type).unwrap();
+            let type_active_reused_count = active_reused_addr_count.get_mut(output_type).unwrap();
+            let type_exposed_count = exposed_addr_count.get_mut(output_type).unwrap();
+            let type_total_exposed_count = total_exposed_addr_count.get_mut(output_type).unwrap();
+            let type_exposed_supply = exposed_supply.get_mut(output_type).unwrap();
             let type_received = received_addrs.get(output_type);
             let type_seen = seen_senders.get_mut_unwrap(output_type);
 
             for (type_index, value) in vec {
                 let addr_data = lookup.get_for_send(output_type, type_index);
 
+                // "Input from a reused address" event: the sending
+                // address is in the reused set (lifetime
+                // funded_txo_count > 1). Checked once per input. The
+                // spend itself doesn't touch funded_txo_count so the
+                // predicate is stable before/after `cohort_state.send`.
+                if addr_data.is_reused() {
+                    *type_input_from_reused_count += 1;
+                }
+
                 let prev_balance = addr_data.balance();
                 let new_balance = prev_balance.checked_sub(value).unwrap();
 
@@ -70,14 +95,29 @@ pub(crate) fn process_sent(
                 if type_seen.insert(type_index) {
                     type_activity.sending += 1;
 
-                    // Track "both" - addresses that sent AND received this block
+                    let also_received = type_received.is_some_and(|s| s.contains(&type_index));
-                    if type_received.is_some_and(|s| s.contains(&type_index)) {
+                    // Track "bidirectional": addresses that sent AND
-                        type_activity.both += 1;
+                    // received this block.
+                    if also_received {
+                        type_activity.bidirectional += 1;
+                    }
+
+                    // Block-level "active reused address" count: count
+                    // every distinct sender that's reused, but skip
+                    // those that also received this block (already
+                    // counted in process_received).
+                    if !also_received && addr_data.is_reused() {
+                        *type_active_reused_count += 1;
                     }
                 }
 
                 let will_be_empty = addr_data.has_1_utxos();
 
+                // Capture exposed state BEFORE the spend mutates spent_txo_count.
+                let was_pubkey_exposed = addr_data.is_pubkey_exposed(output_type);
+                let exposed_contribution_before =
+                    addr_data.exposed_supply_contribution(output_type);
+
                 // Compute buckets once
                 let prev_bucket = AmountBucket::from(prev_balance);
                 let new_bucket = AmountBucket::from(new_balance);
@@ -91,6 +131,28 @@ pub(crate) fn process_sent(
                     .unwrap();
 
                 cohort_state.send(addr_data, value, current_price, prev_price, peak_price, age)?;
+                // addr_data.spent_txo_count is now incremented by 1.
+
+                // Update exposed supply via post-spend contribution delta.
+                let exposed_contribution_after = addr_data.exposed_supply_contribution(output_type);
+                if exposed_contribution_after >= exposed_contribution_before {
+                    *type_exposed_supply +=
+                        exposed_contribution_after - exposed_contribution_before;
+                } else {
+                    *type_exposed_supply -=
+                        exposed_contribution_before - exposed_contribution_after;
+                }
+
+                // Update exposed counts on first-ever pubkey exposure.
+                // For non-pk-exposed types this fires on the first spend; for
+                // pk-exposed types it never fires here (was_pubkey_exposed was
+                // already true at first receive in process_received).
+                if !was_pubkey_exposed {
+                    *type_total_exposed_count += 1;
+                    if !will_be_empty {
+                        *type_exposed_count += 1;
+                    }
+                }
 
                 // If crossing a bucket boundary, remove the (now-updated) address from old bucket
                 if will_be_empty || crossing_boundary {
@@ -101,6 +163,17 @@ pub(crate) fn process_sent(
                 if will_be_empty {
                     *type_addr_count -= 1;
                     *type_empty_count += 1;
+                    // Reused addr leaving the funded reused set
+                    if addr_data.is_reused() {
+                        *type_reused_count -= 1;
+                    }
+                    // Exposed addr leaving the funded exposed set: was in set
+                    // iff its pubkey was exposed pre-spend (since it was funded
+                    // to be in process_sent in the first place), and now leaves
+                    // because it's empty.
+                    if was_pubkey_exposed {
+                        *type_exposed_count -= 1;
+                    }
                     lookup.move_to_empty(output_type, type_index);
                 } else if crossing_boundary {
                     cohorts

crates/brk_computer/src/distribution/block/cohort/tx_counts.rs

@@ -41,10 +41,7 @@ pub(crate) fn update_tx_counts(
             .get_mut(&type_index)
         {
             addr_data.tx_count += tx_count;
-        } else if let Some(addr_data) = empty_cache
+        } else if let Some(addr_data) = empty_cache.get_mut(addr_type).unwrap().get_mut(&type_index)
-            .get_mut(addr_type)
-            .unwrap()
-            .get_mut(&type_index)
         {
             addr_data.tx_count += tx_count;
         }

crates/brk_computer/src/distribution/block/utxo/inputs.rs

@@ -90,9 +90,7 @@ pub(crate) fn process_inputs(
     };
 
     let items: Vec<_> = if input_count < 128 {
-        (0..input_count)
+        (0..input_count).map(map_fn).collect::<Result<Vec<_>>>()?
-            .map(map_fn)
-            .collect::<Result<Vec<_>>>()?
     } else {
         (0..input_count)
             .into_par_iter()
@@ -109,10 +107,9 @@ pub(crate) fn process_inputs(
         Default::default(),
     );
     let mut sent_data = HeightToAddrTypeToVec::with_capacity(estimated_unique_heights);
-    let mut addr_data =
+    let mut addr_data = AddrTypeToTypeIndexMap::<WithAddrDataSource<FundedAddrData>>::with_capacity(
-        AddrTypeToTypeIndexMap::<WithAddrDataSource<FundedAddrData>>::with_capacity(
+        estimated_per_type,
-            estimated_per_type,
+    );
-        );
     let mut tx_index_vecs =
         AddrTypeToTypeIndexMap::<SmallVec<[TxIndex; 4]>>::with_capacity(estimated_per_type);
 

crates/brk_computer/src/distribution/block/utxo/outputs.rs

@@ -5,9 +5,7 @@ use rayon::prelude::*;
 use smallvec::SmallVec;
 
 use crate::distribution::{
-    addr::{
+    addr::{AddrTypeToTypeIndexMap, AddrTypeToVec, AddrsDataVecs, AnyAddrIndexesVecs},
-        AddrTypeToTypeIndexMap, AddrTypeToVec, AddrsDataVecs, AnyAddrIndexesVecs,
-    },
     compute::{TxOutData, VecsReaders},
     state::Transacted,
 };
@@ -79,9 +77,7 @@ pub(crate) fn process_outputs(
     };
 
     let items: Vec<_> = if output_count < 128 {
-        (0..output_count)
+        (0..output_count).map(map_fn).collect::<Result<Vec<_>>>()?
-            .map(map_fn)
-            .collect::<Result<Vec<_>>>()?
     } else {
         (0..output_count)
             .into_par_iter()
@@ -93,10 +89,9 @@ pub(crate) fn process_outputs(
     let estimated_per_type = (output_count / 8).max(8);
     let mut transacted = Transacted::default();
     let mut received_data = AddrTypeToVec::with_capacity(estimated_per_type);
-    let mut addr_data =
+    let mut addr_data = AddrTypeToTypeIndexMap::<WithAddrDataSource<FundedAddrData>>::with_capacity(
-        AddrTypeToTypeIndexMap::<WithAddrDataSource<FundedAddrData>>::with_capacity(
+        estimated_per_type,
-            estimated_per_type,
+    );
-        );
     let mut tx_index_vecs =
         AddrTypeToTypeIndexMap::<SmallVec<[TxIndex; 4]>>::with_capacity(estimated_per_type);