@opencodehub/ingestion 0.1.0

package/dist/pipeline/orchestrator.d.ts

@@ -0,0 +1,105 @@
+/**
+ * Top-level orchestrator that constructs a `PipelineContext`, runs the
+ * configured phase set, and returns a summary plus the hashed graph.
+ *
+ * The orchestrator does not touch storage — the returned
+ * `KnowledgeGraph` is in-memory only. Persisting it (DuckDB / embeddings)
+ * is a CLI concern (see `codehub analyze`, which opens a writable store
+ * and calls `bulkLoad`).
+ */
+import { KnowledgeGraph } from "@opencodehub/core-types";
+import { type CochangeOutput } from "./phases/cochange.js";
+import { type EmbedderPhaseOutput, type EmbeddingHashCacheAdapter } from "./phases/embeddings.js";
+import { type IncrementalScopeOutput } from "./phases/incremental-scope.js";
+import { type ScanOutput } from "./phases/scan.js";
+import { type SummarizePhaseOutput, type SummaryCacheAdapter } from "./phases/summarize.js";
+import type { PipelineOptions, PipelinePhase, ProgressEvent } from "./types.js";
+export interface RunPipelineResult {
+    readonly graph: KnowledgeGraph;
+    readonly graphHash: string;
+    readonly stats: {
+        readonly totalMs: number;
+        readonly phases: readonly {
+            readonly name: string;
+            readonly durationMs: number;
+        }[];
+        readonly nodeCount: number;
+        readonly edgeCount: number;
+        readonly byKind?: Record<string, number>;
+        readonly byRelation?: Record<string, number>;
+        readonly schemaVersion?: string;
+        readonly currentCommit?: string;
+        /**
+         * Cache hit / miss counts from the parse phase. Populated on every run
+         * because `parsePhase` tracks them unconditionally; a fully-cold run
+         * reports `{hits: 0, misses: N}` which produces `cacheHitRatio === 0`.
+         */
+        readonly parseCache?: {
+            readonly hits: number;
+            readonly misses: number;
+            readonly ratio: number;
+        };
+    };
+    readonly warnings: readonly string[];
+    /**
+     * Output of the `embeddings` phase when present. `undefined` only in the
+     * pathological case of a custom phase set that omits it entirely. When
+     * embeddings were disabled the phase still ran and returned a zero-ed
+     * output so this field remains defined.
+     */
+    readonly embeddings?: EmbedderPhaseOutput;
+    /**
+     * Output of the `incremental-scope` phase when it ran. Absent when the
+     * operator supplied a custom phase set that excluded it. Downstream code
+     * (analyze CLI, doctor) reads `mode` and `closureRatio` for reporting.
+     */
+    readonly incrementalScope?: IncrementalScopeOutput;
+    /**
+     * Output of the `scan` phase. Exposed so CLIs can persist the post-scan
+     * state (content hashes per file) for the next run's incremental pass.
+     * Always present under the default phase set; absent only when a custom
+     * set omits scan entirely (which would fail validation anyway).
+     */
+    readonly scan?: ScanOutput;
+    /**
+     * Output of the `cochange` phase. Carries the file×file association-rule
+     * rows that the CLI persists into the dedicated `cochanges` storage table
+     * after `bulkLoad`. Absent only when a custom phase set omits cochange.
+     */
+    readonly cochange?: CochangeOutput;
+    /**
+     * Output of the `summarize` phase. Carries any fresh `SymbolSummaryRow`
+     * entries the CLI persists into the `symbol_summaries` table. Absent only
+     * when a custom phase set omits summarize; the default phase set always
+     * runs it (the phase internally short-circuits when gated off).
+     */
+    readonly summarize?: SummarizePhaseOutput;
+}
+export interface RunIngestionOptions extends PipelineOptions {
+    readonly phases?: readonly PipelinePhase[];
+    readonly onProgress?: (ev: ProgressEvent) => void;
+    /**
+     * Optional adapter the summarize phase probes before issuing work.
+     * Production wires this to the DuckDB store's `lookupSymbolSummary`
+     * implementation so re-indexes become free when source hasn't drifted.
+     * Tests inject an in-memory fake. Absent by default — the phase degrades
+     * to "every candidate is a miss" which is still correct, just more
+     * expensive.
+     */
+    readonly summaryCacheAdapter?: SummaryCacheAdapter;
+    /**
+     * Optional adapter the embeddings phase probes before issuing embedder
+     * calls. Production wires this to the DuckDB store's
+     * `listEmbeddingHashes` implementation so re-analyze runs skip chunks
+     * whose `content_hash` matches a prior row. Absent by default —
+     * the phase degrades to "every chunk is new" which is still correct,
+     * just more expensive. Ignored when `options.force === true`.
+     */
+    readonly embeddingHashCacheAdapter?: EmbeddingHashCacheAdapter;
+}
+/**
+ * Kick off an end-to-end pipeline run. When `options.phases` is omitted
+ * the default phase set is used.
+ */
+export declare function runIngestion(repoPath: string, options?: RunIngestionOptions): Promise<RunPipelineResult>;
+//# sourceMappingURL=orchestrator.d.ts.map

package/dist/pipeline/orchestrator.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"orchestrator.d.ts","sourceRoot":"","sources":["../../src/pipeline/orchestrator.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAEH,OAAO,EAAa,cAAc,EAAE,MAAM,yBAAyB,CAAC;AAEpE,OAAO,EAAuB,KAAK,cAAc,EAAE,MAAM,sBAAsB,CAAC;AAEhF,OAAO,EAGL,KAAK,mBAAmB,EACxB,KAAK,yBAAyB,EAC/B,MAAM,wBAAwB,CAAC;AAChC,OAAO,EAEL,KAAK,sBAAsB,EAC5B,MAAM,+BAA+B,CAAC;AAEvC,OAAO,EAAmB,KAAK,UAAU,EAAE,MAAM,kBAAkB,CAAC;AACpE,OAAO,EAGL,KAAK,oBAAoB,EACzB,KAAK,mBAAmB,EACzB,MAAM,uBAAuB,CAAC;AAE/B,OAAO,KAAK,EAGV,eAAe,EACf,aAAa,EAEb,aAAa,EACd,MAAM,YAAY,CAAC;AAEpB,MAAM,WAAW,iBAAiB;IAChC,QAAQ,CAAC,KAAK,EAAE,cAAc,CAAC;IAC/B,QAAQ,CAAC,SAAS,EAAE,MAAM,CAAC;IAC3B,QAAQ,CAAC,KAAK,EAAE;QACd,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAC;QACzB,QAAQ,CAAC,MAAM,EAAE,SAAS;YAAE,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;YAAC,QAAQ,CAAC,UAAU,EAAE,MAAM,CAAA;SAAE,EAAE,CAAC;QACnF,QAAQ,CAAC,SAAS,EAAE,MAAM,CAAC;QAC3B,QAAQ,CAAC,SAAS,EAAE,MAAM,CAAC;QAC3B,QAAQ,CAAC,MAAM,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;QACzC,QAAQ,CAAC,UAAU,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;QAC7C,QAAQ,CAAC,aAAa,CAAC,EAAE,MAAM,CAAC;QAChC,QAAQ,CAAC,aAAa,CAAC,EAAE,MAAM,CAAC;QAChC;;;;WAIG;QACH,QAAQ,CAAC,UAAU,CAAC,EAAE;YACpB,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;YACtB,QAAQ,CAAC,MAAM,EAAE,MAAM,CAAC;YACxB,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;SACxB,CAAC;KACH,CAAC;IACF,QAAQ,CAAC,QAAQ,EAAE,SAAS,MAAM,EAAE,CAAC;IACrC;;;;;OAKG;IACH,QAAQ,CAAC,UAAU,CAAC,EAAE,mBAAmB,CAAC;IAC1C;;;;OAIG;IACH,QAAQ,CAAC,gBAAgB,CAAC,EAAE,sBAAsB,CAAC;IACnD;;;;;OAKG;IACH,QAAQ,CAAC,IAAI,CAAC,EAAE,UAAU,CAAC;IAC3B;;;;OAIG;IACH,QAAQ,CAAC,QAAQ,CAAC,EAAE,cAAc,CAAC;IACnC;;;;;OAKG;IACH,QAAQ,CAAC,SAAS,CAAC,EAAE,oBAAoB,CAAC;CAC3C;AAED,MAAM,WAAW,mBAAoB,SAAQ,eAAe;IAC1D,QAAQ,CAAC,MAAM,CAAC,EAAE,SAAS,aAAa,EAAE,CAAC;IAC3C,QAAQ,CAAC,UAAU,CAAC,EAAE,CAAC,EAAE,EAAE,aAAa,KAAK,IAAI,CAAC;IAClD;;;;;;;OAOG;IACH,QAAQ,CAAC,mBAAmB,CAAC,EAAE,mBAAmB,CAAC;IACnD;;;;;;;OAOG;IACH,QAAQ,CAAC,yBAAyB,CAAC,EAAE,yBAAyB,CAAC;CAChE;AAED;;;GAGG;AACH,wBAAsB,YAAY,CAChC,QAAQ,EAAE,MAAM,EAChB,OAAO,GAAE,mBAAwB,GAChC,OAAO,CAAC,iBAAiB,CAAC,CA4H5B"}

package/dist/pipeline/orchestrator.js

@@ -0,0 +1,175 @@
+/**
+ * Top-level orchestrator that constructs a `PipelineContext`, runs the
+ * configured phase set, and returns a summary plus the hashed graph.
+ *
+ * The orchestrator does not touch storage — the returned
+ * `KnowledgeGraph` is in-memory only. Persisting it (DuckDB / embeddings)
+ * is a CLI concern (see `codehub analyze`, which opens a writable store
+ * and calls `bulkLoad`).
+ */
+import { graphHash, KnowledgeGraph } from "@opencodehub/core-types";
+import { ANNOTATE_PHASE_NAME } from "./phases/annotate.js";
+import { COCHANGE_PHASE_NAME } from "./phases/cochange.js";
+import { DEFAULT_PHASES } from "./phases/default-set.js";
+import { EMBEDDER_PHASE_NAME, EMBEDDING_HASH_CACHE_OPTIONS_KEY, } from "./phases/embeddings.js";
+import { INCREMENTAL_SCOPE_PHASE_NAME, } from "./phases/incremental-scope.js";
+import { PARSE_PHASE_NAME } from "./phases/parse.js";
+import { SCAN_PHASE_NAME } from "./phases/scan.js";
+import { SUMMARIZE_PHASE_NAME, SUMMARY_CACHE_OPTIONS_KEY, } from "./phases/summarize.js";
+import { runPipeline } from "./runner.js";
+/**
+ * Kick off an end-to-end pipeline run. When `options.phases` is omitted
+ * the default phase set is used.
+ */
+export async function runIngestion(repoPath, options = {}) {
+    const phases = options.phases ?? DEFAULT_PHASES;
+    const normalizedOptions = stripPhaseKeys(options);
+    // Attach the optional summary-cache adapter onto the options bag via a
+    // well-known key. The `summarize` phase reads it back via an unchecked
+    // cast; keeping the attach-point here (rather than inside stripPhaseKeys)
+    // keeps the typed fields in stripPhaseKeys honest.
+    if (options.summaryCacheAdapter !== undefined) {
+        normalizedOptions[SUMMARY_CACHE_OPTIONS_KEY] =
+            options.summaryCacheAdapter;
+    }
+    // Same trick for the embeddings phase's content-hash cache.
+    // Attached here (not in stripPhaseKeys) so the typed option shape stays
+    // minimal — this is a well-known extension point, not a first-class
+    // `PipelineOptions` field.
+    if (options.embeddingHashCacheAdapter !== undefined) {
+        normalizedOptions[EMBEDDING_HASH_CACHE_OPTIONS_KEY] =
+            options.embeddingHashCacheAdapter;
+    }
+    const graph = new KnowledgeGraph();
+    const warnings = [];
+    const ctx = {
+        repoPath,
+        options: normalizedOptions,
+        graph,
+        phaseOutputs: new Map(),
+        ...(options.onProgress !== undefined
+            ? {
+                onProgress: (ev) => {
+                    if (ev.kind === "warn" && ev.message !== undefined) {
+                        warnings.push(ev.message);
+                    }
+                    options.onProgress?.(ev);
+                },
+            }
+            : {
+                onProgress: (ev) => {
+                    if (ev.kind === "warn" && ev.message !== undefined) {
+                        warnings.push(ev.message);
+                    }
+                },
+            }),
+    };
+    const started = Date.now();
+    const results = await runPipeline(phases, ctx);
+    const totalMs = Date.now() - started;
+    // If `annotate` ran, pull the richer stats and schema metadata out onto
+    // the orchestrator-level result so the CLI can surface them without
+    // reaching into the phase-output map.
+    const annotate = results.find((r) => r.name === ANNOTATE_PHASE_NAME)?.output;
+    const embeddings = results.find((r) => r.name === EMBEDDER_PHASE_NAME)?.output;
+    const incrementalScope = results.find((r) => r.name === INCREMENTAL_SCOPE_PHASE_NAME)?.output;
+    const parse = results.find((r) => r.name === PARSE_PHASE_NAME)?.output;
+    const scan = results.find((r) => r.name === SCAN_PHASE_NAME)?.output;
+    const cochange = results.find((r) => r.name === COCHANGE_PHASE_NAME)?.output;
+    const summarize = results.find((r) => r.name === SUMMARIZE_PHASE_NAME)?.output;
+    const parseCache = parse !== undefined
+        ? {
+            hits: parse.cacheHits,
+            misses: parse.cacheMisses,
+            ratio: parse.cacheHits + parse.cacheMisses === 0
+                ? 0
+                : parse.cacheHits / (parse.cacheHits + parse.cacheMisses),
+        }
+        : undefined;
+    // DEBUG_PHASE_MEM=1 brackets the graphHash call so large monorepos (where
+    // the hash can take several seconds) don't look like a hang relative to
+    // the per-phase telemetry emitted by the runner. Entirely gated — safe to
+    // leave in; costs are two stderr writes + one Date.now() sample.
+    const phaseMemDebug = process.env["DEBUG_PHASE_MEM"] === "1";
+    if (phaseMemDebug) {
+        process.stderr.write(`[phase-telemetry] graphHash-start nodes=${graph.nodeCount()} edges=${graph.edgeCount()}\n`);
+    }
+    const hashStart = Date.now();
+    const hashed = graphHash(graph);
+    if (phaseMemDebug) {
+        process.stderr.write(`[phase-telemetry] graphHash-end dur=${Date.now() - hashStart}ms\n`);
+    }
+    return {
+        graph,
+        graphHash: hashed,
+        stats: {
+            totalMs,
+            phases: results.map(summarizePhase),
+            nodeCount: graph.nodeCount(),
+            edgeCount: graph.edgeCount(),
+            ...(annotate !== undefined
+                ? {
+                    byKind: annotate.stats.byKind,
+                    byRelation: annotate.stats.byRelation,
+                    schemaVersion: annotate.schemaVersion,
+                    ...(annotate.currentCommit !== undefined
+                        ? { currentCommit: annotate.currentCommit }
+                        : {}),
+                }
+                : {}),
+            ...(parseCache !== undefined ? { parseCache } : {}),
+        },
+        warnings,
+        ...(embeddings !== undefined ? { embeddings } : {}),
+        ...(incrementalScope !== undefined ? { incrementalScope } : {}),
+        ...(scan !== undefined ? { scan } : {}),
+        ...(cochange !== undefined ? { cochange } : {}),
+        ...(summarize !== undefined ? { summarize } : {}),
+    };
+}
+function summarizePhase(r) {
+    return { name: r.name, durationMs: r.durationMs };
+}
+function stripPhaseKeys(options) {
+    // Copy only the fields PipelineOptions declares — phase overrides and
+    // progress callbacks are orchestrator-level concerns and must not leak
+    // into the normalized per-phase options.
+    const typed = {};
+    if (options.force !== undefined)
+        typed.force = options.force;
+    if (options.offline !== undefined)
+        typed.offline = options.offline;
+    if (options.verbose !== undefined)
+        typed.verbose = options.verbose;
+    if (options.skipGit !== undefined)
+        typed.skipGit = options.skipGit;
+    if (options.byteCapPerFile !== undefined)
+        typed.byteCapPerFile = options.byteCapPerFile;
+    if (options.maxTotalFiles !== undefined)
+        typed.maxTotalFiles = options.maxTotalFiles;
+    if (options.embeddings !== undefined)
+        typed.embeddings = options.embeddings;
+    if (options.embeddingsVariant !== undefined)
+        typed.embeddingsVariant = options.embeddingsVariant;
+    if (options.sbom !== undefined)
+        typed.sbom = options.sbom;
+    if (options.reproducibleSbom !== undefined)
+        typed.reproducibleSbom = options.reproducibleSbom;
+    if (options.embeddingsModelDir !== undefined) {
+        typed.embeddingsModelDir = options.embeddingsModelDir;
+    }
+    if (options.embeddingsGranularity !== undefined) {
+        typed.embeddingsGranularity = options.embeddingsGranularity;
+    }
+    if (options.incrementalFrom !== undefined)
+        typed.incrementalFrom = options.incrementalFrom;
+    if (options.summaries !== undefined)
+        typed.summaries = options.summaries;
+    if (options.maxSummariesPerRun !== undefined) {
+        typed.maxSummariesPerRun = options.maxSummariesPerRun;
+    }
+    if (options.summaryModel !== undefined)
+        typed.summaryModel = options.summaryModel;
+    return typed;
+}
+//# sourceMappingURL=orchestrator.js.map

package/dist/pipeline/orchestrator.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"orchestrator.js","sourceRoot":"","sources":["../../src/pipeline/orchestrator.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAEH,OAAO,EAAE,SAAS,EAAE,cAAc,EAAE,MAAM,yBAAyB,CAAC;AACpE,OAAO,EAAE,mBAAmB,EAAuB,MAAM,sBAAsB,CAAC;AAChF,OAAO,EAAE,mBAAmB,EAAuB,MAAM,sBAAsB,CAAC;AAChF,OAAO,EAAE,cAAc,EAAE,MAAM,yBAAyB,CAAC;AACzD,OAAO,EACL,mBAAmB,EACnB,gCAAgC,GAGjC,MAAM,wBAAwB,CAAC;AAChC,OAAO,EACL,4BAA4B,GAE7B,MAAM,+BAA+B,CAAC;AACvC,OAAO,EAAE,gBAAgB,EAAoB,MAAM,mBAAmB,CAAC;AACvE,OAAO,EAAE,eAAe,EAAmB,MAAM,kBAAkB,CAAC;AACpE,OAAO,EACL,oBAAoB,EACpB,yBAAyB,GAG1B,MAAM,uBAAuB,CAAC;AAC/B,OAAO,EAAE,WAAW,EAAE,MAAM,aAAa,CAAC;AA4F1C;;;GAGG;AACH,MAAM,CAAC,KAAK,UAAU,YAAY,CAChC,QAAgB,EAChB,UAA+B,EAAE;IAEjC,MAAM,MAAM,GAAG,OAAO,CAAC,MAAM,IAAI,cAAc,CAAC;IAChD,MAAM,iBAAiB,GAAoB,cAAc,CAAC,OAAO,CAAC,CAAC;IACnE,uEAAuE;IACvE,uEAAuE;IACvE,0EAA0E;IAC1E,mDAAmD;IACnD,IAAI,OAAO,CAAC,mBAAmB,KAAK,SAAS,EAAE,CAAC;QAC7C,iBAAwD,CAAC,yBAAyB,CAAC;YAClF,OAAO,CAAC,mBAAmB,CAAC;IAChC,CAAC;IACD,4DAA4D;IAC5D,wEAAwE;IACxE,oEAAoE;IACpE,2BAA2B;IAC3B,IAAI,OAAO,CAAC,yBAAyB,KAAK,SAAS,EAAE,CAAC;QACnD,iBAAwD,CAAC,gCAAgC,CAAC;YACzF,OAAO,CAAC,yBAAyB,CAAC;IACtC,CAAC;IACD,MAAM,KAAK,GAAG,IAAI,cAAc,EAAE,CAAC;IACnC,MAAM,QAAQ,GAAa,EAAE,CAAC;IAE9B,MAAM,GAAG,GAAoB;QAC3B,QAAQ;QACR,OAAO,EAAE,iBAAiB;QAC1B,KAAK;QACL,YAAY,EAAE,IAAI,GAAG,EAAE;QACvB,GAAG,CAAC,OAAO,CAAC,UAAU,KAAK,SAAS;YAClC,CAAC,CAAC;gBACE,UAAU,EAAE,CAAC,EAAE,EAAE,EAAE;oBACjB,IAAI,EAAE,CAAC,IAAI,KAAK,MAAM,IAAI,EAAE,CAAC,OAAO,KAAK,SAAS,EAAE,CAAC;wBACnD,QAAQ,CAAC,IAAI,CAAC,EAAE,CAAC,OAAO,CAAC,CAAC;oBAC5B,CAAC;oBACD,OAAO,CAAC,UAAU,EAAE,CAAC,EAAE,CAAC,CAAC;gBAC3B,CAAC;aACF;YACH,CAAC,CAAC;gBACE,UAAU,EAAE,CAAC,EAAE,EAAE,EAAE;oBACjB,IAAI,EAAE,CAAC,IAAI,KAAK,MAAM,IAAI,EAAE,CAAC,OAAO,KAAK,SAAS,EAAE,CAAC;wBACnD,QAAQ,CAAC,IAAI,CAAC,EAAE,CAAC,OAAO,CAAC,CAAC;oBAC5B,CAAC;gBACH,CAAC;aACF,CAAC;KACP,CAAC;IAEF,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;IAC3B,MAAM,OAAO,GAAG,MAAM,WAAW,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC;IAC/C,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,EAAE,GAAG,OAAO,CAAC;IAErC,wEAAwE;IACxE,oEAAoE;IACpE,sCAAsC;IACtC,MAAM,QAAQ,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,mBAAmB,CAAC,EAAE,MAEzD,CAAC;IACd,MAAM,UAAU,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,mBAAmB,CAAC,EAAE,MAE3D,CAAC;IACd,MAAM,gBAAgB,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,4BAA4B,CAAC,EAAE,MAE1E,CAAC;IACd,MAAM,KAAK,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,gBAAgB,CAAC,EAAE,MAAiC,CAAC;IAClG,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,eAAe,CAAC,EAAE,MAAgC,CAAC;IAC/F,MAAM,QAAQ,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,mBAAmB,CAAC,EAAE,MAEzD,CAAC;IACd,MAAM,SAAS,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,oBAAoB,CAAC,EAAE,MAE3D,CAAC;IAEd,MAAM,UAAU,GACd,KAAK,KAAK,SAAS;QACjB,CAAC,CAAC;YACE,IAAI,EAAE,KAAK,CAAC,SAAS;YACrB,MAAM,EAAE,KAAK,CAAC,WAAW;YACzB,KAAK,EACH,KAAK,CAAC,SAAS,GAAG,KAAK,CAAC,WAAW,KAAK,CAAC;gBACvC,CAAC,CAAC,CAAC;gBACH,CAAC,CAAC,KAAK,CAAC,SAAS,GAAG,CAAC,KAAK,CAAC,SAAS,GAAG,KAAK,CAAC,WAAW,CAAC;SAC9D;QACH,CAAC,CAAC,SAAS,CAAC;IAEhB,0EAA0E;IAC1E,wEAAwE;IACxE,0EAA0E;IAC1E,iEAAiE;IACjE,MAAM,aAAa,GAAG,OAAO,CAAC,GAAG,CAAC,iBAAiB,CAAC,KAAK,GAAG,CAAC;IAC7D,IAAI,aAAa,EAAE,CAAC;QAClB,OAAO,CAAC,MAAM,CAAC,KAAK,CAClB,2CAA2C,KAAK,CAAC,SAAS,EAAE,UAAU,KAAK,CAAC,SAAS,EAAE,IAAI,CAC5F,CAAC;IACJ,CAAC;IACD,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;IAC7B,MAAM,MAAM,GAAG,SAAS,CAAC,KAAK,CAAC,CAAC;IAChC,IAAI,aAAa,EAAE,CAAC;QAClB,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,uCAAuC,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS,MAAM,CAAC,CAAC;IAC5F,CAAC;IACD,OAAO;QACL,KAAK;QACL,SAAS,EAAE,MAAM;QACjB,KAAK,EAAE;YACL,OAAO;YACP,MAAM,EAAE,OAAO,CAAC,GAAG,CAAC,cAAc,CAAC;YACnC,SAAS,EAAE,KAAK,CAAC,SAAS,EAAE;YAC5B,SAAS,EAAE,KAAK,CAAC,SAAS,EAAE;YAC5B,GAAG,CAAC,QAAQ,KAAK,SAAS;gBACxB,CAAC,CAAC;oBACE,MAAM,EAAE,QAAQ,CAAC,KAAK,CAAC,MAAM;oBAC7B,UAAU,EAAE,QAAQ,CAAC,KAAK,CAAC,UAAU;oBACrC,aAAa,EAAE,QAAQ,CAAC,aAAa;oBACrC,GAAG,CAAC,QAAQ,CAAC,aAAa,KAAK,SAAS;wBACtC,CAAC,CAAC,EAAE,aAAa,EAAE,QAAQ,CAAC,aAAa,EAAE;wBAC3C,CAAC,CAAC,EAAE,CAAC;iBACR;gBACH,CAAC,CAAC,EAAE,CAAC;YACP,GAAG,CAAC,UAAU,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,UAAU,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;SACpD;QACD,QAAQ;QACR,GAAG,CAAC,UAAU,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,UAAU,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QACnD,GAAG,CAAC,gBAAgB,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,gBAAgB,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAC/D,GAAG,CAAC,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QACvC,GAAG,CAAC,QAAQ,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAC/C,GAAG,CAAC,SAAS,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,SAAS,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;KAClD,CAAC;AACJ,CAAC;AAED,SAAS,cAAc,CAAC,CAAc;IACpC,OAAO,EAAE,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,UAAU,EAAE,CAAC,CAAC,UAAU,EAAE,CAAC;AACpD,CAAC;AAED,SAAS,cAAc,CAAC,OAA4B;IAClD,sEAAsE;IACtE,uEAAuE;IACvE,yCAAyC;IACzC,MAAM,KAAK,GAiBP,EAAE,CAAC;IACP,IAAI,OAAO,CAAC,KAAK,KAAK,SAAS;QAAE,KAAK,CAAC,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC;IAC7D,IAAI,OAAO,CAAC,OAAO,KAAK,SAAS;QAAE,KAAK,CAAC,OAAO,GAAG,OAAO,CAAC,OAAO,CAAC;IACnE,IAAI,OAAO,CAAC,OAAO,KAAK,SAAS;QAAE,KAAK,CAAC,OAAO,GAAG,OAAO,CAAC,OAAO,CAAC;IACnE,IAAI,OAAO,CAAC,OAAO,KAAK,SAAS;QAAE,KAAK,CAAC,OAAO,GAAG,OAAO,CAAC,OAAO,CAAC;IACnE,IAAI,OAAO,CAAC,cAAc,KAAK,SAAS;QAAE,KAAK,CAAC,cAAc,GAAG,OAAO,CAAC,cAAc,CAAC;IACxF,IAAI,OAAO,CAAC,aAAa,KAAK,SAAS;QAAE,KAAK,CAAC,aAAa,GAAG,OAAO,CAAC,aAAa,CAAC;IACrF,IAAI,OAAO,CAAC,UAAU,KAAK,SAAS;QAAE,KAAK,CAAC,UAAU,GAAG,OAAO,CAAC,UAAU,CAAC;IAC5E,IAAI,OAAO,CAAC,iBAAiB,KAAK,SAAS;QAAE,KAAK,CAAC,iBAAiB,GAAG,OAAO,CAAC,iBAAiB,CAAC;IACjG,IAAI,OAAO,CAAC,IAAI,KAAK,SAAS;QAAE,KAAK,CAAC,IAAI,GAAG,OAAO,CAAC,IAAI,CAAC;IAC1D,IAAI,OAAO,CAAC,gBAAgB,KAAK,SAAS;QAAE,KAAK,CAAC,gBAAgB,GAAG,OAAO,CAAC,gBAAgB,CAAC;IAC9F,IAAI,OAAO,CAAC,kBAAkB,KAAK,SAAS,EAAE,CAAC;QAC7C,KAAK,CAAC,kBAAkB,GAAG,OAAO,CAAC,kBAAkB,CAAC;IACxD,CAAC;IACD,IAAI,OAAO,CAAC,qBAAqB,KAAK,SAAS,EAAE,CAAC;QAChD,KAAK,CAAC,qBAAqB,GAAG,OAAO,CAAC,qBAAqB,CAAC;IAC9D,CAAC;IACD,IAAI,OAAO,CAAC,eAAe,KAAK,SAAS;QAAE,KAAK,CAAC,eAAe,GAAG,OAAO,CAAC,eAAe,CAAC;IAC3F,IAAI,OAAO,CAAC,SAAS,KAAK,SAAS;QAAE,KAAK,CAAC,SAAS,GAAG,OAAO,CAAC,SAAS,CAAC;IACzE,IAAI,OAAO,CAAC,kBAAkB,KAAK,SAAS,EAAE,CAAC;QAC7C,KAAK,CAAC,kBAAkB,GAAG,OAAO,CAAC,kBAAkB,CAAC;IACxD,CAAC;IACD,IAAI,OAAO,CAAC,YAAY,KAAK,SAAS;QAAE,KAAK,CAAC,YAAY,GAAG,OAAO,CAAC,YAAY,CAAC;IAClF,OAAO,KAAK,CAAC;AACf,CAAC"}

package/dist/pipeline/ownership-helpers/drift.d.ts

@@ -0,0 +1,41 @@
+/**
+ * Ownership drift helpers.
+ *
+ * For each of three rolling windows (30, 90, 365 days) we compute the stddev
+ * of the top-3 contributors' line share across the window. A high stddev
+ * implies ownership changed hands inside the window; a low stddev implies
+ * steady ownership.
+ *
+ * Sampling plan (see research notes §9):
+ *   - 30d window: 4 weekly samples.
+ *   - 90d window: 13 weekly samples.
+ *   - 365d window: 52 weekly samples.
+ *
+ * Samples are anchored to `nowEpochSec`. At each sample tick we project the
+ * set of commits whose `ct <= tick` down onto each contributor's cumulative
+ * line share and record the top-3 contributors' shares. Stddev is taken over
+ * the concatenation of the three contributors' share vectors across ticks.
+ */
+export interface CommitContribution {
+    /** Epoch seconds of the committer timestamp. */
+    readonly ctEpochSec: number;
+    /** Per-contributor line counts this commit applied. */
+    readonly contributions: ReadonlyMap<string, number>;
+}
+export interface OwnershipDriftInput {
+    /** History of per-commit contribution events across the target (file or community). */
+    readonly commits: readonly CommitContribution[];
+    /** Reference "now" in epoch seconds. */
+    readonly nowEpochSec: number;
+}
+export interface OwnershipDriftResult {
+    readonly drift30d: number;
+    readonly drift90d: number;
+    readonly drift365d: number;
+}
+/**
+ * Compute the three rolling drift values. Windows with fewer than two samples
+ * yield drift = 0 (no variation detectable).
+ */
+export declare function computeOwnershipDrift(input: OwnershipDriftInput): OwnershipDriftResult;
+//# sourceMappingURL=drift.d.ts.map

package/dist/pipeline/ownership-helpers/drift.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"drift.d.ts","sourceRoot":"","sources":["../../../src/pipeline/ownership-helpers/drift.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;GAiBG;AAEH,MAAM,WAAW,kBAAkB;IACjC,gDAAgD;IAChD,QAAQ,CAAC,UAAU,EAAE,MAAM,CAAC;IAC5B,uDAAuD;IACvD,QAAQ,CAAC,aAAa,EAAE,WAAW,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;CACrD;AAED,MAAM,WAAW,mBAAmB;IAClC,uFAAuF;IACvF,QAAQ,CAAC,OAAO,EAAE,SAAS,kBAAkB,EAAE,CAAC;IAChD,wCAAwC;IACxC,QAAQ,CAAC,WAAW,EAAE,MAAM,CAAC;CAC9B;AAED,MAAM,WAAW,oBAAoB;IACnC,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAC;IAC1B,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAC;IAC1B,QAAQ,CAAC,SAAS,EAAE,MAAM,CAAC;CAC5B;AAKD;;;GAGG;AACH,wBAAgB,qBAAqB,CAAC,KAAK,EAAE,mBAAmB,GAAG,oBAAoB,CAMtF"}

package/dist/pipeline/ownership-helpers/drift.js

@@ -0,0 +1,122 @@
+/**
+ * Ownership drift helpers.
+ *
+ * For each of three rolling windows (30, 90, 365 days) we compute the stddev
+ * of the top-3 contributors' line share across the window. A high stddev
+ * implies ownership changed hands inside the window; a low stddev implies
+ * steady ownership.
+ *
+ * Sampling plan (see research notes §9):
+ *   - 30d window: 4 weekly samples.
+ *   - 90d window: 13 weekly samples.
+ *   - 365d window: 52 weekly samples.
+ *
+ * Samples are anchored to `nowEpochSec`. At each sample tick we project the
+ * set of commits whose `ct <= tick` down onto each contributor's cumulative
+ * line share and record the top-3 contributors' shares. Stddev is taken over
+ * the concatenation of the three contributors' share vectors across ticks.
+ */
+const DAY_SEC = 86_400;
+const WEEK_SEC = 7 * DAY_SEC;
+/**
+ * Compute the three rolling drift values. Windows with fewer than two samples
+ * yield drift = 0 (no variation detectable).
+ */
+export function computeOwnershipDrift(input) {
+    return {
+        drift30d: driftForWindow(input, 30 * DAY_SEC, 4),
+        drift90d: driftForWindow(input, 90 * DAY_SEC, 13),
+        drift365d: driftForWindow(input, 365 * DAY_SEC, 52),
+    };
+}
+function driftForWindow(input, windowSec, sampleCount) {
+    if (sampleCount < 2)
+        return 0;
+    const windowStart = input.nowEpochSec - windowSec;
+    // Filter commits inside the window; we still want older commits for the
+    // "prior share" baseline but they don't contribute new line-share movement
+    // within the window.
+    const relevantCommits = input.commits.filter((c) => c.ctEpochSec >= windowStart);
+    if (relevantCommits.length === 0)
+        return 0;
+    const sortedCommits = [...input.commits].sort((a, b) => a.ctEpochSec - b.ctEpochSec);
+    const sampleTicks = [];
+    // Tick spacing: one per week, capped at `sampleCount` ticks, newest last.
+    for (let i = sampleCount - 1; i >= 0; i -= 1) {
+        const tickSec = input.nowEpochSec - i * WEEK_SEC;
+        if (tickSec < windowStart)
+            continue;
+        sampleTicks.push(tickSec);
+    }
+    if (sampleTicks.length < 2)
+        return 0;
+    // For each tick, compute cumulative shares and pick top-3 by cumulative
+    // lines. Track each selected contributor's share at the tick.
+    const perContributorShares = new Map();
+    const cumulative = new Map();
+    let sortedIdx = 0;
+    for (const tickSec of sampleTicks) {
+        while (sortedIdx < sortedCommits.length) {
+            const commit = sortedCommits[sortedIdx];
+            if (commit === undefined)
+                break;
+            if (commit.ctEpochSec > tickSec)
+                break;
+            for (const [email, lines] of commit.contributions) {
+                cumulative.set(email, (cumulative.get(email) ?? 0) + lines);
+            }
+            sortedIdx += 1;
+        }
+        let total = 0;
+        for (const v of cumulative.values())
+            total += v;
+        if (total === 0)
+            continue;
+        const top3 = [...cumulative.entries()]
+            .sort((a, b) => {
+            if (a[1] !== b[1])
+                return b[1] - a[1];
+            return a[0] < b[0] ? -1 : a[0] > b[0] ? 1 : 0;
+        })
+            .slice(0, 3);
+        for (const [email, lines] of top3) {
+            const share = lines / total;
+            const arr = perContributorShares.get(email);
+            if (arr === undefined)
+                perContributorShares.set(email, [share]);
+            else
+                arr.push(share);
+        }
+    }
+    // Flatten shares across top-3 contributors and compute stddev.
+    const flat = [];
+    for (const arr of perContributorShares.values()) {
+        for (const v of arr)
+            flat.push(v);
+    }
+    if (flat.length < 2)
+        return 0;
+    return round4(stddev(flat));
+}
+function stddev(values) {
+    const n = values.length;
+    if (n < 2)
+        return 0;
+    let mean = 0;
+    for (const v of values)
+        mean += v;
+    mean /= n;
+    let variance = 0;
+    for (const v of values) {
+        const d = v - mean;
+        variance += d * d;
+    }
+    variance /= n; // population stddev — fixture tests prefer this form.
+    return Math.sqrt(variance);
+}
+function round4(n) {
+    if (!Number.isFinite(n))
+        return 0;
+    return Math.round(n * 10_000) / 10_000;
+}
+//# sourceMappingURL=drift.js.map

package/dist/pipeline/ownership-helpers/drift.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"drift.js","sourceRoot":"","sources":["../../../src/pipeline/ownership-helpers/drift.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;GAiBG;AAsBH,MAAM,OAAO,GAAG,MAAM,CAAC;AACvB,MAAM,QAAQ,GAAG,CAAC,GAAG,OAAO,CAAC;AAE7B;;;GAGG;AACH,MAAM,UAAU,qBAAqB,CAAC,KAA0B;IAC9D,OAAO;QACL,QAAQ,EAAE,cAAc,CAAC,KAAK,EAAE,EAAE,GAAG,OAAO,EAAE,CAAC,CAAC;QAChD,QAAQ,EAAE,cAAc,CAAC,KAAK,EAAE,EAAE,GAAG,OAAO,EAAE,EAAE,CAAC;QACjD,SAAS,EAAE,cAAc,CAAC,KAAK,EAAE,GAAG,GAAG,OAAO,EAAE,EAAE,CAAC;KACpD,CAAC;AACJ,CAAC;AAED,SAAS,cAAc,CACrB,KAA0B,EAC1B,SAAiB,EACjB,WAAmB;IAEnB,IAAI,WAAW,GAAG,CAAC;QAAE,OAAO,CAAC,CAAC;IAC9B,MAAM,WAAW,GAAG,KAAK,CAAC,WAAW,GAAG,SAAS,CAAC;IAClD,wEAAwE;IACxE,2EAA2E;IAC3E,qBAAqB;IACrB,MAAM,eAAe,GAAG,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,UAAU,IAAI,WAAW,CAAC,CAAC;IACjF,IAAI,eAAe,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,CAAC,CAAC;IAE3C,MAAM,aAAa,GAAG,CAAC,GAAG,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,UAAU,GAAG,CAAC,CAAC,UAAU,CAAC,CAAC;IACrF,MAAM,WAAW,GAAa,EAAE,CAAC;IACjC,0EAA0E;IAC1E,KAAK,IAAI,CAAC,GAAG,WAAW,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;QAC7C,MAAM,OAAO,GAAG,KAAK,CAAC,WAAW,GAAG,CAAC,GAAG,QAAQ,CAAC;QACjD,IAAI,OAAO,GAAG,WAAW;YAAE,SAAS;QACpC,WAAW,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;IAC5B,CAAC;IACD,IAAI,WAAW,CAAC,MAAM,GAAG,CAAC;QAAE,OAAO,CAAC,CAAC;IAErC,wEAAwE;IACxE,8DAA8D;IAC9D,MAAM,oBAAoB,GAAG,IAAI,GAAG,EAAoB,CAAC;IACzD,MAAM,UAAU,GAAG,IAAI,GAAG,EAAkB,CAAC;IAC7C,IAAI,SAAS,GAAG,CAAC,CAAC;IAElB,KAAK,MAAM,OAAO,IAAI,WAAW,EAAE,CAAC;QAClC,OAAO,SAAS,GAAG,aAAa,CAAC,MAAM,EAAE,CAAC;YACxC,MAAM,MAAM,GAAG,aAAa,CAAC,SAAS,CAAC,CAAC;YACxC,IAAI,MAAM,KAAK,SAAS;gBAAE,MAAM;YAChC,IAAI,MAAM,CAAC,UAAU,GAAG,OAAO;gBAAE,MAAM;YACvC,KAAK,MAAM,CAAC,KAAK,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,aAAa,EAAE,CAAC;gBAClD,UAAU,CAAC,GAAG,CAAC,KAAK,EAAE,CAAC,UAAU,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC;YAC9D,CAAC;YACD,SAAS,IAAI,CAAC,CAAC;QACjB,CAAC;QACD,IAAI,KAAK,GAAG,CAAC,CAAC;QACd,KAAK,MAAM,CAAC,IAAI,UAAU,CAAC,MAAM,EAAE;YAAE,KAAK,IAAI,CAAC,CAAC;QAChD,IAAI,KAAK,KAAK,CAAC;YAAE,SAAS;QAC1B,MAAM,IAAI,GAAG,CAAC,GAAG,UAAU,CAAC,OAAO,EAAE,CAAC;aACnC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE;YACb,IAAI,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;gBAAE,OAAO,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;YACtC,OAAO,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;QAChD,CAAC,CAAC;aACD,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;QACf,KAAK,MAAM,CAAC,KAAK,EAAE,KAAK,CAAC,IAAI,IAAI,EAAE,CAAC;YAClC,MAAM,KAAK,GAAG,KAAK,GAAG,KAAK,CAAC;YAC5B,MAAM,GAAG,GAAG,oBAAoB,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;YAC5C,IAAI,GAAG,KAAK,SAAS;gBAAE,oBAAoB,CAAC,GAAG,CAAC,KAAK,EAAE,CAAC,KAAK,CAAC,CAAC,CAAC;;gBAC3D,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QACvB,CAAC;IACH,CAAC;IACD,+DAA+D;IAC/D,MAAM,IAAI,GAAa,EAAE,CAAC;IAC1B,KAAK,MAAM,GAAG,IAAI,oBAAoB,CAAC,MAAM,EAAE,EAAE,CAAC;QAChD,KAAK,MAAM,CAAC,IAAI,GAAG;YAAE,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IACpC,CAAC;IACD,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC;QAAE,OAAO,CAAC,CAAC;IAC9B,OAAO,MAAM,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC;AAC9B,CAAC;AAED,SAAS,MAAM,CAAC,MAAyB;IACvC,MAAM,CAAC,GAAG,MAAM,CAAC,MAAM,CAAC;IACxB,IAAI,CAAC,GAAG,CAAC;QAAE,OAAO,CAAC,CAAC;IACpB,IAAI,IAAI,GAAG,CAAC,CAAC;IACb,KAAK,MAAM,CAAC,IAAI,MAAM;QAAE,IAAI,IAAI,CAAC,CAAC;IAClC,IAAI,IAAI,CAAC,CAAC;IACV,IAAI,QAAQ,GAAG,CAAC,CAAC;IACjB,KAAK,MAAM,CAAC,IAAI,MAAM,EAAE,CAAC;QACvB,MAAM,CAAC,GAAG,CAAC,GAAG,IAAI,CAAC;QACnB,QAAQ,IAAI,CAAC,GAAG,CAAC,CAAC;IACpB,CAAC;IACD,QAAQ,IAAI,CAAC,CAAC,CAAC,sDAAsD;IACrE,OAAO,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;AAC7B,CAAC;AAED,SAAS,MAAM,CAAC,CAAS;IACvB,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAC;QAAE,OAAO,CAAC,CAAC;IAClC,OAAO,IAAI,CAAC,KAAK,CAAC,CAAC,GAAG,MAAM,CAAC,GAAG,MAAM,CAAC;AACzC,CAAC"}

package/dist/pipeline/ownership-helpers/gini-community.d.ts

@@ -0,0 +1,24 @@
+/**
+ * Community-level truck-factor aggregation.
+ *
+ * Strategy: union every member File's per-contributor line count into a single
+ * vector, then compute the corrected Gini → bus-factor mapping the temporal
+ * phase already uses for File nodes. This collapses "how many people does the
+ * community depend on" into a single integer without the simple-min failure
+ * mode (one single-owner test file would otherwise tank an entire community).
+ *
+ * The vector is the sum of each contributor's lines across all member files,
+ * so a contributor who owns 50 lines in file A and 30 lines in file B
+ * contributes 80 lines to the community vector.
+ */
+import type { ContributorWeight } from "./line-overlap.js";
+export interface CommunityTruckFactorInput {
+    /** One entry per member File — the contributor line-share vector for that file. */
+    readonly memberFiles: ReadonlyArray<readonly ContributorWeight[]>;
+}
+/**
+ * Compute the community truck factor. Empty input → 1 (a community with no
+ * ownership signals is maximally concentrated by default).
+ */
+export declare function communityTruckFactor(input: CommunityTruckFactorInput): number;
+//# sourceMappingURL=gini-community.d.ts.map

package/dist/pipeline/ownership-helpers/gini-community.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"gini-community.d.ts","sourceRoot":"","sources":["../../../src/pipeline/ownership-helpers/gini-community.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;GAYG;AAGH,OAAO,KAAK,EAAE,iBAAiB,EAAE,MAAM,mBAAmB,CAAC;AAE3D,MAAM,WAAW,yBAAyB;IACxC,mFAAmF;IACnF,QAAQ,CAAC,WAAW,EAAE,aAAa,CAAC,SAAS,iBAAiB,EAAE,CAAC,CAAC;CACnE;AAED;;;GAGG;AACH,wBAAgB,oBAAoB,CAAC,KAAK,EAAE,yBAAyB,GAAG,MAAM,CAU7E"}

package/dist/pipeline/ownership-helpers/gini-community.js

@@ -0,0 +1,32 @@
+/**
+ * Community-level truck-factor aggregation.
+ *
+ * Strategy: union every member File's per-contributor line count into a single
+ * vector, then compute the corrected Gini → bus-factor mapping the temporal
+ * phase already uses for File nodes. This collapses "how many people does the
+ * community depend on" into a single integer without the simple-min failure
+ * mode (one single-owner test file would otherwise tank an entire community).
+ *
+ * The vector is the sum of each contributor's lines across all member files,
+ * so a contributor who owns 50 lines in file A and 30 lines in file B
+ * contributes 80 lines to the community vector.
+ */
+import { busFactor } from "../phases/temporal-helpers/gini.js";
+/**
+ * Compute the community truck factor. Empty input → 1 (a community with no
+ * ownership signals is maximally concentrated by default).
+ */
+export function communityTruckFactor(input) {
+    const totals = new Map();
+    for (const fileContribs of input.memberFiles) {
+        for (const contrib of fileContribs) {
+            if (contrib.lines <= 0)
+                continue;
+            totals.set(contrib.email, (totals.get(contrib.email) ?? 0) + contrib.lines);
+        }
+    }
+    if (totals.size === 0)
+        return 1;
+    return busFactor([...totals.values()]);
+}
+//# sourceMappingURL=gini-community.js.map

package/dist/pipeline/ownership-helpers/gini-community.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"gini-community.js","sourceRoot":"","sources":["../../../src/pipeline/ownership-helpers/gini-community.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;GAYG;AAEH,OAAO,EAAE,SAAS,EAAE,MAAM,oCAAoC,CAAC;AAQ/D;;;GAGG;AACH,MAAM,UAAU,oBAAoB,CAAC,KAAgC;IACnE,MAAM,MAAM,GAAG,IAAI,GAAG,EAAkB,CAAC;IACzC,KAAK,MAAM,YAAY,IAAI,KAAK,CAAC,WAAW,EAAE,CAAC;QAC7C,KAAK,MAAM,OAAO,IAAI,YAAY,EAAE,CAAC;YACnC,IAAI,OAAO,CAAC,KAAK,IAAI,CAAC;gBAAE,SAAS;YACjC,MAAM,CAAC,GAAG,CAAC,OAAO,CAAC,KAAK,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC;QAC9E,CAAC;IACH,CAAC;IACD,IAAI,MAAM,CAAC,IAAI,KAAK,CAAC;QAAE,OAAO,CAAC,CAAC;IAChC,OAAO,SAAS,CAAC,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC;AACzC,CAAC"}

package/dist/pipeline/ownership-helpers/git-blame-batcher.d.ts

@@ -0,0 +1,71 @@
+/**
+ * Batch-blame helper for the ownership phase.
+ *
+ * `git` exposes no native multi-file blame command, so we parallelise per-file
+ * `git blame --porcelain` invocations behind a bounded concurrency pool. Two
+ * one-time set-up steps precede the fleet to minimise per-file cost:
+ *   1. `git commit-graph write --reachable --changed-paths` materialises Bloom
+ *      filters the blame engine uses to skip unrelated commits (2-6× speedup).
+ *   2. We resolve parallelism to `max(2, availableParallelism - 2)` capped at
+ *      32 — past that, I/O contention dominates.
+ *
+ * The porcelain format is deterministic line-by-line:
+ *   <commit-sha> <orig-line> <final-line> [<group-size>]
+ *   author <name>
+ *   author-mail <<email>>
+ *   ... other headers, terminated by a `\t<literal source line>` marker.
+ *
+ * A commit-sha repeats its headers only on the first line of each group, so a
+ * cache of sha → author metadata short-circuits repeat headers. Output is a
+ * deterministic `Map<relPath, LineOwner[]>`; the map preserves caller-provided
+ * insertion order but each `LineOwner[]` is ordered by line number.
+ */
+/**
+ * Per-line attribution emitted by `git blame --porcelain`. The email is
+ * lowercased at parse time for consistent hashing downstream.
+ */
+export interface LineOwner {
+    readonly line: number;
+    readonly email: string;
+    readonly authorName: string;
+    readonly sha: string;
+}
+export interface BatchBlameResult {
+    /** Per-file line-by-line attribution map. */
+    readonly byFile: ReadonlyMap<string, readonly LineOwner[]>;
+    /** Number of `git blame` subprocesses spawned (excludes the commit-graph warm-up). */
+    readonly subprocessCount: number;
+    /** Number of files skipped because blame failed (never-committed, binary, etc). */
+    readonly skippedCount: number;
+}
+export interface BatchBlameOptions {
+    /** Override concurrency; defaults to `max(2, availableParallelism - 2)` capped at 32. */
+    readonly concurrency?: number;
+    /**
+     * If `false`, skip the `git commit-graph write` warm-up. Useful when callers
+     * know the graph is already current or when running against a shallow clone
+     * where `--reachable` would fail.
+     */
+    readonly warmCommitGraph?: boolean;
+    /** Optional sink for per-file warnings. */
+    readonly onWarn?: (relPath: string, message: string) => void;
+}
+/**
+ * Run `git blame --porcelain` against each `relPaths` entry in parallel and
+ * return the aggregated attribution map.
+ *
+ * Determinism: outputs are inserted into the map in sorted `relPath` order
+ * regardless of completion order, and each per-file `LineOwner[]` is sorted
+ * by line number.
+ */
+export declare function batchBlame(repoPath: string, relPaths: readonly string[], opts?: BatchBlameOptions): Promise<BatchBlameResult>;
+/**
+ * Parse `git blame --porcelain` output into per-line owners.
+ *
+ * The format groups consecutive lines attributed to the same commit; only the
+ * first line of each group includes the `author` / `author-mail` headers. We
+ * cache per-sha author metadata across groups so later groups can reuse the
+ * first group's headers.
+ */
+export declare function parsePorcelainBlame(stdout: string): LineOwner[];
+//# sourceMappingURL=git-blame-batcher.d.ts.map

package/dist/pipeline/ownership-helpers/git-blame-batcher.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"git-blame-batcher.d.ts","sourceRoot":"","sources":["../../../src/pipeline/ownership-helpers/git-blame-batcher.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;GAqBG;AAWH;;;GAGG;AACH,MAAM,WAAW,SAAS;IACxB,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;IACvB,QAAQ,CAAC,UAAU,EAAE,MAAM,CAAC;IAC5B,QAAQ,CAAC,GAAG,EAAE,MAAM,CAAC;CACtB;AAED,MAAM,WAAW,gBAAgB;IAC/B,6CAA6C;IAC7C,QAAQ,CAAC,MAAM,EAAE,WAAW,CAAC,MAAM,EAAE,SAAS,SAAS,EAAE,CAAC,CAAC;IAC3D,sFAAsF;IACtF,QAAQ,CAAC,eAAe,EAAE,MAAM,CAAC;IACjC,mFAAmF;IACnF,QAAQ,CAAC,YAAY,EAAE,MAAM,CAAC;CAC/B;AAED,MAAM,WAAW,iBAAiB;IAChC,yFAAyF;IACzF,QAAQ,CAAC,WAAW,CAAC,EAAE,MAAM,CAAC;IAC9B;;;;OAIG;IACH,QAAQ,CAAC,eAAe,CAAC,EAAE,OAAO,CAAC;IACnC,2CAA2C;IAC3C,QAAQ,CAAC,MAAM,CAAC,EAAE,CAAC,OAAO,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,KAAK,IAAI,CAAC;CAC9D;AAED;;;;;;;GAOG;AACH,wBAAsB,UAAU,CAC9B,QAAQ,EAAE,MAAM,EAChB,QAAQ,EAAE,SAAS,MAAM,EAAE,EAC3B,IAAI,GAAE,iBAAsB,GAC3B,OAAO,CAAC,gBAAgB,CAAC,CA8D3B;AAYD;;;;;;;GAOG;AACH,wBAAgB,mBAAmB,CAAC,MAAM,EAAE,MAAM,GAAG,SAAS,EAAE,CAiD/D"}