homegraph 1.2.0 → 1.5.0

package/dist/spec/build/git-scanner.js

@@ -0,0 +1,254 @@
+"use strict";
+/**
+ * Git commit history scanning and spec↔commit pairing.
+ *
+ * Replaces `commit4spec/reverse_engineer/git_scanner.py` and
+ * `commit4spec/reverse_engineer/git_utils.py`. Discovers spec↔commit pairs via
+ * conventional-commit scope extraction: extracts scopes from commit messages
+ * and resolves them to spec IDs on disk.
+ *
+ * All Git operations use `execFileSync` with an args array (never template
+ * strings) to avoid shell injection risks — matching the HomeGraph convention
+ * in `src/sync/git-hooks.ts` and `src/sync/worktree.ts`.
+ *
+ * @module spec/build/git-scanner
+ */
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.isGitRepo = isGitRepo;
+exports.getAllCommits = getAllCommits;
+exports.getCommitInfo = getCommitInfo;
+exports.getCommitDiff = getCommitDiff;
+exports.getCommitRange = getCommitRange;
+exports.scan = scan;
+const child_process_1 = require("child_process");
+const scope_resolver_1 = require("./scope-resolver");
+const spec_extractor_1 = require("./spec-extractor");
+const errors_1 = require("../../errors");
+const git_utils_1 = require("../git-utils");
+// ---------------------------------------------------------------------------
+// Helpers
+// ---------------------------------------------------------------------------
+/**
+ * Convert an ISO-8601 string to Unix epoch milliseconds.
+ * Returns 0 for invalid or unparseable strings.
+ */
+function parseISOTimestamp(isoString) {
+    const ts = new Date(isoString).getTime();
+    return isNaN(ts) ? 0 : ts;
+}
+// ---------------------------------------------------------------------------
+// isGitRepo
+// ---------------------------------------------------------------------------
+/**
+ * Check whether `repoPath` is a git repository.
+ *
+ * Runs `git rev-parse --git-dir` and returns `true` on success, `false`
+ * when the command fails (not a repo, git not installed, etc.).
+ */
+function isGitRepo(repoPath) {
+    try {
+        (0, child_process_1.execFileSync)('git', ['rev-parse', '--git-dir'], (0, git_utils_1.gitExecOptions)(repoPath));
+        return true;
+    }
+    catch {
+        return false;
+    }
+}
+// ---------------------------------------------------------------------------
+// getAllCommits
+// ---------------------------------------------------------------------------
+/**
+ * Return every commit in `repoPath` as `CommitInfo` objects.
+ *
+ * Runs `git log --format='%H%n%aI%n%an%n%s%n---END---'` and parses the
+ * output one five-line block per commit.  Commits are returned in reverse
+ * chronological order (most recent first), matching Git's default.
+ *
+ * An empty repo yields an empty array.
+ */
+function getAllCommits(repoPath) {
+    let stdout;
+    try {
+        stdout = (0, child_process_1.execFileSync)('git', ['log', "--format=%H%x00%aI%x00%an%x00%s%x00"], (0, git_utils_1.gitExecOptions)(repoPath));
+    }
+    catch {
+        return [];
+    }
+    const parts = stdout.replace(/\0+$/, '').trim().split('\0');
+    const commits = [];
+    // Each commit occupies 4 NUL-separated fields: hash, timestamp, author, message
+    for (let i = 0; i + 3 < parts.length; i += 4) {
+        const hash = parts[i].trim();
+        const isoString = parts[i + 1].trim();
+        const author = parts[i + 2].trim();
+        const message = parts[i + 3].trim();
+        if (!hash)
+            continue;
+        commits.push({
+            hash,
+            message,
+            author,
+            timestamp: parseISOTimestamp(isoString),
+        });
+    }
+    return commits;
+}
+// ---------------------------------------------------------------------------
+// getCommitInfo
+// ---------------------------------------------------------------------------
+/**
+ * Retrieve metadata for a single commit.
+ *
+ * Runs `git show -s --format='%H%n%aI%n%an%n%s' <commitHash>`.  Returns
+ * `null` when the command fails (hash not found, invalid repo, etc.).
+ */
+function getCommitInfo(repoPath, commitHash) {
+    let stdout;
+    try {
+        stdout = (0, child_process_1.execFileSync)('git', ['show', '-s', "--format=%H%n%aI%n%an%n%s", commitHash], (0, git_utils_1.gitExecOptions)(repoPath));
+    }
+    catch {
+        return null;
+    }
+    const lines = stdout.trim().split('\n');
+    if (lines.length < 4)
+        return null;
+    const hash = lines[0].trim();
+    const isoString = lines[1].trim();
+    const author = lines[2].trim();
+    const message = lines[3].trim();
+    if (!hash)
+        return null;
+    return {
+        hash,
+        message,
+        author,
+        timestamp: parseISOTimestamp(isoString),
+    };
+}
+// ---------------------------------------------------------------------------
+// getCommitDiff
+// ---------------------------------------------------------------------------
+/**
+ * Return the full unified diff introduced by `commitHash`.
+ *
+ * Strategy:
+ * 1. Query the parent hash(es) via `git log --pretty=%P -n 1 <commitHash>`.
+ * 2. If a parent exists, diff it against the commit: `git diff <parent> <commitHash>`.
+ * 3. For an initial (root) commit with no parent, use `git diff <commitHash>`
+ *    which shows all files added in that commit.
+ *
+ * Returns an empty string when any Git command fails.
+ */
+function getCommitDiff(repoPath, commitHash) {
+    // Step 1 — get parent hashes.
+    let parentStdout;
+    try {
+        parentStdout = (0, child_process_1.execFileSync)('git', ['log', '--pretty=%P', '-n', '1', commitHash], (0, git_utils_1.gitExecOptions)(repoPath));
+    }
+    catch {
+        return '';
+    }
+    const parents = parentStdout.trim().split(/\s+/).filter(Boolean);
+    // Step 2 & 3 — run the appropriate diff.
+    try {
+        if (parents.length > 0) {
+            const diffs = [];
+            for (const parent of parents) {
+                try {
+                    diffs.push((0, child_process_1.execFileSync)('git', ['diff', parent, commitHash], (0, git_utils_1.gitExecOptions)(repoPath)));
+                }
+                catch {
+                    // Skip diffs that fail for an individual parent
+                }
+            }
+            return diffs.join('\n');
+        }
+        // Initial commit — no parent.
+        return (0, child_process_1.execFileSync)('git', ['diff', commitHash], (0, git_utils_1.gitExecOptions)(repoPath));
+    }
+    catch {
+        return '';
+    }
+}
+// ---------------------------------------------------------------------------
+// getCommitRange
+// ---------------------------------------------------------------------------
+/**
+ * Return every commit between `fromHash` (exclusive) and `toHash`
+ * (inclusive) in chronological order (oldest first).
+ *
+ * Runs `git log --reverse --format="..." fromHash..toHash`. Returns an empty
+ * array when either hash is not found or the range is empty.
+ *
+ * Callers should validate that `fromHash` is an ancestor of `toHash` before
+ * calling this function (use `git merge-base --is-ancestor`).  Results are
+ * undefined when the two hashes are on unrelated branches.
+ */
+function getCommitRange(repoPath, fromHash, toHash) {
+    let stdout;
+    try {
+        stdout = (0, child_process_1.execFileSync)('git', ['log', '--no-merges', '--reverse', `--format=%H%x00%aI%x00%an%x00%s%x00`, `${fromHash}..${toHash}`], (0, git_utils_1.gitExecOptions)(repoPath));
+    }
+    catch {
+        return [];
+    }
+    const parts = stdout.replace(/\0+$/, '').trim().split('\0');
+    const commits = [];
+    // Each commit occupies 4 NUL-separated fields: hash, timestamp, author, message
+    for (let i = 0; i + 3 < parts.length; i += 4) {
+        const hash = parts[i].trim();
+        const isoString = parts[i + 1].trim();
+        const author = parts[i + 2].trim();
+        const message = parts[i + 3].trim();
+        if (!hash)
+            continue;
+        commits.push({
+            hash,
+            message,
+            author,
+            timestamp: parseISOTimestamp(isoString),
+        });
+    }
+    return commits;
+}
+// ---------------------------------------------------------------------------
+// scan
+// ---------------------------------------------------------------------------
+/**
+ * Main scanning function — discovers spec↔commit pairs from commit history.
+ *
+ * Iterates every commit in the repo, extracts a conventional-commit scope from
+ * its message, normalizes it, and checks whether the resulting spec ID exists
+ * on disk. When a match is found the pair is recorded with full spec and
+ * commit metadata.
+ *
+ * Duplicates (identical specId + commitHash) are de-duplicated via a
+ * `Map` keyed on `"${specId}|${commitHash}"`.
+ */
+function scan(repoPath, specStoragePath, config) {
+    const pairs = new Map();
+    const commits = getAllCommits(repoPath);
+    for (const commit of commits) {
+        const specId = (0, scope_resolver_1.resolveScopeToSpec)(commit.message, specStoragePath, config);
+        if (!specId) {
+            (0, errors_1.logDebug)('scan: no scope match', {
+                commitHash: commit.hash.slice(0, 7),
+                message: commit.message,
+            });
+            continue;
+        }
+        const specMetadata = (0, spec_extractor_1.extractSpecMetadata)(specStoragePath, specId, config);
+        const key = `${specId}|${commit.hash}`;
+        if (!pairs.has(key)) {
+            pairs.set(key, {
+                specId,
+                commitHash: commit.hash,
+                specMetadata,
+                commitMetadata: commit,
+            });
+        }
+    }
+    return Array.from(pairs.values());
+}
+//# sourceMappingURL=git-scanner.js.map

package/dist/spec/build/git-scanner.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"git-scanner.js","sourceRoot":"","sources":["../../../src/spec/build/git-scanner.ts"],"names":[],"mappings":";AAAA;;;;;;;;;;;;;GAaG;;AAsDH,8BAOC;AAeD,sCAiCC;AAYD,sCA+BC;AAiBD,sCA4CC;AAiBD,wCAqCC;AAiBD,oBAiCC;AA3TD,iDAA6C;AAE7C,qDAAsD;AACtD,qDAAqE;AACrE,yCAAwC;AACxC,4CAA8C;AAwB9C,8EAA8E;AAC9E,UAAU;AACV,8EAA8E;AAE9E;;;GAGG;AACH,SAAS,iBAAiB,CAAC,SAAiB;IAC1C,MAAM,EAAE,GAAG,IAAI,IAAI,CAAC,SAAS,CAAC,CAAC,OAAO,EAAE,CAAC;IACzC,OAAO,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;AAC5B,CAAC;AAED,8EAA8E;AAC9E,YAAY;AACZ,8EAA8E;AAE9E;;;;;GAKG;AACH,SAAgB,SAAS,CAAC,QAAgB;IACxC,IAAI,CAAC;QACH,IAAA,4BAAY,EAAC,KAAK,EAAE,CAAC,WAAW,EAAE,WAAW,CAAC,EAAE,IAAA,0BAAc,EAAC,QAAQ,CAAC,CAAC,CAAC;QAC1E,OAAO,IAAI,CAAC;IACd,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,KAAK,CAAC;IACf,CAAC;AACH,CAAC;AAED,8EAA8E;AAC9E,gBAAgB;AAChB,8EAA8E;AAE9E;;;;;;;;GAQG;AACH,SAAgB,aAAa,CAAC,QAAgB;IAC5C,IAAI,MAAc,CAAC;IACnB,IAAI,CAAC;QACH,MAAM,GAAG,IAAA,4BAAY,EACnB,KAAK,EACL,CAAC,KAAK,EAAE,qCAAqC,CAAC,EAC9C,IAAA,0BAAc,EAAC,QAAQ,CAAC,CACzB,CAAC;IACJ,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,MAAM,KAAK,GAAG,MAAM,CAAC,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAC5D,MAAM,OAAO,GAAiB,EAAE,CAAC;IAEjC,gFAAgF;IAChF,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;QAC7C,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC;QAC9B,MAAM,SAAS,GAAG,KAAK,CAAC,CAAC,GAAG,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC;QACvC,MAAM,MAAM,GAAG,KAAK,CAAC,CAAC,GAAG,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC;QACpC,MAAM,OAAO,GAAG,KAAK,CAAC,CAAC,GAAG,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC;QAErC,IAAI,CAAC,IAAI;YAAE,SAAS;QAEpB,OAAO,CAAC,IAAI,CAAC;YACX,IAAI;YACJ,OAAO;YACP,MAAM;YACN,SAAS,EAAE,iBAAiB,CAAC,SAAS,CAAC;SACxC,CAAC,CAAC;IACL,CAAC;IAED,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,8EAA8E;AAC9E,gBAAgB;AAChB,8EAA8E;AAE9E;;;;;GAKG;AACH,SAAgB,aAAa,CAC3B,QAAgB,EAChB,UAAkB;IAElB,IAAI,MAAc,CAAC;IACnB,IAAI,CAAC;QACH,MAAM,GAAG,IAAA,4BAAY,EACnB,KAAK,EACL,CAAC,MAAM,EAAE,IAAI,EAAE,2BAA2B,EAAE,UAAU,CAAC,EACvD,IAAA,0BAAc,EAAC,QAAQ,CAAC,CACzB,CAAC;IACJ,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,IAAI,CAAC;IACd,CAAC;IAED,MAAM,KAAK,GAAG,MAAM,CAAC,IAAI,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IACxC,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC;QAAE,OAAO,IAAI,CAAC;IAElC,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC;IAC9B,MAAM,SAAS,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC;IACnC,MAAM,MAAM,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC;IAChC,MAAM,OAAO,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC;IAEjC,IAAI,CAAC,IAAI;QAAE,OAAO,IAAI,CAAC;IAEvB,OAAO;QACL,IAAI;QACJ,OAAO;QACP,MAAM;QACN,SAAS,EAAE,iBAAiB,CAAC,SAAS,CAAC;KACxC,CAAC;AACJ,CAAC;AAED,8EAA8E;AAC9E,gBAAgB;AAChB,8EAA8E;AAE9E;;;;;;;;;;GAUG;AACH,SAAgB,aAAa,CAAC,QAAgB,EAAE,UAAkB;IAChE,8BAA8B;IAC9B,IAAI,YAAoB,CAAC;IACzB,IAAI,CAAC;QACH,YAAY,GAAG,IAAA,4BAAY,EACzB,KAAK,EACL,CAAC,KAAK,EAAE,aAAa,EAAE,IAAI,EAAE,GAAG,EAAE,UAAU,CAAC,EAC7C,IAAA,0BAAc,EAAC,QAAQ,CAAC,CACzB,CAAC;IACJ,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,MAAM,OAAO,GAAG,YAAY,CAAC,IAAI,EAAE,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;IAEjE,yCAAyC;IACzC,IAAI,CAAC;QACH,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACvB,MAAM,KAAK,GAAa,EAAE,CAAC;YAC3B,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;gBAC7B,IAAI,CAAC;oBACH,KAAK,CAAC,IAAI,CACR,IAAA,4BAAY,EACV,KAAK,EACL,CAAC,MAAM,EAAE,MAAM,EAAE,UAAU,CAAC,EAC5B,IAAA,0BAAc,EAAC,QAAQ,CAAC,CACzB,CACF,CAAC;gBACJ,CAAC;gBAAC,MAAM,CAAC;oBACP,gDAAgD;gBAClD,CAAC;YACH,CAAC;YACD,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAC1B,CAAC;QAED,8BAA8B;QAC9B,OAAO,IAAA,4BAAY,EACjB,KAAK,EACL,CAAC,MAAM,EAAE,UAAU,CAAC,EACpB,IAAA,0BAAc,EAAC,QAAQ,CAAC,CACzB,CAAC;IACJ,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,EAAE,CAAC;IACZ,CAAC;AACH,CAAC;AAED,8EAA8E;AAC9E,iBAAiB;AACjB,8EAA8E;AAE9E;;;;;;;;;;GAUG;AACH,SAAgB,cAAc,CAC5B,QAAgB,EAChB,QAAgB,EAChB,MAAc;IAEd,IAAI,MAAc,CAAC;IACnB,IAAI,CAAC;QACH,MAAM,GAAG,IAAA,4BAAY,EACnB,KAAK,EACL,CAAC,KAAK,EAAE,aAAa,EAAE,WAAW,EAAE,qCAAqC,EAAE,GAAG,QAAQ,KAAK,MAAM,EAAE,CAAC,EACpG,IAAA,0BAAc,EAAC,QAAQ,CAAC,CACzB,CAAC;IACJ,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,MAAM,KAAK,GAAG,MAAM,CAAC,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAC5D,MAAM,OAAO,GAAiB,EAAE,CAAC;IAEjC,gFAAgF;IAChF,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;QAC7C,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC;QAC9B,MAAM,SAAS,GAAG,KAAK,CAAC,CAAC,GAAG,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC;QACvC,MAAM,MAAM,GAAG,KAAK,CAAC,CAAC,GAAG,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC;QACpC,MAAM,OAAO,GAAG,KAAK,CAAC,CAAC,GAAG,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC;QAErC,IAAI,CAAC,IAAI;YAAE,SAAS;QAEpB,OAAO,CAAC,IAAI,CAAC;YACX,IAAI;YACJ,OAAO;YACP,MAAM;YACN,SAAS,EAAE,iBAAiB,CAAC,SAAS,CAAC;SACxC,CAAC,CAAC;IACL,CAAC;IAED,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,8EAA8E;AAC9E,OAAO;AACP,8EAA8E;AAE9E;;;;;;;;;;GAUG;AACH,SAAgB,IAAI,CAClB,QAAgB,EAChB,eAAuB,EACvB,MAAkB;IAElB,MAAM,KAAK,GAAG,IAAI,GAAG,EAA0B,CAAC;IAEhD,MAAM,OAAO,GAAG,aAAa,CAAC,QAAQ,CAAC,CAAC;IAExC,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC7B,MAAM,MAAM,GAAG,IAAA,mCAAkB,EAAC,MAAM,CAAC,OAAO,EAAE,eAAe,EAAE,MAAM,CAAC,CAAC;QAC3E,IAAI,CAAC,MAAM,EAAE,CAAC;YACZ,IAAA,iBAAQ,EAAC,sBAAsB,EAAE;gBAC/B,UAAU,EAAE,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC;gBACnC,OAAO,EAAE,MAAM,CAAC,OAAO;aACxB,CAAC,CAAC;YACH,SAAS;QACX,CAAC;QAED,MAAM,YAAY,GAAG,IAAA,oCAAmB,EAAC,eAAe,EAAE,MAAM,EAAE,MAAM,CAAC,CAAC;QAE1E,MAAM,GAAG,GAAG,GAAG,MAAM,IAAI,MAAM,CAAC,IAAI,EAAE,CAAC;QACvC,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC;YACpB,KAAK,CAAC,GAAG,CAAC,GAAG,EAAE;gBACb,MAAM;gBACN,UAAU,EAAE,MAAM,CAAC,IAAI;gBACvB,YAAY;gBACZ,cAAc,EAAE,MAAM;aACvB,CAAC,CAAC;QACL,CAAC;IACH,CAAC;IAED,OAAO,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,EAAE,CAAC,CAAC;AACpC,CAAC"}

package/dist/spec/build/pipeline.d.ts

@@ -0,0 +1,54 @@
+/**
+ * Build pipeline orchestrator — main entry point for building spec
+ * knowledge from Git history.
+ *
+ * Replaces `commit4spec/reverse_engineer/pipeline.py`.  Discovers spec↔commit
+ * pairs via `scan()`, persists them into the SQLite knowledge graph, and
+ * writes a `.homegraph/commit4spec/meta.json` anchor so the self-evolve pipeline can
+ * pick up where build left off.
+ *
+ * @module spec/build/pipeline
+ */
+import { SqliteDatabase } from '../../db/sqlite-adapter';
+import type { ProgressCallback } from '../ui';
+export interface BuildResult {
+    specsFound: number;
+    commitsFound: number;
+    fragmentsFound: number;
+    relationsCreated: number;
+    totalEntries: number;
+    skippedEntries: Array<{
+        specId: string;
+        reason: string;
+    }>;
+    upToDate?: boolean;
+}
+/**
+ * Run the full build pipeline.
+ *
+ * Steps (ported from `pipeline.py:49-151`):
+ *
+ * 1. Initialise the spec knowledge-graph schema in the database.
+ * 2. Discover all specs on disk (for the `totalEntries` count).
+ * 3. Scan the repository for spec↔commit pairs.
+ * 4. Sort pairs by commit timestamp ascending so that `INSERT OR REPLACE`
+ *    gives the most-recent commit the final word for each spec node.
+ * 5. Process each pair:
+ *    - Insert a SpecNode (once per specId; deduplicated by commit timestamp).
+ *    - Insert a CommitNode (once per commit hash).
+ *    - Insert a SUMMARIZED_FROM relation.
+ *    - Parse the commit diff into code fragments and persist them.
+ * 6. Build a list of specs discovered on disk but never matched to a commit.
+ * 7. Write `.homegraph/commit4spec/meta.json`.
+ * 8. Return the `BuildResult` with all counts.
+ *
+ * Edge cases:
+ * - Empty pairs → zero counts, write meta anyway, return result.
+ * - Pair with commit metadata but no spec metadata → commit and fragments
+ *   are still persisted; SpecNode and spec_commit_relation are skipped.
+ * - Pair with spec metadata but no commit metadata → entire pair is skipped.
+ * - Same spec in multiple pairs → only counted once in `specsFound`, but
+ *   every commit relation is created.
+ */
+export declare function runBuildPipeline(repoPath: string, specStoragePath: string, db: SqliteDatabase, onProgress?: ProgressCallback): BuildResult;
+//# sourceMappingURL=pipeline.d.ts.map

package/dist/spec/build/pipeline.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"pipeline.d.ts","sourceRoot":"","sources":["../../../src/spec/build/pipeline.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;GAUG;AAEH,OAAO,EAAE,cAAc,EAAE,MAAM,yBAAyB,CAAC;AAWzD,OAAO,KAAK,EAAE,gBAAgB,EAAE,MAAM,OAAO,CAAC;AAM9C,MAAM,WAAW,WAAW;IAC1B,UAAU,EAAE,MAAM,CAAC;IACnB,YAAY,EAAE,MAAM,CAAC;IACrB,cAAc,EAAE,MAAM,CAAC;IACvB,gBAAgB,EAAE,MAAM,CAAC;IACzB,YAAY,EAAE,MAAM,CAAC;IACrB,cAAc,EAAE,KAAK,CAAC;QAAE,MAAM,EAAE,MAAM,CAAC;QAAC,MAAM,EAAE,MAAM,CAAA;KAAE,CAAC,CAAC;IAC1D,QAAQ,CAAC,EAAE,OAAO,CAAC;CACpB;AAMD;;;;;;;;;;;;;;;;;;;;;;;;;;GA0BG;AACH,wBAAgB,gBAAgB,CAC9B,QAAQ,EAAE,MAAM,EAChB,eAAe,EAAE,MAAM,EACvB,EAAE,EAAE,cAAc,EAClB,UAAU,CAAC,EAAE,gBAAgB,GAC5B,WAAW,CAoIb"}

package/dist/spec/build/pipeline.js

@@ -0,0 +1,170 @@
+"use strict";
+/**
+ * Build pipeline orchestrator — main entry point for building spec
+ * knowledge from Git history.
+ *
+ * Replaces `commit4spec/reverse_engineer/pipeline.py`.  Discovers spec↔commit
+ * pairs via `scan()`, persists them into the SQLite knowledge graph, and
+ * writes a `.homegraph/commit4spec/meta.json` anchor so the self-evolve pipeline can
+ * pick up where build left off.
+ *
+ * @module spec/build/pipeline
+ */
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.runBuildPipeline = runBuildPipeline;
+const config_1 = require("../config");
+const utils_1 = require("../utils");
+const schema_1 = require("../db/schema");
+const commit_node_1 = require("../db/commit-node");
+const relations_1 = require("../db/relations");
+const persist_1 = require("../db/persist");
+const scan_1 = require("./scan");
+const git_1 = require("../git");
+const diff_parser_1 = require("./diff-parser");
+const errors_1 = require("../../errors");
+// ---------------------------------------------------------------------------
+// runBuildPipeline
+// ---------------------------------------------------------------------------
+/**
+ * Run the full build pipeline.
+ *
+ * Steps (ported from `pipeline.py:49-151`):
+ *
+ * 1. Initialise the spec knowledge-graph schema in the database.
+ * 2. Discover all specs on disk (for the `totalEntries` count).
+ * 3. Scan the repository for spec↔commit pairs.
+ * 4. Sort pairs by commit timestamp ascending so that `INSERT OR REPLACE`
+ *    gives the most-recent commit the final word for each spec node.
+ * 5. Process each pair:
+ *    - Insert a SpecNode (once per specId; deduplicated by commit timestamp).
+ *    - Insert a CommitNode (once per commit hash).
+ *    - Insert a SUMMARIZED_FROM relation.
+ *    - Parse the commit diff into code fragments and persist them.
+ * 6. Build a list of specs discovered on disk but never matched to a commit.
+ * 7. Write `.homegraph/commit4spec/meta.json`.
+ * 8. Return the `BuildResult` with all counts.
+ *
+ * Edge cases:
+ * - Empty pairs → zero counts, write meta anyway, return result.
+ * - Pair with commit metadata but no spec metadata → commit and fragments
+ *   are still persisted; SpecNode and spec_commit_relation are skipped.
+ * - Pair with spec metadata but no commit metadata → entire pair is skipped.
+ * - Same spec in multiple pairs → only counted once in `specsFound`, but
+ *   every commit relation is created.
+ */
+function runBuildPipeline(repoPath, specStoragePath, db, onProgress) {
+    // Pre-check: if meta.json exists and HEAD hasn't changed since the last
+    // build, the knowledge graph is already current — skip the full rebuild.
+    const existingMeta = (0, utils_1.readMeta)(repoPath);
+    if (existingMeta?.currentCommitID) {
+        const headHash = (0, git_1.getHeadHash)(repoPath);
+        if (headHash && headHash === existingMeta.currentCommitID) {
+            (0, errors_1.logDebug)('Skipping spec build: knowledge graph is up to date', {
+                currentCommitID: existingMeta.currentCommitID,
+            });
+            return {
+                specsFound: 0,
+                commitsFound: 0,
+                fragmentsFound: 0,
+                relationsCreated: 0,
+                totalEntries: (0, utils_1.discoverSpecs)(specStoragePath).length,
+                skippedEntries: [],
+                upToDate: true,
+            };
+        }
+    }
+    // ---- Step 1: init schema ----
+    (0, schema_1.initSpecSchema)(db);
+    // ---- Step 2: discover specs (for totalEntries) ----
+    const allEntries = (0, utils_1.discoverSpecs)(specStoragePath);
+    const totalEntries = allEntries.length;
+    // ---- Step 3: scan for pairs ----
+    const pairs = (0, scan_1.scan)(repoPath, specStoragePath, (0, config_1.loadSpecConfig)(repoPath), onProgress);
+    // ---- Step 4: sort by commit timestamp ascending (oldest first) ----
+    // INSERT OR REPLACE is used for SpecNode — when the same spec appears
+    // in multiple pairs, the LAST write (most-recent commit) naturally wins.
+    pairs.sort((a, b) => {
+        const aTs = a.commitMetadata?.timestamp ?? 0;
+        const bTs = b.commitMetadata?.timestamp ?? 0;
+        return aTs - bTs;
+    });
+    // ---- Step 5: process pairs ----
+    const writtenSpecIds = new Set();
+    const seenCommitHashes = new Set();
+    let specsFound = 0;
+    let commitsFound = 0;
+    let fragmentsFound = 0;
+    let relationsCreated = 0;
+    const pairsTotal = pairs.length;
+    let pairsCurrent = 0;
+    for (const pair of pairs) {
+        pairsCurrent++;
+        onProgress?.({
+            phase: 'persisting',
+            current: pairsCurrent,
+            total: pairsTotal,
+            message: `${pair.specId} @ ${pair.commitHash.slice(0, 7)}`,
+        });
+        // Skip if no commit metadata (no data to insert).
+        const cm = pair.commitMetadata;
+        if (!cm) {
+            continue;
+        }
+        const sm = pair.specMetadata;
+        // Insert SpecNode if spec metadata is present.
+        // INSERT OR REPLACE lets the most-recent commit (last in ascending
+        // order) overwrite the timestamp. We still track writtenSpecIds to
+        // correctly count distinct specs (not overwrites).
+        if (sm) {
+            (0, errors_1.logDebug)('Processing pair: inserting SpecNode', {
+                specId: sm.specId,
+                title: sm.title,
+            });
+            (0, persist_1.upsertSpecFromMetadata)(db, sm, cm.timestamp);
+            if (!writtenSpecIds.has(pair.specId)) {
+                writtenSpecIds.add(pair.specId);
+                specsFound++;
+            }
+        }
+        // Insert CommitNode if not yet seen.
+        if (!seenCommitHashes.has(cm.hash)) {
+            (0, commit_node_1.insertCommitNode)(db, {
+                hash: cm.hash,
+                message: cm.message,
+                author: cm.author,
+                timestamp: cm.timestamp,
+            });
+            seenCommitHashes.add(cm.hash);
+            commitsFound++;
+        }
+        // Insert spec_commit_relation (only when spec metadata exists).
+        if (sm) {
+            (0, relations_1.insertSpecCommitRelation)(db, pair.specId, cm.hash, 'SUMMARIZED_FROM');
+            relationsCreated++;
+        }
+        // ---- Analyze diff and insert fragments ----
+        const preFetchedDiff = (0, git_1.getCommitDiff)(repoPath, cm.hash);
+        const fragments = (0, diff_parser_1.analyzeCommitDiff)(repoPath, cm.hash, preFetchedDiff);
+        const persisted = (0, persist_1.persistCommitFragments)(db, cm.hash, fragments);
+        fragmentsFound += persisted.fragmentsInserted;
+        relationsCreated += persisted.relationsCreated;
+    }
+    // ---- Step 6: build skipped entries ----
+    const skippedEntries = allEntries
+        .filter((e) => !writtenSpecIds.has(e.specId))
+        .map((e) => ({ specId: e.specId, reason: 'No matching commits found' }));
+    // ---- Step 7: write meta ----
+    const headHash = (0, git_1.getHeadHash)(repoPath) ?? undefined;
+    (0, utils_1.writeMeta)(repoPath, specStoragePath, headHash);
+    onProgress?.({ phase: 'done', current: 0, total: 0 });
+    // ---- Step 8: return result ----
+    return {
+        specsFound,
+        commitsFound,
+        fragmentsFound,
+        relationsCreated,
+        totalEntries,
+        skippedEntries,
+    };
+}
+//# sourceMappingURL=pipeline.js.map

package/dist/spec/build/pipeline.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"pipeline.js","sourceRoot":"","sources":["../../../src/spec/build/pipeline.ts"],"names":[],"mappings":";AAAA;;;;;;;;;;GAUG;;AA4DH,4CAyIC;AAlMD,sCAA2C;AAC3C,oCAA8D;AAC9D,yCAA8C;AAC9C,mDAAqD;AACrD,+CAA2D;AAC3D,2CAA+E;AAC/E,iCAA8B;AAC9B,gCAAoD;AACpD,+CAAkD;AAClD,yCAAwC;AAiBxC,8EAA8E;AAC9E,mBAAmB;AACnB,8EAA8E;AAE9E;;;;;;;;;;;;;;;;;;;;;;;;;;GA0BG;AACH,SAAgB,gBAAgB,CAC9B,QAAgB,EAChB,eAAuB,EACvB,EAAkB,EAClB,UAA6B;IAE7B,wEAAwE;IACxE,yEAAyE;IACzE,MAAM,YAAY,GAAG,IAAA,gBAAQ,EAAC,QAAQ,CAAC,CAAC;IACxC,IAAI,YAAY,EAAE,eAAe,EAAE,CAAC;QAClC,MAAM,QAAQ,GAAG,IAAA,iBAAW,EAAC,QAAQ,CAAC,CAAC;QACvC,IAAI,QAAQ,IAAI,QAAQ,KAAK,YAAY,CAAC,eAAe,EAAE,CAAC;YAC1D,IAAA,iBAAQ,EAAC,oDAAoD,EAAE;gBAC7D,eAAe,EAAE,YAAY,CAAC,eAAe;aAC9C,CAAC,CAAC;YACH,OAAO;gBACL,UAAU,EAAE,CAAC;gBACb,YAAY,EAAE,CAAC;gBACf,cAAc,EAAE,CAAC;gBACjB,gBAAgB,EAAE,CAAC;gBACnB,YAAY,EAAE,IAAA,qBAAa,EAAC,eAAe,CAAC,CAAC,MAAM;gBACnD,cAAc,EAAE,EAAE;gBAClB,QAAQ,EAAE,IAAI;aACf,CAAC;QACJ,CAAC;IACH,CAAC;IAED,gCAAgC;IAChC,IAAA,uBAAc,EAAC,EAAE,CAAC,CAAC;IAEnB,sDAAsD;IACtD,MAAM,UAAU,GAAG,IAAA,qBAAa,EAAC,eAAe,CAAC,CAAC;IAClD,MAAM,YAAY,GAAG,UAAU,CAAC,MAAM,CAAC;IAEvC,mCAAmC;IACnC,MAAM,KAAK,GAAG,IAAA,WAAI,EAAC,QAAQ,EAAE,eAAe,EAAE,IAAA,uBAAc,EAAC,QAAQ,CAAC,EAAE,UAAU,CAAC,CAAC;IAEpF,sEAAsE;IACtE,sEAAsE;IACtE,yEAAyE;IACzE,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE;QAClB,MAAM,GAAG,GAAG,CAAC,CAAC,cAAc,EAAE,SAAS,IAAI,CAAC,CAAC;QAC7C,MAAM,GAAG,GAAG,CAAC,CAAC,cAAc,EAAE,SAAS,IAAI,CAAC,CAAC;QAC7C,OAAO,GAAG,GAAG,GAAG,CAAC;IACnB,CAAC,CAAC,CAAC;IAEH,kCAAkC;IAClC,MAAM,cAAc,GAAG,IAAI,GAAG,EAAU,CAAC;IACzC,MAAM,gBAAgB,GAAG,IAAI,GAAG,EAAU,CAAC;IAC3C,IAAI,UAAU,GAAG,CAAC,CAAC;IACnB,IAAI,YAAY,GAAG,CAAC,CAAC;IACrB,IAAI,cAAc,GAAG,CAAC,CAAC;IACvB,IAAI,gBAAgB,GAAG,CAAC,CAAC;IAEzB,MAAM,UAAU,GAAG,KAAK,CAAC,MAAM,CAAC;IAChC,IAAI,YAAY,GAAG,CAAC,CAAC;IAErB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,YAAY,EAAE,CAAC;QACf,UAAU,EAAE,CAAC;YACX,KAAK,EAAE,YAAY;YACnB,OAAO,EAAE,YAAY;YACrB,KAAK,EAAE,UAAU;YACjB,OAAO,EAAE,GAAG,IAAI,CAAC,MAAM,MAAM,IAAI,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE;SAC3D,CAAC,CAAC;QAEH,kDAAkD;QAClD,MAAM,EAAE,GAAG,IAAI,CAAC,cAAc,CAAC;QAC/B,IAAI,CAAC,EAAE,EAAE,CAAC;YACR,SAAS;QACX,CAAC;QAED,MAAM,EAAE,GAAG,IAAI,CAAC,YAAY,CAAC;QAE7B,+CAA+C;QAC/C,mEAAmE;QACnE,mEAAmE;QACnE,mDAAmD;QACnD,IAAI,EAAE,EAAE,CAAC;YACP,IAAA,iBAAQ,EAAC,qCAAqC,EAAE;gBAC9C,MAAM,EAAE,EAAE,CAAC,MAAM;gBACjB,KAAK,EAAE,EAAE,CAAC,KAAK;aAChB,CAAC,CAAC;YAEH,IAAA,gCAAsB,EAAC,EAAE,EAAE,EAAE,EAAE,EAAE,CAAC,SAAS,CAAC,CAAC;YAC7C,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC;gBACrC,cAAc,CAAC,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;gBAChC,UAAU,EAAE,CAAC;YACf,CAAC;QACH,CAAC;QAED,qCAAqC;QACrC,IAAI,CAAC,gBAAgB,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;YACnC,IAAA,8BAAgB,EAAC,EAAE,EAAE;gBACnB,IAAI,EAAE,EAAE,CAAC,IAAI;gBACb,OAAO,EAAE,EAAE,CAAC,OAAO;gBACnB,MAAM,EAAE,EAAE,CAAC,MAAM;gBACjB,SAAS,EAAE,EAAE,CAAC,SAAS;aACxB,CAAC,CAAC;YACH,gBAAgB,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,CAAC;YAC9B,YAAY,EAAE,CAAC;QACjB,CAAC;QAED,gEAAgE;QAChE,IAAI,EAAE,EAAE,CAAC;YACP,IAAA,oCAAwB,EAAC,EAAE,EAAE,IAAI,CAAC,MAAM,EAAE,EAAE,CAAC,IAAI,EAAE,iBAAiB,CAAC,CAAC;YACtE,gBAAgB,EAAE,CAAC;QACrB,CAAC;QAED,8CAA8C;QAC9C,MAAM,cAAc,GAAG,IAAA,mBAAa,EAAC,QAAQ,EAAE,EAAE,CAAC,IAAI,CAAC,CAAC;QACxD,MAAM,SAAS,GAAG,IAAA,+BAAiB,EAAC,QAAQ,EAAE,EAAE,CAAC,IAAI,EAAE,cAAc,CAAC,CAAC;QACvE,MAAM,SAAS,GAAG,IAAA,gCAAsB,EAAC,EAAE,EAAE,EAAE,CAAC,IAAI,EAAE,SAAS,CAAC,CAAC;QACjE,cAAc,IAAI,SAAS,CAAC,iBAAiB,CAAC;QAC9C,gBAAgB,IAAI,SAAS,CAAC,gBAAgB,CAAC;IACjD,CAAC;IAED,0CAA0C;IAC1C,MAAM,cAAc,GAAG,UAAU;SAC9B,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,cAAc,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC;SAC5C,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,MAAM,EAAE,CAAC,CAAC,MAAM,EAAE,MAAM,EAAE,2BAA2B,EAAE,CAAC,CAAC,CAAC;IAE3E,+BAA+B;IAC/B,MAAM,QAAQ,GAAG,IAAA,iBAAW,EAAC,QAAQ,CAAC,IAAI,SAAS,CAAC;IACpD,IAAA,iBAAS,EAAC,QAAQ,EAAE,eAAe,EAAE,QAAQ,CAAC,CAAC;IAE/C,UAAU,EAAE,CAAC,EAAE,KAAK,EAAE,MAAM,EAAE,OAAO,EAAE,CAAC,EAAE,KAAK,EAAE,CAAC,EAAE,CAAC,CAAC;IAEtD,kCAAkC;IAClC,OAAO;QACL,UAAU;QACV,YAAY;QACZ,cAAc;QACd,gBAAgB;QAChB,YAAY;QACZ,cAAc;KACf,CAAC;AACJ,CAAC"}

package/dist/spec/build/scan.d.ts

@@ -0,0 +1,48 @@
+/**
+ * Spec↔commit pairing — discovers which commits reference which specs via
+ * conventional-commit scope extraction.
+ *
+ * Replaces `commit4spec/reverse_engineer/git_scanner.py`'s `scan()` (Strategy
+ * A). Shared by the build pipeline and the evolve pipeline's incremental
+ * commit analyzer.
+ *
+ * @module spec/build/scan
+ */
+import { SpecConfig } from '../config';
+import { CommitInfo } from '../git';
+import type { ProgressCallback } from '../ui';
+import { SpecMetadata } from './spec-extractor';
+export interface SpecCommitPair {
+    specId: string;
+    commitHash: string;
+    specMetadata: SpecMetadata | null;
+    commitMetadata: CommitInfo | null;
+}
+/** Result of matching a single commit message against the on-disk specs. */
+export interface SpecMatch {
+    /** Resolved spec ID, or null when the commit scope matched nothing. */
+    specId: string | null;
+    /** Extracted spec metadata (null when the spec doc could not be read). */
+    metadata: SpecMetadata | null;
+}
+/**
+ * Match a single commit message to an on-disk spec: resolve the
+ * conventional-commit scope to a spec ID, then extract that spec's metadata.
+ *
+ * This is the shared per-commit matching step used by both {@link scan}
+ * (build) and the evolve pipeline's incremental analyzer.
+ */
+export declare function matchCommitToSpec(message: string, specIds: Set<string>, specStoragePath: string, config: SpecConfig): SpecMatch;
+/**
+ * Main scanning function — discovers spec↔commit pairs from commit history.
+ *
+ * Iterates every commit in the repo, extracts a conventional-commit scope from
+ * its message, normalizes it, and checks whether the resulting spec ID exists
+ * on disk. When a match is found the pair is recorded with full spec and
+ * commit metadata.
+ *
+ * Duplicates (identical specId + commitHash) are de-duplicated via a
+ * `Map` keyed on `"${specId}|${commitHash}"`.
+ */
+export declare function scan(repoPath: string, specStoragePath: string, config: SpecConfig, onProgress?: ProgressCallback): SpecCommitPair[];
+//# sourceMappingURL=scan.d.ts.map

package/dist/spec/build/scan.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"scan.d.ts","sourceRoot":"","sources":["../../../src/spec/build/scan.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAEH,OAAO,EAAE,UAAU,EAAE,MAAM,WAAW,CAAC;AACvC,OAAO,EAAE,UAAU,EAAiB,MAAM,QAAQ,CAAC;AAEnD,OAAO,KAAK,EAAE,gBAAgB,EAAE,MAAM,OAAO,CAAC;AAE9C,OAAO,EAAuB,YAAY,EAAE,MAAM,kBAAkB,CAAC;AAOrE,MAAM,WAAW,cAAc;IAC7B,MAAM,EAAE,MAAM,CAAC;IACf,UAAU,EAAE,MAAM,CAAC;IACnB,YAAY,EAAE,YAAY,GAAG,IAAI,CAAC;IAClC,cAAc,EAAE,UAAU,GAAG,IAAI,CAAC;CACnC;AAED,4EAA4E;AAC5E,MAAM,WAAW,SAAS;IACxB,uEAAuE;IACvE,MAAM,EAAE,MAAM,GAAG,IAAI,CAAC;IACtB,0EAA0E;IAC1E,QAAQ,EAAE,YAAY,GAAG,IAAI,CAAC;CAC/B;AAMD;;;;;;GAMG;AACH,wBAAgB,iBAAiB,CAC/B,OAAO,EAAE,MAAM,EACf,OAAO,EAAE,GAAG,CAAC,MAAM,CAAC,EACpB,eAAe,EAAE,MAAM,EACvB,MAAM,EAAE,UAAU,GACjB,SAAS,CASX;AAMD;;;;;;;;;;GAUG;AACH,wBAAgB,IAAI,CAClB,QAAQ,EAAE,MAAM,EAChB,eAAe,EAAE,MAAM,EACvB,MAAM,EAAE,UAAU,EAClB,UAAU,CAAC,EAAE,gBAAgB,GAC5B,cAAc,EAAE,CA0ClB"}

package/dist/spec/build/scan.js

@@ -0,0 +1,89 @@
+"use strict";
+/**
+ * Spec↔commit pairing — discovers which commits reference which specs via
+ * conventional-commit scope extraction.
+ *
+ * Replaces `commit4spec/reverse_engineer/git_scanner.py`'s `scan()` (Strategy
+ * A). Shared by the build pipeline and the evolve pipeline's incremental
+ * commit analyzer.
+ *
+ * @module spec/build/scan
+ */
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.matchCommitToSpec = matchCommitToSpec;
+exports.scan = scan;
+const git_1 = require("../git");
+const utils_1 = require("../utils");
+const scope_resolver_1 = require("./scope-resolver");
+const spec_extractor_1 = require("./spec-extractor");
+const errors_1 = require("../../errors");
+// ---------------------------------------------------------------------------
+// matchCommitToSpec
+// ---------------------------------------------------------------------------
+/**
+ * Match a single commit message to an on-disk spec: resolve the
+ * conventional-commit scope to a spec ID, then extract that spec's metadata.
+ *
+ * This is the shared per-commit matching step used by both {@link scan}
+ * (build) and the evolve pipeline's incremental analyzer.
+ */
+function matchCommitToSpec(message, specIds, specStoragePath, config) {
+    const specId = (0, scope_resolver_1.resolveScopeToSpec)(message, specIds, config);
+    if (!specId) {
+        return { specId: null, metadata: null };
+    }
+    return {
+        specId,
+        metadata: (0, spec_extractor_1.extractSpecMetadata)(specStoragePath, specId, config),
+    };
+}
+// ---------------------------------------------------------------------------
+// scan
+// ---------------------------------------------------------------------------
+/**
+ * Main scanning function — discovers spec↔commit pairs from commit history.
+ *
+ * Iterates every commit in the repo, extracts a conventional-commit scope from
+ * its message, normalizes it, and checks whether the resulting spec ID exists
+ * on disk. When a match is found the pair is recorded with full spec and
+ * commit metadata.
+ *
+ * Duplicates (identical specId + commitHash) are de-duplicated via a
+ * `Map` keyed on `"${specId}|${commitHash}"`.
+ */
+function scan(repoPath, specStoragePath, config, onProgress) {
+    const pairs = new Map();
+    // Hoist the directory read out of the per-commit loop.
+    const specIds = new Set((0, utils_1.discoverSpecs)(specStoragePath).map((e) => e.specId));
+    const commits = (0, git_1.getAllCommits)(repoPath);
+    const total = commits.length;
+    let current = 0;
+    for (const commit of commits) {
+        current++;
+        onProgress?.({
+            phase: 'scanning',
+            current,
+            total,
+            message: `${commit.hash.slice(0, 7)} ${commit.message.slice(0, 30)}`,
+        });
+        const { specId, metadata } = matchCommitToSpec(commit.message, specIds, specStoragePath, config);
+        if (!specId) {
+            (0, errors_1.logDebug)('scan: no scope match', {
+                commitHash: commit.hash.slice(0, 7),
+                message: commit.message,
+            });
+            continue;
+        }
+        const key = `${specId}|${commit.hash}`;
+        if (!pairs.has(key)) {
+            pairs.set(key, {
+                specId,
+                commitHash: commit.hash,
+                specMetadata: metadata,
+                commitMetadata: commit,
+            });
+        }
+    }
+    return Array.from(pairs.values());
+}
+//# sourceMappingURL=scan.js.map

package/dist/spec/build/scan.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"scan.js","sourceRoot":"","sources":["../../../src/spec/build/scan.ts"],"names":[],"mappings":";AAAA;;;;;;;;;GASG;;AAwCH,8CAcC;AAiBD,oBA+CC;AAnHD,gCAAmD;AACnD,oCAAyC;AAEzC,qDAAsD;AACtD,qDAAqE;AACrE,yCAAwC;AAqBxC,8EAA8E;AAC9E,oBAAoB;AACpB,8EAA8E;AAE9E;;;;;;GAMG;AACH,SAAgB,iBAAiB,CAC/B,OAAe,EACf,OAAoB,EACpB,eAAuB,EACvB,MAAkB;IAElB,MAAM,MAAM,GAAG,IAAA,mCAAkB,EAAC,OAAO,EAAE,OAAO,EAAE,MAAM,CAAC,CAAC;IAC5D,IAAI,CAAC,MAAM,EAAE,CAAC;QACZ,OAAO,EAAE,MAAM,EAAE,IAAI,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAC;IAC1C,CAAC;IACD,OAAO;QACL,MAAM;QACN,QAAQ,EAAE,IAAA,oCAAmB,EAAC,eAAe,EAAE,MAAM,EAAE,MAAM,CAAC;KAC/D,CAAC;AACJ,CAAC;AAED,8EAA8E;AAC9E,OAAO;AACP,8EAA8E;AAE9E;;;;;;;;;;GAUG;AACH,SAAgB,IAAI,CAClB,QAAgB,EAChB,eAAuB,EACvB,MAAkB,EAClB,UAA6B;IAE7B,MAAM,KAAK,GAAG,IAAI,GAAG,EAA0B,CAAC;IAEhD,uDAAuD;IACvD,MAAM,OAAO,GAAG,IAAI,GAAG,CAAC,IAAA,qBAAa,EAAC,eAAe,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC;IAE7E,MAAM,OAAO,GAAG,IAAA,mBAAa,EAAC,QAAQ,CAAC,CAAC;IACxC,MAAM,KAAK,GAAG,OAAO,CAAC,MAAM,CAAC;IAC7B,IAAI,OAAO,GAAG,CAAC,CAAC;IAEhB,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC7B,OAAO,EAAE,CAAC;QACV,UAAU,EAAE,CAAC;YACX,KAAK,EAAE,UAAU;YACjB,OAAO;YACP,KAAK;YACL,OAAO,EAAE,GAAG,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE;SACrE,CAAC,CAAC;QAEH,MAAM,EAAE,MAAM,EAAE,QAAQ,EAAE,GAAG,iBAAiB,CAC5C,MAAM,CAAC,OAAO,EAAE,OAAO,EAAE,eAAe,EAAE,MAAM,CACjD,CAAC;QACF,IAAI,CAAC,MAAM,EAAE,CAAC;YACZ,IAAA,iBAAQ,EAAC,sBAAsB,EAAE;gBAC/B,UAAU,EAAE,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC;gBACnC,OAAO,EAAE,MAAM,CAAC,OAAO;aACxB,CAAC,CAAC;YACH,SAAS;QACX,CAAC;QAED,MAAM,GAAG,GAAG,GAAG,MAAM,IAAI,MAAM,CAAC,IAAI,EAAE,CAAC;QACvC,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC;YACpB,KAAK,CAAC,GAAG,CAAC,GAAG,EAAE;gBACb,MAAM;gBACN,UAAU,EAAE,MAAM,CAAC,IAAI;gBACvB,YAAY,EAAE,QAAQ;gBACtB,cAAc,EAAE,MAAM;aACvB,CAAC,CAAC;QACL,CAAC;IACH,CAAC;IAED,OAAO,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,EAAE,CAAC,CAAC;AACpC,CAAC"}