tryassay 0.28.2 → 0.30.1

package/dist/lib/rule-harvester/pr-discovery.js

@@ -0,0 +1,168 @@
+import { execSync } from 'node:child_process';
+import { parsePatch } from './diff-parser.js';
+// ── Constants ──────────────────────────────────────────────────
+/** Default labels that indicate a bug-fix PR. Matched case-insensitively as substrings. */
+export const DEFAULT_BUG_LABELS = [
+    'bug',
+    'fix',
+    'bugfix',
+    'hotfix',
+    'security',
+    'patch',
+    'defect',
+    'regression',
+];
+/** File extensions we mine patterns from. */
+const TARGET_EXTENSIONS = /\.(tsx?|jsx?|py)$/;
+/** Paths that indicate test files — excluded from mining. */
+const TEST_PATH_PATTERNS = /\.(test|spec)\.[tj]sx?$|__tests__[/\\]|tests?[/\\]|_test\.py$|test_[^/\\]+\.py$/;
+/** Declaration files — excluded from mining. */
+const DECLARATION_PATTERN = /\.d\.ts$|\.pyi$/;
+// ── Pure Filter Functions (unit-testable) ──────────────────────
+/**
+ * Returns true if any label in the array contains a bug-related keyword.
+ * Matches case-insensitively. Accepts optional custom bug label list.
+ */
+export function matchesBugLabels(labels, bugLabels = DEFAULT_BUG_LABELS) {
+    const lower = labels.map((l) => l.toLowerCase());
+    return lower.some((label) => bugLabels.some((keyword) => label.includes(keyword)));
+}
+/**
+ * Returns true if the PR title matches common bug-fix prefixes.
+ * Looks for fix/bug/resolve/patch at the start or after a conventional commit prefix.
+ */
+export function matchesBugTitle(title) {
+    return /\b(fix|bug|resolve|hotfix|patch)\b/i.test(title);
+}
+/**
+ * Returns true if the file path is a TypeScript/JavaScript source file
+ * that is NOT a test file or declaration file.
+ */
+export function isTargetFile(path) {
+    if (!TARGET_EXTENSIONS.test(path))
+        return false;
+    if (TEST_PATH_PATTERNS.test(path))
+        return false;
+    if (DECLARATION_PATTERN.test(path))
+        return false;
+    return true;
+}
+/**
+ * Call the `gh api` CLI and return parsed JSON.
+ * Relies on gh's built-in auth and rate-limit handling.
+ */
+function ghApi(endpoint) {
+    const raw = execSync(`gh api "${endpoint}" --paginate`, {
+        encoding: 'utf-8',
+        maxBuffer: 50 * 1024 * 1024, // 50 MB — large repos can have big diffs
+        timeout: 60_000,
+    });
+    return JSON.parse(raw);
+}
+/**
+ * Fetch a single page from the GitHub API (no --paginate).
+ */
+function ghApiPage(endpoint) {
+    const raw = execSync(`gh api "${endpoint}"`, {
+        encoding: 'utf-8',
+        maxBuffer: 50 * 1024 * 1024,
+        timeout: 60_000,
+    });
+    return JSON.parse(raw);
+}
+/**
+ * Discover merged bug-fix PRs from a GitHub repo.
+ *
+ * Flow:
+ * 1. Fetch closed PRs (paginated)
+ * 2. Filter to merged-only (merged_at !== null)
+ * 3. Filter by bug labels (with optional title fallback)
+ * 4. Optionally enrich with review comments
+ * 5. Collect changed file paths
+ */
+export function discoverBugFixPRs(owner, name, config = {}) {
+    const { perPage = 30, maxPages = 3, bugLabels = DEFAULT_BUG_LABELS, titleFallback = true, fetchComments = true, } = config;
+    const discovered = [];
+    for (let page = 1; page <= maxPages; page++) {
+        const endpoint = `/repos/${owner}/${name}/pulls?state=closed&sort=updated&direction=desc&per_page=${perPage}&page=${page}`;
+        let prs;
+        try {
+            prs = ghApiPage(endpoint);
+        }
+        catch {
+            // If the API call fails (auth, network, etc.), stop pagination
+            break;
+        }
+        if (!prs || prs.length === 0)
+            break;
+        for (const pr of prs) {
+            // Only merged PRs
+            if (!pr.merged_at || !pr.merge_commit_sha)
+                continue;
+            const labelNames = pr.labels.map((l) => l.name);
+            const hasBugLabel = matchesBugLabels(labelNames, bugLabels);
+            const hasBugTitle = titleFallback && matchesBugTitle(pr.title);
+            if (!hasBugLabel && !hasBugTitle)
+                continue;
+            // Fetch file list for this PR
+            let files = [];
+            try {
+                const prFiles = ghApi(`/repos/${owner}/${name}/pulls/${pr.number}/files`);
+                files = prFiles.map((f) => f.filename);
+            }
+            catch {
+                // If we can't get files, still include the PR with empty files
+            }
+            // Fetch review comments if requested
+            let reviewComments = [];
+            if (fetchComments) {
+                try {
+                    const comments = ghApi(`/repos/${owner}/${name}/pulls/${pr.number}/comments`);
+                    reviewComments = comments.map((c) => ({
+                        body: c.body,
+                        path: c.path,
+                        line: c.line ?? c.original_line ?? null,
+                    }));
+                }
+                catch {
+                    // Review comments are optional enrichment
+                }
+            }
+            discovered.push({
+                repo: `${owner}/${name}`,
+                prNumber: pr.number,
+                title: pr.title,
+                labels: labelNames,
+                mergeCommit: pr.merge_commit_sha,
+                reviewComments,
+                files,
+            });
+        }
+    }
+    return discovered;
+}
+/**
+ * Fetch and parse diffs for a discovered PR.
+ *
+ * Calls the files endpoint, filters to target files,
+ * and parses each file's patch into typed DiffHunks.
+ */
+export function fetchPRDiffs(owner, name, pr) {
+    let allHunks = [];
+    try {
+        const files = ghApi(`/repos/${owner}/${name}/pulls/${pr.prNumber}/files`);
+        for (const file of files) {
+            if (!isTargetFile(file.filename))
+                continue;
+            if (!file.patch)
+                continue;
+            const hunks = parsePatch(file.filename, file.patch);
+            allHunks = allHunks.concat(hunks);
+        }
+    }
+    catch {
+        // Return empty hunks if API call fails
+    }
+    return { pr, hunks: allHunks };
+}
+//# sourceMappingURL=pr-discovery.js.map

package/dist/lib/rule-harvester/pr-discovery.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"pr-discovery.js","sourceRoot":"","sources":["../../../src/lib/rule-harvester/pr-discovery.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,MAAM,oBAAoB,CAAC;AAE9C,OAAO,EAAE,UAAU,EAAE,MAAM,kBAAkB,CAAC;AAE9C,kEAAkE;AAElE,2FAA2F;AAC3F,MAAM,CAAC,MAAM,kBAAkB,GAAsB;IACnD,KAAK;IACL,KAAK;IACL,QAAQ;IACR,QAAQ;IACR,UAAU;IACV,OAAO;IACP,QAAQ;IACR,YAAY;CACJ,CAAC;AAEX,6CAA6C;AAC7C,MAAM,iBAAiB,GAAG,mBAAmB,CAAC;AAE9C,6DAA6D;AAC7D,MAAM,kBAAkB,GAAG,iFAAiF,CAAC;AAE7G,gDAAgD;AAChD,MAAM,mBAAmB,GAAG,iBAAiB,CAAC;AAE9C,kEAAkE;AAElE;;;GAGG;AACH,MAAM,UAAU,gBAAgB,CAC9B,MAAgB,EAChB,YAA+B,kBAAkB;IAEjD,MAAM,KAAK,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,WAAW,EAAE,CAAC,CAAC;IACjD,OAAO,KAAK,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAC1B,SAAS,CAAC,IAAI,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,KAAK,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC,CACrD,CAAC;AACJ,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,eAAe,CAAC,KAAa;IAC3C,OAAO,qCAAqC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;AAC3D,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,YAAY,CAAC,IAAY;IACvC,IAAI,CAAC,iBAAiB,CAAC,IAAI,CAAC,IAAI,CAAC;QAAE,OAAO,KAAK,CAAC;IAChD,IAAI,kBAAkB,CAAC,IAAI,CAAC,IAAI,CAAC;QAAE,OAAO,KAAK,CAAC;IAChD,IAAI,mBAAmB,CAAC,IAAI,CAAC,IAAI,CAAC;QAAE,OAAO,KAAK,CAAC;IACjD,OAAO,IAAI,CAAC;AACd,CAAC;AAyCD;;;GAGG;AACH,SAAS,KAAK,CAAI,QAAgB;IAChC,MAAM,GAAG,GAAG,QAAQ,CAAC,WAAW,QAAQ,cAAc,EAAE;QACtD,QAAQ,EAAE,OAAO;QACjB,SAAS,EAAE,EAAE,GAAG,IAAI,GAAG,IAAI,EAAE,yCAAyC;QACtE,OAAO,EAAE,MAAM;KAChB,CAAC,CAAC;IACH,OAAO,IAAI,CAAC,KAAK,CAAC,GAAG,CAAM,CAAC;AAC9B,CAAC;AAED;;GAEG;AACH,SAAS,SAAS,CAAI,QAAgB;IACpC,MAAM,GAAG,GAAG,QAAQ,CAAC,WAAW,QAAQ,GAAG,EAAE;QAC3C,QAAQ,EAAE,OAAO;QACjB,SAAS,EAAE,EAAE,GAAG,IAAI,GAAG,IAAI;QAC3B,OAAO,EAAE,MAAM;KAChB,CAAC,CAAC;IACH,OAAO,IAAI,CAAC,KAAK,CAAC,GAAG,CAAM,CAAC;AAC9B,CAAC;AAED;;;;;;;;;GASG;AACH,MAAM,UAAU,iBAAiB,CAC/B,KAAa,EACb,IAAY,EACZ,SAA0B,EAAE;IAE5B,MAAM,EACJ,OAAO,GAAG,EAAE,EACZ,QAAQ,GAAG,CAAC,EACZ,SAAS,GAAG,kBAAkB,EAC9B,aAAa,GAAG,IAAI,EACpB,aAAa,GAAG,IAAI,GACrB,GAAG,MAAM,CAAC;IAEX,MAAM,UAAU,GAAmB,EAAE,CAAC;IAEtC,KAAK,IAAI,IAAI,GAAG,CAAC,EAAE,IAAI,IAAI,QAAQ,EAAE,IAAI,EAAE,EAAE,CAAC;QAC5C,MAAM,QAAQ,GACZ,UAAU,KAAK,IAAI,IAAI,4DAA4D,OAAO,SAAS,IAAI,EAAE,CAAC;QAE5G,IAAI,GAAe,CAAC;QACpB,IAAI,CAAC;YACH,GAAG,GAAG,SAAS,CAAa,QAAQ,CAAC,CAAC;QACxC,CAAC;QAAC,MAAM,CAAC;YACP,+DAA+D;YAC/D,MAAM;QACR,CAAC;QAED,IAAI,CAAC,GAAG,IAAI,GAAG,CAAC,MAAM,KAAK,CAAC;YAAE,MAAM;QAEpC,KAAK,MAAM,EAAE,IAAI,GAAG,EAAE,CAAC;YACrB,kBAAkB;YAClB,IAAI,CAAC,EAAE,CAAC,SAAS,IAAI,CAAC,EAAE,CAAC,gBAAgB;gBAAE,SAAS;YAEpD,MAAM,UAAU,GAAG,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;YAChD,MAAM,WAAW,GAAG,gBAAgB,CAAC,UAAU,EAAE,SAAS,CAAC,CAAC;YAC5D,MAAM,WAAW,GAAG,aAAa,IAAI,eAAe,CAAC,EAAE,CAAC,KAAK,CAAC,CAAC;YAE/D,IAAI,CAAC,WAAW,IAAI,CAAC,WAAW;gBAAE,SAAS;YAE3C,8BAA8B;YAC9B,IAAI,KAAK,GAAa,EAAE,CAAC;YACzB,IAAI,CAAC;gBACH,MAAM,OAAO,GAAG,KAAK,CACnB,UAAU,KAAK,IAAI,IAAI,UAAU,EAAE,CAAC,MAAM,QAAQ,CACnD,CAAC;gBACF,KAAK,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC;YACzC,CAAC;YAAC,MAAM,CAAC;gBACP,+DAA+D;YACjE,CAAC;YAED,qCAAqC;YACrC,IAAI,cAAc,GAAsB,EAAE,CAAC;YAC3C,IAAI,aAAa,EAAE,CAAC;gBAClB,IAAI,CAAC;oBACH,MAAM,QAAQ,GAAG,KAAK,CACpB,UAAU,KAAK,IAAI,IAAI,UAAU,EAAE,CAAC,MAAM,WAAW,CACtD,CAAC;oBACF,cAAc,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;wBACpC,IAAI,EAAE,CAAC,CAAC,IAAI;wBACZ,IAAI,EAAE,CAAC,CAAC,IAAI;wBACZ,IAAI,EAAE,CAAC,CAAC,IAAI,IAAI,CAAC,CAAC,aAAa,IAAI,IAAI;qBACxC,CAAC,CAAC,CAAC;gBACN,CAAC;gBAAC,MAAM,CAAC;oBACP,0CAA0C;gBAC5C,CAAC;YACH,CAAC;YAED,UAAU,CAAC,IAAI,CAAC;gBACd,IAAI,EAAE,GAAG,KAAK,IAAI,IAAI,EAAE;gBACxB,QAAQ,EAAE,EAAE,CAAC,MAAM;gBACnB,KAAK,EAAE,EAAE,CAAC,KAAK;gBACf,MAAM,EAAE,UAAU;gBAClB,WAAW,EAAE,EAAE,CAAC,gBAAgB;gBAChC,cAAc;gBACd,KAAK;aACN,CAAC,CAAC;QACL,CAAC;IACH,CAAC;IAED,OAAO,UAAU,CAAC;AACpB,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,YAAY,CAC1B,KAAa,EACb,IAAY,EACZ,EAAgB;IAEhB,IAAI,QAAQ,GAAe,EAAE,CAAC;IAE9B,IAAI,CAAC;QACH,MAAM,KAAK,GAAG,KAAK,CACjB,UAAU,KAAK,IAAI,IAAI,UAAU,EAAE,CAAC,QAAQ,QAAQ,CACrD,CAAC;QAEF,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YACzB,IAAI,CAAC,YAAY,CAAC,IAAI,CAAC,QAAQ,CAAC;gBAAE,SAAS;YAC3C,IAAI,CAAC,IAAI,CAAC,KAAK;gBAAE,SAAS;YAE1B,MAAM,KAAK,GAAG,UAAU,CAAC,IAAI,CAAC,QAAQ,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC;YACpD,QAAQ,GAAG,QAAQ,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;QACpC,CAAC;IACH,CAAC;IAAC,MAAM,CAAC;QACP,uCAAuC;IACzC,CAAC;IAED,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,QAAQ,EAAE,CAAC;AACjC,CAAC"}

package/dist/lib/rule-harvester/pr-harvest.d.ts

@@ -0,0 +1,53 @@
+/**
+ * PR Harvest Orchestrator — ties discovery -> extraction -> generalization -> catalog
+ * into a complete pipeline for mining bug-fix PRs from public GitHub repos.
+ *
+ * PR-sourced rules are validated through the validation harness before promotion.
+ * Fuzzy deduplication via Dice coefficient prevents near-duplicate rules.
+ */
+import type { LearnedRule } from '../learned-rules/types.js';
+/** Per-repo set of processed PR numbers. */
+export type PRProgress = Record<string, number[]>;
+export declare function createPRProgress(): PRProgress;
+export declare function markPRProcessed(progress: PRProgress, repo: string, prNumber: number): PRProgress;
+export declare function isPRProcessed(progress: PRProgress, repo: string, prNumber: number): boolean;
+export declare function getPRProgressStats(progress: PRProgress, repo: string): {
+    processedPRs: number;
+};
+export declare function loadPRProgress(dir: string): Promise<PRProgress>;
+export declare function savePRProgress(dir: string, progress: PRProgress): Promise<void>;
+/**
+ * Dice coefficient (bigram overlap) for fuzzy string similarity.
+ * Returns a value between 0 (no overlap) and 1 (identical).
+ */
+export declare function diceCoefficient(a: string, b: string): number;
+/**
+ * Find the most similar existing rule by description, if above threshold.
+ */
+export declare function findFuzzyDuplicate(existingRules: LearnedRule[], description: string, category: string): LearnedRule | null;
+export interface PRHarvestConfig {
+    /** List of repos to mine PRs from. */
+    repoList: Array<{
+        owner: string;
+        name: string;
+    }>;
+    /** Assay project root where rules go. */
+    catalogPath: string;
+    /** Where pr-progress.json lives. */
+    progressDir: string;
+    /** Where run reports go. */
+    runsDir: string;
+    /** Model name for reporting. */
+    model: string;
+    /** Max repos to process. */
+    limit?: number;
+    /** Only scan these repos (owner/name format). */
+    repoFilter?: string[];
+    /** Skip already-processed PRs. */
+    resume?: boolean;
+    /** Max PRs per repo. */
+    maxPRsPerRepo?: number;
+    /** Logging callback. */
+    onLog?: (msg: string) => void;
+}
+export declare function harvestPRs(config: PRHarvestConfig): Promise<void>;

package/dist/lib/rule-harvester/pr-harvest.js

@@ -0,0 +1,326 @@
+/**
+ * PR Harvest Orchestrator — ties discovery -> extraction -> generalization -> catalog
+ * into a complete pipeline for mining bug-fix PRs from public GitHub repos.
+ *
+ * PR-sourced rules are validated through the validation harness before promotion.
+ * Fuzzy deduplication via Dice coefficient prevents near-duplicate rules.
+ */
+import { readFile, writeFile, mkdir } from 'node:fs/promises';
+import { join } from 'node:path';
+import { discoverBugFixPRs, fetchPRDiffs } from './pr-discovery.js';
+import { filterHunks } from './diff-parser.js';
+import { generalizeHunk, mapToExtractedPattern } from './rule-generalizer.js';
+import { addRule, updateRule, loadRules, loadStats } from '../learned-rules/learned-catalog.js';
+import { validateRule } from '../learned-rules/validation-harness.js';
+import { saveRunReport } from './reporter.js';
+export function createPRProgress() {
+    return {};
+}
+export function markPRProcessed(progress, repo, prNumber) {
+    const existing = progress[repo] ?? [];
+    if (existing.includes(prNumber))
+        return progress;
+    return { ...progress, [repo]: [...existing, prNumber] };
+}
+export function isPRProcessed(progress, repo, prNumber) {
+    return (progress[repo] ?? []).includes(prNumber);
+}
+export function getPRProgressStats(progress, repo) {
+    return { processedPRs: (progress[repo] ?? []).length };
+}
+export async function loadPRProgress(dir) {
+    const filePath = join(dir, 'pr-progress.json');
+    try {
+        const content = await readFile(filePath, 'utf-8');
+        return JSON.parse(content);
+    }
+    catch {
+        return {};
+    }
+}
+export async function savePRProgress(dir, progress) {
+    await mkdir(dir, { recursive: true });
+    const filePath = join(dir, 'pr-progress.json');
+    await writeFile(filePath, JSON.stringify(progress, null, 2), 'utf-8');
+}
+// ── Fuzzy Deduplication ─────────────────────────────────────
+/**
+ * Dice coefficient (bigram overlap) for fuzzy string similarity.
+ * Returns a value between 0 (no overlap) and 1 (identical).
+ */
+export function diceCoefficient(a, b) {
+    if (a === b)
+        return 1;
+    if (a.length < 2 || b.length < 2)
+        return 0;
+    const bigramsA = new Map();
+    for (let i = 0; i < a.length - 1; i++) {
+        const bigram = a.substring(i, i + 2);
+        bigramsA.set(bigram, (bigramsA.get(bigram) ?? 0) + 1);
+    }
+    const bigramsB = new Map();
+    for (let i = 0; i < b.length - 1; i++) {
+        const bigram = b.substring(i, i + 2);
+        bigramsB.set(bigram, (bigramsB.get(bigram) ?? 0) + 1);
+    }
+    let intersectionSize = 0;
+    for (const [bigram, countA] of bigramsA) {
+        const countB = bigramsB.get(bigram) ?? 0;
+        intersectionSize += Math.min(countA, countB);
+    }
+    return (2 * intersectionSize) / (a.length - 1 + (b.length - 1));
+}
+/** Similarity threshold above which two descriptions are considered duplicates. */
+const DEDUP_THRESHOLD = 0.7;
+/**
+ * Find the most similar existing rule by description, if above threshold.
+ */
+export function findFuzzyDuplicate(existingRules, description, category) {
+    let bestMatch = null;
+    let bestScore = 0;
+    for (const rule of existingRules) {
+        // Only compare within the same category
+        if (rule.pattern.claimCategory !== category)
+            continue;
+        if (rule.status === 'rejected' || rule.status === 'deprecated')
+            continue;
+        const score = diceCoefficient(rule.pattern.description.toLowerCase(), description.toLowerCase());
+        if (score > bestScore) {
+            bestScore = score;
+            bestMatch = rule;
+        }
+    }
+    return bestScore >= DEDUP_THRESHOLD ? bestMatch : null;
+}
+// ── Rule Construction ───────────────────────────────────────
+function buildLearnedRule(generalizedRule, patternId, repo, prNumber, hunkFile, buggyCode) {
+    const now = new Date().toISOString();
+    const pattern = mapToExtractedPattern(generalizedRule, patternId);
+    // Use actual buggy code from the diff hunk so validation has real test material
+    const codeSnippet = buggyCode && buggyCode.trim().length > 10
+        ? buggyCode
+        : `PR #${prNumber}: ${generalizedRule.fix.description}`;
+    const sourceFinding = {
+        claimId: `pr_${repo.replace('/', '_')}_${prNumber}`,
+        claimDescription: generalizedRule.description,
+        codeSnippet,
+        filePath: hunkFile,
+        language: generalizedRule.detection.language,
+        timestamp: now,
+    };
+    return {
+        id: patternId,
+        pattern,
+        status: 'promoted', // Default; overridden by validation harness result
+        createdAt: now,
+        updatedAt: now,
+        fireCount: 0,
+        truePositiveCount: 0,
+        falsePositiveCount: 0,
+        sourceFindings: [sourceFinding],
+        source: 'pr',
+        fixDescription: generalizedRule.fix.description,
+        fixPattern: generalizedRule.fix.pattern,
+        confirmationCount: 1,
+    };
+}
+// ── Main Orchestrator ───────────────────────────────────────
+function log(config, msg) {
+    if (config.onLog) {
+        config.onLog(msg);
+    }
+    else {
+        console.log(msg);
+    }
+}
+export async function harvestPRs(config) {
+    const startTime = Date.now();
+    // 1. Load progress if resuming
+    let progress = config.resume
+        ? await loadPRProgress(config.progressDir)
+        : createPRProgress();
+    if (config.resume) {
+        const totalTracked = Object.values(progress).reduce((sum, prs) => sum + prs.length, 0);
+        log(config, `Resumed PR progress: ${totalTracked} PRs tracked`);
+    }
+    // 2. Filter repos
+    let repos = config.repoList;
+    if (config.repoFilter && config.repoFilter.length > 0) {
+        const filterSet = new Set(config.repoFilter);
+        repos = repos.filter((r) => filterSet.has(`${r.owner}/${r.name}`));
+    }
+    if (config.limit !== undefined && config.limit > 0) {
+        repos = repos.slice(0, config.limit);
+    }
+    log(config, `Processing ${repos.length} repos in PR mode`);
+    // Tracking for the run report
+    let totalPRsDiscovered = 0;
+    let totalHunksProcessed = 0;
+    let rulesLearned = 0;
+    let rulesRejected = 0;
+    let rulesDuplicate = 0;
+    const unmatchedCategories = {};
+    for (const repo of repos) {
+        const key = `${repo.owner}/${repo.name}`;
+        log(config, `[PR-DISCOVER] ${key}`);
+        // 3. Discover bug-fix PRs
+        let discoveredPRs;
+        try {
+            const maxPRs = config.maxPRsPerRepo ?? 100;
+            const allDiscoveredPRs = discoverBugFixPRs(repo.owner, repo.name, {
+                perPage: 30,
+                maxPages: Math.ceil(maxPRs / 30),
+                titleFallback: true,
+                fetchComments: true,
+            });
+            discoveredPRs = allDiscoveredPRs.slice(0, maxPRs);
+        }
+        catch (err) {
+            log(config, `[FAIL] ${key} — discovery failed: ${String(err)}`);
+            continue;
+        }
+        log(config, `[FOUND] ${key} — ${discoveredPRs.length} bug-fix PRs`);
+        totalPRsDiscovered += discoveredPRs.length;
+        for (const pr of discoveredPRs) {
+            // Skip already-processed PRs on resume
+            if (config.resume && isPRProcessed(progress, key, pr.prNumber)) {
+                log(config, `  [SKIP] PR #${pr.prNumber} — already processed`);
+                continue;
+            }
+            log(config, `  [PR] #${pr.prNumber}: ${pr.title}`);
+            // 4. Fetch and parse diffs
+            let parsedDiff;
+            try {
+                parsedDiff = fetchPRDiffs(repo.owner, repo.name, pr);
+            }
+            catch (err) {
+                log(config, `  [DIFF_ERR] PR #${pr.prNumber} — ${String(err)}`);
+                progress = markPRProcessed(progress, key, pr.prNumber);
+                await savePRProgress(config.progressDir, progress);
+                continue;
+            }
+            // Filter to meaningful hunks
+            const meaningfulHunks = filterHunks(parsedDiff.hunks);
+            if (meaningfulHunks.length === 0) {
+                log(config, `  [SKIP] PR #${pr.prNumber} — no meaningful hunks`);
+                progress = markPRProcessed(progress, key, pr.prNumber);
+                await savePRProgress(config.progressDir, progress);
+                continue;
+            }
+            log(config, `  [HUNKS] ${meaningfulHunks.length} meaningful (of ${parsedDiff.hunks.length} total)`);
+            totalHunksProcessed += meaningfulHunks.length;
+            // Load catalog once before processing hunks (refreshed after writes)
+            let existingRules = await loadRules(config.catalogPath);
+            // 5. Generalize each hunk
+            for (const hunk of meaningfulHunks) {
+                // Find matching review comment for this hunk's file
+                const matchingComment = pr.reviewComments.find((c) => c.path === hunk.file);
+                let generalizedRule;
+                try {
+                    generalizedRule = await generalizeHunk(hunk, key, pr.prNumber, pr.title, matchingComment?.body);
+                }
+                catch (err) {
+                    log(config, `    [GEN_ERR] ${hunk.file}:${hunk.startLine} — ${String(err)}`);
+                    rulesRejected++;
+                    continue;
+                }
+                if (!generalizedRule) {
+                    log(config, `    [SKIP] ${hunk.file}:${hunk.startLine} — too context-specific`);
+                    rulesRejected++;
+                    continue;
+                }
+                // 6. Check for fuzzy duplicates in catalog
+                const duplicate = findFuzzyDuplicate(existingRules, generalizedRule.description, generalizedRule.category);
+                if (duplicate) {
+                    // Merge provenance into existing rule
+                    log(config, `    [DEDUP] "${generalizedRule.description}" → merging into ${duplicate.id}`);
+                    await updateRule(config.catalogPath, duplicate.id, (existing) => ({
+                        ...existing,
+                        confirmationCount: (existing.confirmationCount ?? 1) + 1,
+                        updatedAt: new Date().toISOString(),
+                        sourceFindings: [
+                            ...existing.sourceFindings,
+                            {
+                                claimId: `pr_${key.replace('/', '_')}_${pr.prNumber}`,
+                                claimDescription: generalizedRule.description,
+                                codeSnippet: `PR #${pr.prNumber}: ${generalizedRule.fix.description}`,
+                                filePath: hunk.file,
+                                language: generalizedRule.detection.language,
+                                timestamp: new Date().toISOString(),
+                            },
+                        ],
+                    }));
+                    rulesDuplicate++;
+                    existingRules = await loadRules(config.catalogPath);
+                    continue;
+                }
+                // 7. Build candidate rule and validate before promotion
+                const patternId = `lp_pr_${Date.now()}`;
+                // Pass the buggy code (removed lines + context) so validation has real test material
+                const buggyCode = [...hunk.context, ...hunk.removedLines].join('\n');
+                const candidateRule = buildLearnedRule(generalizedRule, patternId, key, pr.prNumber, hunk.file, buggyCode);
+                // 7b. Run through validation harness
+                const validationResult = validateRule(candidateRule);
+                const finalRule = {
+                    ...candidateRule,
+                    status: validationResult.passed ? 'promoted' : 'rejected',
+                    validationResults: validationResult,
+                };
+                await addRule(config.catalogPath, finalRule);
+                existingRules = await loadRules(config.catalogPath);
+                if (validationResult.passed) {
+                    rulesLearned++;
+                    log(config, `    [LEARNED] ${patternId}: ${generalizedRule.description} (precision=${validationResult.precision.toFixed(2)} recall=${validationResult.recall.toFixed(2)})`);
+                }
+                else {
+                    rulesRejected++;
+                    log(config, `    [REJECTED] ${patternId}: ${generalizedRule.description} — validation failed (precision=${validationResult.precision.toFixed(2)} recall=${validationResult.recall.toFixed(2)})`);
+                }
+            }
+            // Mark PR as processed and persist after each PR
+            progress = markPRProcessed(progress, key, pr.prNumber);
+            await savePRProgress(config.progressDir, progress);
+        }
+        log(config, `[DONE] ${key}`);
+    }
+    // 8. Generate and save run report
+    const stats = await loadStats(config.catalogPath);
+    const allRules = await loadRules(config.catalogPath);
+    const catalogSize = {
+        total: allRules.length,
+        promoted: allRules.filter((r) => r.status === 'promoted').length,
+        rejected: allRules.filter((r) => r.status === 'rejected').length,
+    };
+    const durationMinutes = (Date.now() - startTime) / 60_000;
+    const report = {
+        timestamp: new Date().toISOString(),
+        model: config.model,
+        reposScanned: repos.length,
+        filesScanned: totalHunksProcessed, // hunks are the PR-mode equivalent of files
+        claimsExtracted: totalPRsDiscovered,
+        findingsConfirmed: totalHunksProcessed,
+        rulesLearned,
+        rulesRejected,
+        rulesDuplicate,
+        unmatchedCategories,
+        durationMinutes,
+        catalogSize,
+    };
+    const reportPath = await saveRunReport(config.runsDir, report);
+    // 9. Print summary
+    log(config, '');
+    log(config, '══════════════════════════════════════');
+    log(config, '  PR HARVEST COMPLETE');
+    log(config, '══════════════════════════════════════');
+    log(config, `  Repos processed:    ${repos.length}`);
+    log(config, `  PRs discovered:     ${totalPRsDiscovered}`);
+    log(config, `  Hunks processed:    ${totalHunksProcessed}`);
+    log(config, `  Rules learned:      ${rulesLearned}`);
+    log(config, `  Rules duplicate:    ${rulesDuplicate}`);
+    log(config, `  Rules rejected:     ${rulesRejected}`);
+    log(config, `  Catalog total:      ${catalogSize.total} (${catalogSize.promoted} promoted)`);
+    log(config, `  Duration:           ${durationMinutes.toFixed(1)}m`);
+    log(config, `  Report:             ${reportPath}`);
+    log(config, '══════════════════════════════════════');
+}
+//# sourceMappingURL=pr-harvest.js.map

package/dist/lib/rule-harvester/pr-harvest.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"pr-harvest.js","sourceRoot":"","sources":["../../../src/lib/rule-harvester/pr-harvest.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAEH,OAAO,EAAE,QAAQ,EAAE,SAAS,EAAE,KAAK,EAAE,MAAM,kBAAkB,CAAC;AAC9D,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;AAEjC,OAAO,EAAE,iBAAiB,EAAE,YAAY,EAAE,MAAM,mBAAmB,CAAC;AACpE,OAAO,EAAE,WAAW,EAAE,MAAM,kBAAkB,CAAC;AAC/C,OAAO,EAAE,cAAc,EAAE,qBAAqB,EAAE,MAAM,uBAAuB,CAAC;AAC9E,OAAO,EAAE,OAAO,EAAE,UAAU,EAAE,SAAS,EAAE,SAAS,EAAE,MAAM,qCAAqC,CAAC;AAChG,OAAO,EAAE,YAAY,EAAE,MAAM,wCAAwC,CAAC;AACtE,OAAO,EAAE,aAAa,EAAkB,MAAM,eAAe,CAAC;AAa9D,MAAM,UAAU,gBAAgB;IAC9B,OAAO,EAAE,CAAC;AACZ,CAAC;AAED,MAAM,UAAU,eAAe,CAC7B,QAAoB,EACpB,IAAY,EACZ,QAAgB;IAEhB,MAAM,QAAQ,GAAG,QAAQ,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC;IACtC,IAAI,QAAQ,CAAC,QAAQ,CAAC,QAAQ,CAAC;QAAE,OAAO,QAAQ,CAAC;IACjD,OAAO,EAAE,GAAG,QAAQ,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,GAAG,QAAQ,EAAE,QAAQ,CAAC,EAAE,CAAC;AAC1D,CAAC;AAED,MAAM,UAAU,aAAa,CAC3B,QAAoB,EACpB,IAAY,EACZ,QAAgB;IAEhB,OAAO,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC;AACnD,CAAC;AAED,MAAM,UAAU,kBAAkB,CAChC,QAAoB,EACpB,IAAY;IAEZ,OAAO,EAAE,YAAY,EAAE,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC;AACzD,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,cAAc,CAAC,GAAW;IAC9C,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,EAAE,kBAAkB,CAAC,CAAC;IAC/C,IAAI,CAAC;QACH,MAAM,OAAO,GAAG,MAAM,QAAQ,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC;QAClD,OAAO,IAAI,CAAC,KAAK,CAAC,OAAO,CAAe,CAAC;IAC3C,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,EAAE,CAAC;IACZ,CAAC;AACH,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,cAAc,CAClC,GAAW,EACX,QAAoB;IAEpB,MAAM,KAAK,CAAC,GAAG,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;IACtC,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,EAAE,kBAAkB,CAAC,CAAC;IAC/C,MAAM,SAAS,CAAC,QAAQ,EAAE,IAAI,CAAC,SAAS,CAAC,QAAQ,EAAE,IAAI,EAAE,CAAC,CAAC,EAAE,OAAO,CAAC,CAAC;AACxE,CAAC;AAED,+DAA+D;AAE/D;;;GAGG;AACH,MAAM,UAAU,eAAe,CAAC,CAAS,EAAE,CAAS;IAClD,IAAI,CAAC,KAAK,CAAC;QAAE,OAAO,CAAC,CAAC;IACtB,IAAI,CAAC,CAAC,MAAM,GAAG,CAAC,IAAI,CAAC,CAAC,MAAM,GAAG,CAAC;QAAE,OAAO,CAAC,CAAC;IAE3C,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAkB,CAAC;IAC3C,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;QACtC,MAAM,MAAM,GAAG,CAAC,CAAC,SAAS,CAAC,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;QACrC,QAAQ,CAAC,GAAG,CAAC,MAAM,EAAE,CAAC,QAAQ,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;IACxD,CAAC;IAED,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAkB,CAAC;IAC3C,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;QACtC,MAAM,MAAM,GAAG,CAAC,CAAC,SAAS,CAAC,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;QACrC,QAAQ,CAAC,GAAG,CAAC,MAAM,EAAE,CAAC,QAAQ,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;IACxD,CAAC;IAED,IAAI,gBAAgB,GAAG,CAAC,CAAC;IACzB,KAAK,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,IAAI,QAAQ,EAAE,CAAC;QACxC,MAAM,MAAM,GAAG,QAAQ,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;QACzC,gBAAgB,IAAI,IAAI,CAAC,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IAC/C,CAAC;IAED,OAAO,CAAC,CAAC,GAAG,gBAAgB,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC;AAClE,CAAC;AAED,mFAAmF;AACnF,MAAM,eAAe,GAAG,GAAG,CAAC;AAE5B;;GAEG;AACH,MAAM,UAAU,kBAAkB,CAChC,aAA4B,EAC5B,WAAmB,EACnB,QAAgB;IAEhB,IAAI,SAAS,GAAuB,IAAI,CAAC;IACzC,IAAI,SAAS,GAAG,CAAC,CAAC;IAElB,KAAK,MAAM,IAAI,IAAI,aAAa,EAAE,CAAC;QACjC,wCAAwC;QACxC,IAAI,IAAI,CAAC,OAAO,CAAC,aAAa,KAAK,QAAQ;YAAE,SAAS;QACtD,IAAI,IAAI,CAAC,MAAM,KAAK,UAAU,IAAI,IAAI,CAAC,MAAM,KAAK,YAAY;YAAE,SAAS;QAEzE,MAAM,KAAK,GAAG,eAAe,CAC3B,IAAI,CAAC,OAAO,CAAC,WAAW,CAAC,WAAW,EAAE,EACtC,WAAW,CAAC,WAAW,EAAE,CAC1B,CAAC;QAEF,IAAI,KAAK,GAAG,SAAS,EAAE,CAAC;YACtB,SAAS,GAAG,KAAK,CAAC;YAClB,SAAS,GAAG,IAAI,CAAC;QACnB,CAAC;IACH,CAAC;IAED,OAAO,SAAS,IAAI,eAAe,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC;AACzD,CAAC;AAED,+DAA+D;AAE/D,SAAS,gBAAgB,CACvB,eAAgC,EAChC,SAAiB,EACjB,IAAY,EACZ,QAAgB,EAChB,QAAgB,EAChB,SAAkB;IAElB,MAAM,GAAG,GAAG,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;IACrC,MAAM,OAAO,GAAG,qBAAqB,CAAC,eAAe,EAAE,SAAS,CAAC,CAAC;IAElE,gFAAgF;IAChF,MAAM,WAAW,GAAG,SAAS,IAAI,SAAS,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,EAAE;QAC3D,CAAC,CAAC,SAAS;QACX,CAAC,CAAC,OAAO,QAAQ,KAAK,eAAe,CAAC,GAAG,CAAC,WAAW,EAAE,CAAC;IAE1D,MAAM,aAAa,GAAkB;QACnC,OAAO,EAAE,MAAM,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,GAAG,CAAC,IAAI,QAAQ,EAAE;QACnD,gBAAgB,EAAE,eAAe,CAAC,WAAW;QAC7C,WAAW;QACX,QAAQ,EAAE,QAAQ;QAClB,QAAQ,EAAE,eAAe,CAAC,SAAS,CAAC,QAAQ;QAC5C,SAAS,EAAE,GAAG;KACf,CAAC;IAEF,OAAO;QACL,EAAE,EAAE,SAAS;QACb,OAAO;QACP,MAAM,EAAE,UAAU,EAAE,mDAAmD;QACvE,SAAS,EAAE,GAAG;QACd,SAAS,EAAE,GAAG;QACd,SAAS,EAAE,CAAC;QACZ,iBAAiB,EAAE,CAAC;QACpB,kBAAkB,EAAE,CAAC;QACrB,cAAc,EAAE,CAAC,aAAa,CAAC;QAC/B,MAAM,EAAE,IAAI;QACZ,cAAc,EAAE,eAAe,CAAC,GAAG,CAAC,WAAW;QAC/C,UAAU,EAAE,eAAe,CAAC,GAAG,CAAC,OAAO;QACvC,iBAAiB,EAAE,CAAC;KACrB,CAAC;AACJ,CAAC;AA2BD,+DAA+D;AAE/D,SAAS,GAAG,CAAC,MAAuB,EAAE,GAAW;IAC/C,IAAI,MAAM,CAAC,KAAK,EAAE,CAAC;QACjB,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;IACpB,CAAC;SAAM,CAAC;QACN,OAAO,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;IACnB,CAAC;AACH,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,UAAU,CAAC,MAAuB;IACtD,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;IAE7B,+BAA+B;IAC/B,IAAI,QAAQ,GAAG,MAAM,CAAC,MAAM;QAC1B,CAAC,CAAC,MAAM,cAAc,CAAC,MAAM,CAAC,WAAW,CAAC;QAC1C,CAAC,CAAC,gBAAgB,EAAE,CAAC;IAEvB,IAAI,MAAM,CAAC,MAAM,EAAE,CAAC;QAClB,MAAM,YAAY,GAAG,MAAM,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC,MAAM,CACjD,CAAC,GAAG,EAAE,GAAG,EAAE,EAAE,CAAC,GAAG,GAAG,GAAG,CAAC,MAAM,EAC9B,CAAC,CACF,CAAC;QACF,GAAG,CAAC,MAAM,EAAE,wBAAwB,YAAY,cAAc,CAAC,CAAC;IAClE,CAAC;IAED,kBAAkB;IAClB,IAAI,KAAK,GAAG,MAAM,CAAC,QAAQ,CAAC;IAE5B,IAAI,MAAM,CAAC,UAAU,IAAI,MAAM,CAAC,UAAU,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACtD,MAAM,SAAS,GAAG,IAAI,GAAG,CAAC,MAAM,CAAC,UAAU,CAAC,CAAC;QAC7C,KAAK,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,SAAS,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,KAAK,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;IACrE,CAAC;IAED,IAAI,MAAM,CAAC,KAAK,KAAK,SAAS,IAAI,MAAM,CAAC,KAAK,GAAG,CAAC,EAAE,CAAC;QACnD,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,MAAM,CAAC,KAAK,CAAC,CAAC;IACvC,CAAC;IAED,GAAG,CAAC,MAAM,EAAE,cAAc,KAAK,CAAC,MAAM,mBAAmB,CAAC,CAAC;IAE3D,8BAA8B;IAC9B,IAAI,kBAAkB,GAAG,CAAC,CAAC;IAC3B,IAAI,mBAAmB,GAAG,CAAC,CAAC;IAC5B,IAAI,YAAY,GAAG,CAAC,CAAC;IACrB,IAAI,aAAa,GAAG,CAAC,CAAC;IACtB,IAAI,cAAc,GAAG,CAAC,CAAC;IACvB,MAAM,mBAAmB,GAA2B,EAAE,CAAC;IAEvD,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,GAAG,GAAG,GAAG,IAAI,CAAC,KAAK,IAAI,IAAI,CAAC,IAAI,EAAE,CAAC;QACzC,GAAG,CAAC,MAAM,EAAE,iBAAiB,GAAG,EAAE,CAAC,CAAC;QAEpC,0BAA0B;QAC1B,IAAI,aAA6B,CAAC;QAClC,IAAI,CAAC;YACH,MAAM,MAAM,GAAG,MAAM,CAAC,aAAa,IAAI,GAAG,CAAC;YAC3C,MAAM,gBAAgB,GAAG,iBAAiB,CAAC,IAAI,CAAC,KAAK,EAAE,IAAI,CAAC,IAAI,EAAE;gBAChE,OAAO,EAAE,EAAE;gBACX,QAAQ,EAAE,IAAI,CAAC,IAAI,CAAC,MAAM,GAAG,EAAE,CAAC;gBAChC,aAAa,EAAE,IAAI;gBACnB,aAAa,EAAE,IAAI;aACpB,CAAC,CAAC;YACH,aAAa,GAAG,gBAAgB,CAAC,KAAK,CAAC,CAAC,EAAE,MAAM,CAAC,CAAC;QACpD,CAAC;QAAC,OAAO,GAAG,EAAE,CAAC;YACb,GAAG,CAAC,MAAM,EAAE,UAAU,GAAG,wBAAwB,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;YAChE,SAAS;QACX,CAAC;QAED,GAAG,CAAC,MAAM,EAAE,WAAW,GAAG,MAAM,aAAa,CAAC,MAAM,cAAc,CAAC,CAAC;QACpE,kBAAkB,IAAI,aAAa,CAAC,MAAM,CAAC;QAE3C,KAAK,MAAM,EAAE,IAAI,aAAa,EAAE,CAAC;YAC/B,uCAAuC;YACvC,IAAI,MAAM,CAAC,MAAM,IAAI,aAAa,CAAC,QAAQ,EAAE,GAAG,EAAE,EAAE,CAAC,QAAQ,CAAC,EAAE,CAAC;gBAC/D,GAAG,CAAC,MAAM,EAAE,gBAAgB,EAAE,CAAC,QAAQ,sBAAsB,CAAC,CAAC;gBAC/D,SAAS;YACX,CAAC;YAED,GAAG,CAAC,MAAM,EAAE,WAAW,EAAE,CAAC,QAAQ,KAAK,EAAE,CAAC,KAAK,EAAE,CAAC,CAAC;YAEnD,2BAA2B;YAC3B,IAAI,UAAU,CAAC;YACf,IAAI,CAAC;gBACH,UAAU,GAAG,YAAY,CAAC,IAAI,CAAC,KAAK,EAAE,IAAI,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;YACvD,CAAC;YAAC,OAAO,GAAG,EAAE,CAAC;gBACb,GAAG,CAAC,MAAM,EAAE,oBAAoB,EAAE,CAAC,QAAQ,MAAM,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;gBAChE,QAAQ,GAAG,eAAe,CAAC,QAAQ,EAAE,GAAG,EAAE,EAAE,CAAC,QAAQ,CAAC,CAAC;gBACvD,MAAM,cAAc,CAAC,MAAM,CAAC,WAAW,EAAE,QAAQ,CAAC,CAAC;gBACnD,SAAS;YACX,CAAC;YAED,6BAA6B;YAC7B,MAAM,eAAe,GAAG,WAAW,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC;YACtD,IAAI,eAAe,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;gBACjC,GAAG,CAAC,MAAM,EAAE,gBAAgB,EAAE,CAAC,QAAQ,wBAAwB,CAAC,CAAC;gBACjE,QAAQ,GAAG,eAAe,CAAC,QAAQ,EAAE,GAAG,EAAE,EAAE,CAAC,QAAQ,CAAC,CAAC;gBACvD,MAAM,cAAc,CAAC,MAAM,CAAC,WAAW,EAAE,QAAQ,CAAC,CAAC;gBACnD,SAAS;YACX,CAAC;YAED,GAAG,CACD,MAAM,EACN,aAAa,eAAe,CAAC,MAAM,mBAAmB,UAAU,CAAC,KAAK,CAAC,MAAM,SAAS,CACvF,CAAC;YACF,mBAAmB,IAAI,eAAe,CAAC,MAAM,CAAC;YAE9C,qEAAqE;YACrE,IAAI,aAAa,GAAG,MAAM,SAAS,CAAC,MAAM,CAAC,WAAW,CAAC,CAAC;YAExD,0BAA0B;YAC1B,KAAK,MAAM,IAAI,IAAI,eAAe,EAAE,CAAC;gBACnC,oDAAoD;gBACpD,MAAM,eAAe,GAAG,EAAE,CAAC,cAAc,CAAC,IAAI,CAC5C,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,IAAI,CAAC,IAAI,CAC5B,CAAC;gBAEF,IAAI,eAAuC,CAAC;gBAC5C,IAAI,CAAC;oBACH,eAAe,GAAG,MAAM,cAAc,CACpC,IAAI,EACJ,GAAG,EACH,EAAE,CAAC,QAAQ,EACX,EAAE,CAAC,KAAK,EACR,eAAe,EAAE,IAAI,CACtB,CAAC;gBACJ,CAAC;gBAAC,OAAO,GAAG,EAAE,CAAC;oBACb,GAAG,CACD,MAAM,EACN,iBAAiB,IAAI,CAAC,IAAI,IAAI,IAAI,CAAC,SAAS,MAAM,MAAM,CAAC,GAAG,CAAC,EAAE,CAChE,CAAC;oBACF,aAAa,EAAE,CAAC;oBAChB,SAAS;gBACX,CAAC;gBAED,IAAI,CAAC,eAAe,EAAE,CAAC;oBACrB,GAAG,CACD,MAAM,EACN,cAAc,IAAI,CAAC,IAAI,IAAI,IAAI,CAAC,SAAS,yBAAyB,CACnE,CAAC;oBACF,aAAa,EAAE,CAAC;oBAChB,SAAS;gBACX,CAAC;gBAED,2CAA2C;gBAC3C,MAAM,SAAS,GAAG,kBAAkB,CAClC,aAAa,EACb,eAAe,CAAC,WAAW,EAC3B,eAAe,CAAC,QAAQ,CACzB,CAAC;gBAEF,IAAI,SAAS,EAAE,CAAC;oBACd,sCAAsC;oBACtC,GAAG,CACD,MAAM,EACN,gBAAgB,eAAe,CAAC,WAAW,oBAAoB,SAAS,CAAC,EAAE,EAAE,CAC9E,CAAC;oBACF,MAAM,UAAU,CAAC,MAAM,CAAC,WAAW,EAAE,SAAS,CAAC,EAAE,EAAE,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC;wBAChE,GAAG,QAAQ;wBACX,iBAAiB,EAAE,CAAC,QAAQ,CAAC,iBAAiB,IAAI,CAAC,CAAC,GAAG,CAAC;wBACxD,SAAS,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;wBACnC,cAAc,EAAE;4BACd,GAAG,QAAQ,CAAC,cAAc;4BAC1B;gCACE,OAAO,EAAE,MAAM,GAAG,CAAC,OAAO,CAAC,GAAG,EAAE,GAAG,CAAC,IAAI,EAAE,CAAC,QAAQ,EAAE;gCACrD,gBAAgB,EAAE,eAAgB,CAAC,WAAW;gCAC9C,WAAW,EAAE,OAAO,EAAE,CAAC,QAAQ,KAAK,eAAgB,CAAC,GAAG,CAAC,WAAW,EAAE;gCACtE,QAAQ,EAAE,IAAI,CAAC,IAAI;gCACnB,QAAQ,EAAE,eAAgB,CAAC,SAAS,CAAC,QAAQ;gCAC7C,SAAS,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;6BACpC;yBACF;qBACF,CAAC,CAAC,CAAC;oBACJ,cAAc,EAAE,CAAC;oBACjB,aAAa,GAAG,MAAM,SAAS,CAAC,MAAM,CAAC,WAAW,CAAC,CAAC;oBACpD,SAAS;gBACX,CAAC;gBAED,wDAAwD;gBACxD,MAAM,SAAS,GAAG,SAAS,IAAI,CAAC,GAAG,EAAE,EAAE,CAAC;gBACxC,qFAAqF;gBACrF,MAAM,SAAS,GAAG,CAAC,GAAG,IAAI,CAAC,OAAO,EAAE,GAAG,IAAI,CAAC,YAAY,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;gBACrE,MAAM,aAAa,GAAG,gBAAgB,CACpC,eAAe,EACf,SAAS,EACT,GAAG,EACH,EAAE,CAAC,QAAQ,EACX,IAAI,CAAC,IAAI,EACT,SAAS,CACV,CAAC;gBAEF,qCAAqC;gBACrC,MAAM,gBAAgB,GAAG,YAAY,CAAC,aAAa,CAAC,CAAC;gBACrD,MAAM,SAAS,GAAgB;oBAC7B,GAAG,aAAa;oBAChB,MAAM,EAAE,gBAAgB,CAAC,MAAM,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,UAAU;oBACzD,iBAAiB,EAAE,gBAAgB;iBACpC,CAAC;gBAEF,MAAM,OAAO,CAAC,MAAM,CAAC,WAAW,EAAE,SAAS,CAAC,CAAC;gBAC7C,aAAa,GAAG,MAAM,SAAS,CAAC,MAAM,CAAC,WAAW,CAAC,CAAC;gBAEpD,IAAI,gBAAgB,CAAC,MAAM,EAAE,CAAC;oBAC5B,YAAY,EAAE,CAAC;oBACf,GAAG,CACD,MAAM,EACN,iBAAiB,SAAS,KAAK,eAAe,CAAC,WAAW,eAAe,gBAAgB,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC,CAAC,WAAW,gBAAgB,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG,CAC/J,CAAC;gBACJ,CAAC;qBAAM,CAAC;oBACN,aAAa,EAAE,CAAC;oBAChB,GAAG,CACD,MAAM,EACN,kBAAkB,SAAS,KAAK,eAAe,CAAC,WAAW,mCAAmC,gBAAgB,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC,CAAC,WAAW,gBAAgB,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG,CACpL,CAAC;gBACJ,CAAC;YACH,CAAC;YAED,iDAAiD;YACjD,QAAQ,GAAG,eAAe,CAAC,QAAQ,EAAE,GAAG,EAAE,EAAE,CAAC,QAAQ,CAAC,CAAC;YACvD,MAAM,cAAc,CAAC,MAAM,CAAC,WAAW,EAAE,QAAQ,CAAC,CAAC;QACrD,CAAC;QAED,GAAG,CAAC,MAAM,EAAE,UAAU,GAAG,EAAE,CAAC,CAAC;IAC/B,CAAC;IAED,kCAAkC;IAClC,MAAM,KAAK,GAAG,MAAM,SAAS,CAAC,MAAM,CAAC,WAAW,CAAC,CAAC;IAClD,MAAM,QAAQ,GAAG,MAAM,SAAS,CAAC,MAAM,CAAC,WAAW,CAAC,CAAC;IACrD,MAAM,WAAW,GAAG;QAClB,KAAK,EAAE,QAAQ,CAAC,MAAM;QACtB,QAAQ,EAAE,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,UAAU,CAAC,CAAC,MAAM;QAChE,QAAQ,EAAE,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,UAAU,CAAC,CAAC,MAAM;KACjE,CAAC;IAEF,MAAM,eAAe,GAAG,CAAC,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS,CAAC,GAAG,MAAM,CAAC;IAE1D,MAAM,MAAM,GAAc;QACxB,SAAS,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;QACnC,KAAK,EAAE,MAAM,CAAC,KAAK;QACnB,YAAY,EAAE,KAAK,CAAC,MAAM;QAC1B,YAAY,EAAE,mBAAmB,EAAE,4CAA4C;QAC/E,eAAe,EAAE,kBAAkB;QACnC,iBAAiB,EAAE,mBAAmB;QACtC,YAAY;QACZ,aAAa;QACb,cAAc;QACd,mBAAmB;QACnB,eAAe;QACf,WAAW;KACZ,CAAC;IAEF,MAAM,UAAU,GAAG,MAAM,aAAa,CAAC,MAAM,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;IAE/D,mBAAmB;IACnB,GAAG,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC;IAChB,GAAG,CAAC,MAAM,EAAE,wCAAwC,CAAC,CAAC;IACtD,GAAG,CAAC,MAAM,EAAE,uBAAuB,CAAC,CAAC;IACrC,GAAG,CAAC,MAAM,EAAE,wCAAwC,CAAC,CAAC;IACtD,GAAG,CAAC,MAAM,EAAE,yBAAyB,KAAK,CAAC,MAAM,EAAE,CAAC,CAAC;IACrD,GAAG,CAAC,MAAM,EAAE,yBAAyB,kBAAkB,EAAE,CAAC,CAAC;IAC3D,GAAG,CAAC,MAAM,EAAE,yBAAyB,mBAAmB,EAAE,CAAC,CAAC;IAC5D,GAAG,CAAC,MAAM,EAAE,yBAAyB,YAAY,EAAE,CAAC,CAAC;IACrD,GAAG,CAAC,MAAM,EAAE,yBAAyB,cAAc,EAAE,CAAC,CAAC;IACvD,GAAG,CAAC,MAAM,EAAE,yBAAyB,aAAa,EAAE,CAAC,CAAC;IACtD,GAAG,CAAC,MAAM,EAAE,yBAAyB,WAAW,CAAC,KAAK,KAAK,WAAW,CAAC,QAAQ,YAAY,CAAC,CAAC;IAC7F,GAAG,CAAC,MAAM,EAAE,yBAAyB,eAAe,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC;IACpE,GAAG,CAAC,MAAM,EAAE,yBAAyB,UAAU,EAAE,CAAC,CAAC;IACnD,GAAG,CAAC,MAAM,EAAE,wCAAwC,CAAC,CAAC;AACxD,CAAC"}

package/dist/lib/rule-harvester/progress.d.ts

@@ -0,0 +1,13 @@
+export type FileState = 'pending' | 'scanned' | 'confirmed' | 'learned';
+export interface RepoProgress {
+    status: 'pending' | 'in_progress' | 'complete';
+    startedAt: string;
+    files: Record<string, FileState>;
+}
+export type HarvestProgress = Record<string, RepoProgress>;
+export declare function loadProgress(dir: string): Promise<HarvestProgress>;
+export declare function saveProgress(dir: string, progress: HarvestProgress): Promise<void>;
+export declare function getFileState(progress: HarvestProgress, repoKey: string, filePath: string): FileState;
+export declare function getRepoStatus(files: Record<string, FileState>): 'pending' | 'in_progress' | 'complete';
+export declare function initRepo(progress: HarvestProgress, repoKey: string): HarvestProgress;
+export declare function setFileState(progress: HarvestProgress, repoKey: string, filePath: string, state: FileState): HarvestProgress;

package/dist/lib/rule-harvester/progress.js

@@ -0,0 +1,50 @@
+import { readFile, writeFile, mkdir } from 'node:fs/promises';
+import { join } from 'node:path';
+export async function loadProgress(dir) {
+    const filePath = join(dir, 'progress.json');
+    try {
+        const content = await readFile(filePath, 'utf-8');
+        return JSON.parse(content);
+    }
+    catch {
+        return {};
+    }
+}
+export async function saveProgress(dir, progress) {
+    await mkdir(dir, { recursive: true });
+    const filePath = join(dir, 'progress.json');
+    await writeFile(filePath, JSON.stringify(progress, null, 2), 'utf-8');
+}
+export function getFileState(progress, repoKey, filePath) {
+    return progress[repoKey]?.files[filePath] ?? 'pending';
+}
+export function getRepoStatus(files) {
+    const states = Object.values(files);
+    if (states.length === 0)
+        return 'pending';
+    if (states.every((s) => s === 'learned'))
+        return 'complete';
+    if (states.every((s) => s === 'pending'))
+        return 'pending';
+    return 'in_progress';
+}
+export function initRepo(progress, repoKey) {
+    if (progress[repoKey])
+        return progress;
+    return {
+        ...progress,
+        [repoKey]: { status: 'pending', startedAt: new Date().toISOString(), files: {} },
+    };
+}
+export function setFileState(progress, repoKey, filePath, state) {
+    const existingRepo = progress[repoKey];
+    const existingFiles = existingRepo?.files ?? {};
+    const updatedFiles = { ...existingFiles, [filePath]: state };
+    const updatedRepo = {
+        status: getRepoStatus(updatedFiles),
+        startedAt: existingRepo?.startedAt ?? new Date().toISOString(),
+        files: updatedFiles,
+    };
+    return { ...progress, [repoKey]: updatedRepo };
+}
+//# sourceMappingURL=progress.js.map

package/dist/lib/rule-harvester/progress.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"progress.js","sourceRoot":"","sources":["../../../src/lib/rule-harvester/progress.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,SAAS,EAAE,KAAK,EAAE,MAAM,kBAAkB,CAAC;AAC9D,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;AAYjC,MAAM,CAAC,KAAK,UAAU,YAAY,CAAC,GAAW;IAC5C,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,EAAE,eAAe,CAAC,CAAC;IAC5C,IAAI,CAAC;QACH,MAAM,OAAO,GAAG,MAAM,QAAQ,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC;QAClD,OAAO,IAAI,CAAC,KAAK,CAAC,OAAO,CAAoB,CAAC;IAChD,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,EAAE,CAAC;IACZ,CAAC;AACH,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,YAAY,CAAC,GAAW,EAAE,QAAyB;IACvE,MAAM,KAAK,CAAC,GAAG,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;IACtC,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,EAAE,eAAe,CAAC,CAAC;IAC5C,MAAM,SAAS,CAAC,QAAQ,EAAE,IAAI,CAAC,SAAS,CAAC,QAAQ,EAAE,IAAI,EAAE,CAAC,CAAC,EAAE,OAAO,CAAC,CAAC;AACxE,CAAC;AAED,MAAM,UAAU,YAAY,CAC1B,QAAyB,EACzB,OAAe,EACf,QAAgB;IAEhB,OAAO,QAAQ,CAAC,OAAO,CAAC,EAAE,KAAK,CAAC,QAAQ,CAAC,IAAI,SAAS,CAAC;AACzD,CAAC;AAED,MAAM,UAAU,aAAa,CAC3B,KAAgC;IAEhC,MAAM,MAAM,GAAG,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;IACpC,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,SAAS,CAAC;IAC1C,IAAI,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,KAAK,SAAS,CAAC;QAAE,OAAO,UAAU,CAAC;IAC5D,IAAI,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,KAAK,SAAS,CAAC;QAAE,OAAO,SAAS,CAAC;IAC3D,OAAO,aAAa,CAAC;AACvB,CAAC;AAED,MAAM,UAAU,QAAQ,CACtB,QAAyB,EACzB,OAAe;IAEf,IAAI,QAAQ,CAAC,OAAO,CAAC;QAAE,OAAO,QAAQ,CAAC;IACvC,OAAO;QACL,GAAG,QAAQ;QACX,CAAC,OAAO,CAAC,EAAE,EAAE,MAAM,EAAE,SAAS,EAAE,SAAS,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,EAAE,KAAK,EAAE,EAAE,EAAE;KACjF,CAAC;AACJ,CAAC;AAED,MAAM,UAAU,YAAY,CAC1B,QAAyB,EACzB,OAAe,EACf,QAAgB,EAChB,KAAgB;IAEhB,MAAM,YAAY,GAAG,QAAQ,CAAC,OAAO,CAAC,CAAC;IACvC,MAAM,aAAa,GAAG,YAAY,EAAE,KAAK,IAAI,EAAE,CAAC;IAChD,MAAM,YAAY,GAA8B,EAAE,GAAG,aAAa,EAAE,CAAC,QAAQ,CAAC,EAAE,KAAK,EAAE,CAAC;IAExF,MAAM,WAAW,GAAiB;QAChC,MAAM,EAAE,aAAa,CAAC,YAAY,CAAC;QACnC,SAAS,EAAE,YAAY,EAAE,SAAS,IAAI,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;QAC9D,KAAK,EAAE,YAAY;KACpB,CAAC;IAEF,OAAO,EAAE,GAAG,QAAQ,EAAE,CAAC,OAAO,CAAC,EAAE,WAAW,EAAE,CAAC;AACjD,CAAC"}