@tonycasey/lisa 2.25.4 → 2.27.0

package/dist/lib/application/services/GitExtractorService.js

@@ -0,0 +1,232 @@
+"use strict";
+/**
+ * Git Extractor Service
+ *
+ * Heuristic-based extraction of structured facts from GitHub PR data.
+ * Fetches PR descriptions, review comments, and linked issues, then
+ * applies pattern matching to extract decision, gotcha, and convention facts.
+ *
+ * Complements the LLM-based CommitEnricher with fast, no-cost extraction.
+ *
+ * Part of LISA-9: Phase 3 Git-Powered Memory Extraction.
+ */
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.createGitExtractorService = createGitExtractorService;
+const HeuristicPatterns_1 = require("../../infrastructure/services/patterns/HeuristicPatterns");
+/** Default maximum PRs to process in one call. */
+const DEFAULT_MAX_PRS = 10;
+/** Confidence level ordering for filtering. */
+const CONFIDENCE_ORDER = {
+    verified: 4,
+    high: 3,
+    medium: 2,
+    low: 1,
+    uncertain: 0,
+};
+/**
+ * Create a Git extractor service.
+ *
+ * @param dataFetcher - GitHub data fetcher for PR/issue data
+ * @param logger - Optional logger
+ * @returns A Git extractor service
+ */
+function createGitExtractorService(dataFetcher, logger) {
+    return {
+        async extractFromPRs(prNumbers, repo, options = {}) {
+            const maxPRs = options.maxPRs ?? DEFAULT_MAX_PRS;
+            const skipPRs = new Set(options.skipPRs ?? []);
+            const extractTypes = options.extractTypes;
+            const minConfidence = options.minConfidence ?? 'low';
+            const minConfidenceLevel = CONFIDENCE_ORDER[minConfidence];
+            // Filter and limit PRs to process
+            const toProcess = prNumbers
+                .filter(n => !skipPRs.has(n))
+                .slice(0, maxPRs);
+            if (toProcess.length === 0) {
+                return {
+                    facts: [],
+                    prsProcessed: 0,
+                    prsSkipped: prNumbers.length,
+                    patternsMatched: 0,
+                };
+            }
+            logger?.debug('Starting PR extraction', {
+                repo,
+                prCount: toProcess.length,
+                extractTypes,
+            });
+            const allFacts = [];
+            let prsProcessed = 0;
+            let prsSkipped = prNumbers.length - toProcess.length;
+            let patternsMatched = 0;
+            // Process each PR
+            for (const prNumber of toProcess) {
+                try {
+                    const pr = await dataFetcher.fetchPR(repo, prNumber);
+                    if (!pr) {
+                        prsSkipped++;
+                        continue;
+                    }
+                    prsProcessed++;
+                    // Extract from PR description
+                    const descFacts = extractFromText(pr.body, 'pr-description', prNumber, undefined, extractTypes);
+                    allFacts.push(...descFacts);
+                    patternsMatched += descFacts.length;
+                    // Extract from review comments
+                    for (const comment of pr.reviewComments) {
+                        const commentFacts = extractFromText(comment.body, 'review-comment', prNumber, undefined, extractTypes);
+                        // Boost confidence for resolved review threads (indicates accepted feedback)
+                        if (comment.isResolved) {
+                            for (const fact of commentFacts) {
+                                if (fact.type === 'gotcha' || fact.type === 'convention') {
+                                    // Resolved feedback is more likely to be valuable
+                                    fact.confidence = boostConfidence(fact.confidence);
+                                }
+                            }
+                        }
+                        // Add file path as tag if available
+                        if (comment.path) {
+                            for (const fact of commentFacts) {
+                                fact.tags.push(comment.path);
+                            }
+                        }
+                        allFacts.push(...commentFacts);
+                        patternsMatched += commentFacts.length;
+                    }
+                    // Extract from linked issues
+                    for (const issue of pr.linkedIssues) {
+                        if (issue.body) {
+                            const issueFacts = extractFromText(issue.body, 'issue-body', prNumber, issue.number, extractTypes);
+                            allFacts.push(...issueFacts);
+                            patternsMatched += issueFacts.length;
+                        }
+                    }
+                    logger?.debug('Extracted facts from PR', {
+                        prNumber,
+                        factCount: allFacts.length,
+                    });
+                }
+                catch (error) {
+                    logger?.warn('Failed to process PR', {
+                        repo,
+                        prNumber,
+                        error: error.message,
+                    });
+                    prsSkipped++;
+                }
+            }
+            // Filter by minimum confidence
+            const filteredFacts = allFacts.filter(f => CONFIDENCE_ORDER[f.confidence] >= minConfidenceLevel);
+            // Deduplicate by similar text
+            const uniqueFacts = deduplicateFacts(filteredFacts);
+            logger?.info('PR extraction complete', {
+                repo,
+                prsProcessed,
+                prsSkipped,
+                patternsMatched,
+                factsExtracted: uniqueFacts.length,
+            });
+            return {
+                facts: uniqueFacts,
+                prsProcessed,
+                prsSkipped,
+                patternsMatched,
+            };
+        },
+    };
+}
+/**
+ * Extract facts from text using heuristic patterns.
+ */
+function extractFromText(text, source, prNumber, issueNumber, filterTypes) {
+    if (!text || text.trim().length === 0) {
+        return [];
+    }
+    const matches = filterTypes
+        ? filterTypes.flatMap(type => (0, HeuristicPatterns_1.extractByFactType)(text, type))
+        : (0, HeuristicPatterns_1.extractPatternMatches)(text);
+    return matches.map(match => ({
+        text: cleanExtractedText(match.text),
+        type: match.factType,
+        source,
+        confidence: match.confidence,
+        tags: extractTags(match.text),
+        prNumber,
+        issueNumber,
+        matchedPattern: match.patternName,
+    }));
+}
+/**
+ * Clean extracted text for storage.
+ */
+function cleanExtractedText(text) {
+    return text
+        .replace(/\s+/g, ' ') // Normalize whitespace
+        .replace(/^\s*[-*]\s*/, '') // Remove leading bullet points
+        .trim();
+}
+/**
+ * Extract potential tags from text (file paths, technologies).
+ */
+function extractTags(text) {
+    const tags = [];
+    // Extract file paths
+    const pathPattern = /(?:^|\s)([\w/-]+\.[a-z]{1,4})(?:\s|$|[,:])/gi;
+    let match;
+    while ((match = pathPattern.exec(text)) !== null) {
+        if (match[1] && !match[1].includes('://')) {
+            tags.push(match[1]);
+        }
+    }
+    // Extract technology names (common ones)
+    const techPatterns = [
+        /\b(typescript|javascript|react|node|graphql|rest)\b/gi,
+        /\b(postgres|mongodb|redis|neo4j)\b/gi,
+        /\b(docker|kubernetes|aws|gcp)\b/gi,
+    ];
+    for (const pattern of techPatterns) {
+        while ((match = pattern.exec(text)) !== null) {
+            if (match[1]) {
+                tags.push(match[1].toLowerCase());
+            }
+        }
+    }
+    return [...new Set(tags)]; // Dedupe
+}
+/**
+ * Boost confidence level by one step.
+ */
+function boostConfidence(confidence) {
+    switch (confidence) {
+        case 'low':
+            return 'medium';
+        case 'medium':
+            return 'high';
+        default:
+            return confidence;
+    }
+}
+/**
+ * Deduplicate facts by similar text.
+ * Converts mutable facts to readonly interface.
+ */
+function deduplicateFacts(facts) {
+    const seen = new Map();
+    for (const fact of facts) {
+        // Normalize text for comparison
+        const normalized = fact.text.toLowerCase().replace(/\s+/g, ' ').trim();
+        // Check for exact or near-duplicate
+        const existing = seen.get(normalized);
+        if (existing) {
+            // Keep the one with higher confidence
+            if (CONFIDENCE_ORDER[fact.confidence] > CONFIDENCE_ORDER[existing.confidence]) {
+                seen.set(normalized, fact);
+            }
+        }
+        else {
+            seen.set(normalized, fact);
+        }
+    }
+    return Array.from(seen.values());
+}
+//# sourceMappingURL=GitExtractorService.js.map

package/dist/lib/application/services/GitExtractorService.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"GitExtractorService.js","sourceRoot":"","sources":["../../../../src/lib/application/services/GitExtractorService.ts"],"names":[],"mappings":";AAAA;;;;;;;;;;GAUG;;AAqDH,8DAoJC;AA5LD,gGAGkE;AAElE,kDAAkD;AAClD,MAAM,eAAe,GAAG,EAAE,CAAC;AAkB3B,+CAA+C;AAC/C,MAAM,gBAAgB,GAAoC;IACxD,QAAQ,EAAE,CAAC;IACX,IAAI,EAAE,CAAC;IACP,MAAM,EAAE,CAAC;IACT,GAAG,EAAE,CAAC;IACN,SAAS,EAAE,CAAC;CACb,CAAC;AAEF;;;;;;GAMG;AACH,SAAgB,yBAAyB,CACvC,WAA+B,EAC/B,MAAgB;IAEhB,OAAO;QACL,KAAK,CAAC,cAAc,CAClB,SAA4B,EAC5B,IAAY,EACZ,UAAuC,EAAE;YAEzC,MAAM,MAAM,GAAG,OAAO,CAAC,MAAM,IAAI,eAAe,CAAC;YACjD,MAAM,OAAO,GAAG,IAAI,GAAG,CAAC,OAAO,CAAC,OAAO,IAAI,EAAE,CAAC,CAAC;YAC/C,MAAM,YAAY,GAAG,OAAO,CAAC,YAAY,CAAC;YAC1C,MAAM,aAAa,GAAG,OAAO,CAAC,aAAa,IAAI,KAAK,CAAC;YACrD,MAAM,kBAAkB,GAAG,gBAAgB,CAAC,aAAa,CAAC,CAAC;YAE3D,kCAAkC;YAClC,MAAM,SAAS,GAAG,SAAS;iBACxB,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;iBAC5B,KAAK,CAAC,CAAC,EAAE,MAAM,CAAC,CAAC;YAEpB,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;gBAC3B,OAAO;oBACL,KAAK,EAAE,EAAE;oBACT,YAAY,EAAE,CAAC;oBACf,UAAU,EAAE,SAAS,CAAC,MAAM;oBAC5B,eAAe,EAAE,CAAC;iBACnB,CAAC;YACJ,CAAC;YAED,MAAM,EAAE,KAAK,CAAC,wBAAwB,EAAE;gBACtC,IAAI;gBACJ,OAAO,EAAE,SAAS,CAAC,MAAM;gBACzB,YAAY;aACb,CAAC,CAAC;YAEH,MAAM,QAAQ,GAA4B,EAAE,CAAC;YAC7C,IAAI,YAAY,GAAG,CAAC,CAAC;YACrB,IAAI,UAAU,GAAG,SAAS,CAAC,MAAM,GAAG,SAAS,CAAC,MAAM,CAAC;YACrD,IAAI,eAAe,GAAG,CAAC,CAAC;YAExB,kBAAkB;YAClB,KAAK,MAAM,QAAQ,IAAI,SAAS,EAAE,CAAC;gBACjC,IAAI,CAAC;oBACH,MAAM,EAAE,GAAG,MAAM,WAAW,CAAC,OAAO,CAAC,IAAI,EAAE,QAAQ,CAAC,CAAC;oBAErD,IAAI,CAAC,EAAE,EAAE,CAAC;wBACR,UAAU,EAAE,CAAC;wBACb,SAAS;oBACX,CAAC;oBAED,YAAY,EAAE,CAAC;oBAEf,8BAA8B;oBAC9B,MAAM,SAAS,GAAG,eAAe,CAC/B,EAAE,CAAC,IAAI,EACP,gBAAgB,EAChB,QAAQ,EACR,SAAS,EACT,YAAY,CACb,CAAC;oBACF,QAAQ,CAAC,IAAI,CAAC,GAAG,SAAS,CAAC,CAAC;oBAC5B,eAAe,IAAI,SAAS,CAAC,MAAM,CAAC;oBAEpC,+BAA+B;oBAC/B,KAAK,MAAM,OAAO,IAAI,EAAE,CAAC,cAAc,EAAE,CAAC;wBACxC,MAAM,YAAY,GAAG,eAAe,CAClC,OAAO,CAAC,IAAI,EACZ,gBAAgB,EAChB,QAAQ,EACR,SAAS,EACT,YAAY,CACb,CAAC;wBAEF,6EAA6E;wBAC7E,IAAI,OAAO,CAAC,UAAU,EAAE,CAAC;4BACvB,KAAK,MAAM,IAAI,IAAI,YAAY,EAAE,CAAC;gCAChC,IAAI,IAAI,CAAC,IAAI,KAAK,QAAQ,IAAI,IAAI,CAAC,IAAI,KAAK,YAAY,EAAE,CAAC;oCACzD,kDAAkD;oCAClD,IAAI,CAAC,UAAU,GAAG,eAAe,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;gCACrD,CAAC;4BACH,CAAC;wBACH,CAAC;wBAED,oCAAoC;wBACpC,IAAI,OAAO,CAAC,IAAI,EAAE,CAAC;4BACjB,KAAK,MAAM,IAAI,IAAI,YAAY,EAAE,CAAC;gCAChC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;4BAC/B,CAAC;wBACH,CAAC;wBAED,QAAQ,CAAC,IAAI,CAAC,GAAG,YAAY,CAAC,CAAC;wBAC/B,eAAe,IAAI,YAAY,CAAC,MAAM,CAAC;oBACzC,CAAC;oBAED,6BAA6B;oBAC7B,KAAK,MAAM,KAAK,IAAI,EAAE,CAAC,YAAY,EAAE,CAAC;wBACpC,IAAI,KAAK,CAAC,IAAI,EAAE,CAAC;4BACf,MAAM,UAAU,GAAG,eAAe,CAChC,KAAK,CAAC,IAAI,EACV,YAAY,EACZ,QAAQ,EACR,KAAK,CAAC,MAAM,EACZ,YAAY,CACb,CAAC;4BACF,QAAQ,CAAC,IAAI,CAAC,GAAG,UAAU,CAAC,CAAC;4BAC7B,eAAe,IAAI,UAAU,CAAC,MAAM,CAAC;wBACvC,CAAC;oBACH,CAAC;oBAED,MAAM,EAAE,KAAK,CAAC,yBAAyB,EAAE;wBACvC,QAAQ;wBACR,SAAS,EAAE,QAAQ,CAAC,MAAM;qBAC3B,CAAC,CAAC;gBACL,CAAC;gBAAC,OAAO,KAAK,EAAE,CAAC;oBACf,MAAM,EAAE,IAAI,CAAC,sBAAsB,EAAE;wBACnC,IAAI;wBACJ,QAAQ;wBACR,KAAK,EAAG,KAAe,CAAC,OAAO;qBAChC,CAAC,CAAC;oBACH,UAAU,EAAE,CAAC;gBACf,CAAC;YACH,CAAC;YAED,+BAA+B;YAC/B,MAAM,aAAa,GAAG,QAAQ,CAAC,MAAM,CACnC,CAAC,CAAC,EAAE,CAAC,gBAAgB,CAAC,CAAC,CAAC,UAAU,CAAC,IAAI,kBAAkB,CAC1D,CAAC;YAEF,8BAA8B;YAC9B,MAAM,WAAW,GAAG,gBAAgB,CAAC,aAAa,CAAC,CAAC;YAEpD,MAAM,EAAE,IAAI,CAAC,wBAAwB,EAAE;gBACrC,IAAI;gBACJ,YAAY;gBACZ,UAAU;gBACV,eAAe;gBACf,cAAc,EAAE,WAAW,CAAC,MAAM;aACnC,CAAC,CAAC;YAEH,OAAO;gBACL,KAAK,EAAE,WAAW;gBAClB,YAAY;gBACZ,UAAU;gBACV,eAAe;aAChB,CAAC;QACJ,CAAC;KACF,CAAC;AACJ,CAAC;AAED;;GAEG;AACH,SAAS,eAAe,CACtB,IAAY,EACZ,MAAkB,EAClB,QAAgB,EAChB,WAAoB,EACpB,WAA0C;IAE1C,IAAI,CAAC,IAAI,IAAI,IAAI,CAAC,IAAI,EAAE,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACtC,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,MAAM,OAAO,GAAG,WAAW;QACzB,CAAC,CAAC,WAAW,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,CAAC,IAAA,qCAAiB,EAAC,IAAI,EAAE,IAAI,CAAC,CAAC;QAC5D,CAAC,CAAC,IAAA,yCAAqB,EAAC,IAAI,CAAC,CAAC;IAEhC,OAAO,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;QAC3B,IAAI,EAAE,kBAAkB,CAAC,KAAK,CAAC,IAAI,CAAC;QACpC,IAAI,EAAE,KAAK,CAAC,QAAQ;QACpB,MAAM;QACN,UAAU,EAAE,KAAK,CAAC,UAA6B;QAC/C,IAAI,EAAE,WAAW,CAAC,KAAK,CAAC,IAAI,CAAC;QAC7B,QAAQ;QACR,WAAW;QACX,cAAc,EAAE,KAAK,CAAC,WAAW;KAClC,CAAC,CAAC,CAAC;AACN,CAAC;AAED;;GAEG;AACH,SAAS,kBAAkB,CAAC,IAAY;IACtC,OAAO,IAAI;SACR,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAE,uBAAuB;SAC7C,OAAO,CAAC,aAAa,EAAE,EAAE,CAAC,CAAE,+BAA+B;SAC3D,IAAI,EAAE,CAAC;AACZ,CAAC;AAED;;GAEG;AACH,SAAS,WAAW,CAAC,IAAY;IAC/B,MAAM,IAAI,GAAa,EAAE,CAAC;IAE1B,qBAAqB;IACrB,MAAM,WAAW,GAAG,8CAA8C,CAAC;IACnE,IAAI,KAAK,CAAC;IACV,OAAO,CAAC,KAAK,GAAG,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;QACjD,IAAI,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,KAAK,CAAC,EAAE,CAAC;YAC1C,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;QACtB,CAAC;IACH,CAAC;IAED,yCAAyC;IACzC,MAAM,YAAY,GAAG;QACnB,uDAAuD;QACvD,sCAAsC;QACtC,mCAAmC;KACpC,CAAC;IAEF,KAAK,MAAM,OAAO,IAAI,YAAY,EAAE,CAAC;QACnC,OAAO,CAAC,KAAK,GAAG,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;YAC7C,IAAI,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC;gBACb,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,WAAW,EAAE,CAAC,CAAC;YACpC,CAAC;QACH,CAAC;IACH,CAAC;IAED,OAAO,CAAC,GAAG,IAAI,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,CAAE,SAAS;AACvC,CAAC;AAED;;GAEG;AACH,SAAS,eAAe,CAAC,UAA2B;IAClD,QAAQ,UAAU,EAAE,CAAC;QACnB,KAAK,KAAK;YACR,OAAO,QAAQ,CAAC;QAClB,KAAK,QAAQ;YACX,OAAO,MAAM,CAAC;QAChB;YACE,OAAO,UAAU,CAAC;IACtB,CAAC;AACH,CAAC;AAED;;;GAGG;AACH,SAAS,gBAAgB,CAAC,KAA8B;IACtD,MAAM,IAAI,GAAG,IAAI,GAAG,EAAiC,CAAC;IAEtD,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,gCAAgC;QAChC,MAAM,UAAU,GAAG,IAAI,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;QAEvE,oCAAoC;QACpC,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,CAAC,UAAU,CAAC,CAAC;QACtC,IAAI,QAAQ,EAAE,CAAC;YACb,sCAAsC;YACtC,IAAI,gBAAgB,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,gBAAgB,CAAC,QAAQ,CAAC,UAAU,CAAC,EAAE,CAAC;gBAC9E,IAAI,CAAC,GAAG,CAAC,UAAU,EAAE,IAAI,CAAC,CAAC;YAC7B,CAAC;QACH,CAAC;aAAM,CAAC;YACN,IAAI,CAAC,GAAG,CAAC,UAAU,EAAE,IAAI,CAAC,CAAC;QAC7B,CAAC;IACH,CAAC;IAED,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC,CAAC;AACnC,CAAC"}

package/dist/lib/application/services/GitIndexingService.d.ts

@@ -0,0 +1,25 @@
+/**
+ * Git Indexing Service
+ *
+ * Saves extracted git facts to memory with appropriate quality tags,
+ * lifecycle metadata, and deduplication.
+ *
+ * Quality defaults:
+ * - source: code-analysis
+ * - confidence: from IHeuristicFact (typically medium)
+ * - lifecycle: project
+ *
+ * Part of LISA-10: Phase 4 Indexing.
+ */
+import type { IGitIndexingService } from '../../domain/interfaces/IGitIndexingService';
+import type { IMemoryService } from '../../domain/interfaces/IMemoryService';
+import type { ILogger } from '../../domain/interfaces/ILogger';
+/**
+ * Create a Git indexing service.
+ *
+ * @param memory - Memory service for saving facts
+ * @param logger - Optional logger
+ * @returns A Git indexing service
+ */
+export declare function createGitIndexingService(memory: IMemoryService, logger?: ILogger): IGitIndexingService;
+//# sourceMappingURL=GitIndexingService.d.ts.map

package/dist/lib/application/services/GitIndexingService.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"GitIndexingService.d.ts","sourceRoot":"","sources":["../../../../src/lib/application/services/GitIndexingService.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;GAYG;AAEH,OAAO,KAAK,EACV,mBAAmB,EAIpB,MAAM,6CAA6C,CAAC;AAErD,OAAO,KAAK,EAAE,cAAc,EAAE,MAAM,wCAAwC,CAAC;AAC7E,OAAO,KAAK,EAAE,OAAO,EAAE,MAAM,iCAAiC,CAAC;AA6E/D;;;;;;GAMG;AACH,wBAAgB,wBAAwB,CACtC,MAAM,EAAE,cAAc,EACtB,MAAM,CAAC,EAAE,OAAO,GACf,mBAAmB,CA4GrB"}

package/dist/lib/application/services/GitIndexingService.js

@@ -0,0 +1,175 @@
+"use strict";
+/**
+ * Git Indexing Service
+ *
+ * Saves extracted git facts to memory with appropriate quality tags,
+ * lifecycle metadata, and deduplication.
+ *
+ * Quality defaults:
+ * - source: code-analysis
+ * - confidence: from IHeuristicFact (typically medium)
+ * - lifecycle: project
+ *
+ * Part of LISA-10: Phase 4 Indexing.
+ */
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.createGitIndexingService = createGitIndexingService;
+const IGitIndexingService_1 = require("../../domain/interfaces/IGitIndexingService");
+const IMemoryQuality_1 = require("../../domain/interfaces/types/IMemoryQuality");
+/** Default similarity threshold for deduplication. */
+const DEFAULT_SIMILARITY_THRESHOLD = 0.85;
+/**
+ * Simple string similarity using Jaccard index on word tokens.
+ * Returns a value between 0 (no similarity) and 1 (identical).
+ */
+function calculateSimilarity(a, b) {
+    const tokenize = (s) => new Set(s.toLowerCase().split(/\s+/).filter(Boolean));
+    const setA = tokenize(a);
+    const setB = tokenize(b);
+    if (setA.size === 0 && setB.size === 0)
+        return 1;
+    if (setA.size === 0 || setB.size === 0)
+        return 0;
+    let intersection = 0;
+    for (const word of setA) {
+        if (setB.has(word))
+            intersection++;
+    }
+    const union = setA.size + setB.size - intersection;
+    return union === 0 ? 0 : intersection / union;
+}
+/**
+ * Build tags for a fact following LISA-10 quality defaults.
+ */
+function buildTags(fact, memoryType) {
+    const tags = [
+        // Quality tags
+        (0, IMemoryQuality_1.resolveSourceTag)('code-analysis'),
+        `confidence:${fact.confidence}`,
+        'lifecycle:project',
+        // Extraction metadata (type:heuristic-extraction needed for checkExtractedPRs dedupe)
+        'type:heuristic-extraction',
+        `factType:${fact.type}`,
+        `factSource:${fact.source}`,
+        'extractionMethod:heuristic',
+    ];
+    // Memory type tag (if mapped)
+    if (memoryType) {
+        tags.push(`memoryType:${memoryType}`);
+    }
+    // Linking tags
+    if (fact.prNumber !== undefined) {
+        tags.push(`pr:${fact.prNumber}`);
+    }
+    if (fact.issueNumber !== undefined) {
+        tags.push(`issue:${fact.issueNumber}`);
+    }
+    if (fact.commitSha) {
+        tags.push(`commit:${fact.commitSha}`);
+    }
+    if (fact.matchedPattern) {
+        tags.push(`pattern:${fact.matchedPattern}`);
+    }
+    // Original tags from extraction (prefixed to avoid collisions)
+    for (const tag of fact.tags) {
+        tags.push(`tag:${tag}`);
+    }
+    return tags;
+}
+/**
+ * Create a Git indexing service.
+ *
+ * @param memory - Memory service for saving facts
+ * @param logger - Optional logger
+ * @returns A Git indexing service
+ */
+function createGitIndexingService(memory, logger) {
+    return {
+        async indexFacts(facts, options = {}) {
+            const { conventionalType, skipDeduplication = false, similarityThreshold = DEFAULT_SIMILARITY_THRESHOLD, } = options;
+            // Map conventional type to memory type
+            const memoryType = conventionalType
+                ? IGitIndexingService_1.COMMIT_TYPE_TO_MEMORY_TYPE[conventionalType]
+                : null;
+            // Check if this commit type should be skipped entirely
+            if (conventionalType && memoryType === null) {
+                logger?.debug('Skipping indexing for commit type', { conventionalType });
+                return {
+                    indexed: 0,
+                    skipped: facts.length,
+                    duplicates: 0,
+                };
+            }
+            let indexed = 0;
+            let skipped = 0;
+            let duplicates = 0;
+            // Load existing facts for deduplication
+            let existingFacts = [];
+            if (!skipDeduplication && facts.length > 0) {
+                try {
+                    // Search for existing facts with similar metadata
+                    existingFacts = await memory.searchFacts('source:code-analysis extractionMethod:heuristic', 100);
+                    logger?.debug('Loaded existing facts for deduplication', {
+                        count: existingFacts.length,
+                    });
+                }
+                catch (error) {
+                    // Continue without deduplication if search fails
+                    logger?.warn('Failed to load existing facts for deduplication', {
+                        error: error.message,
+                    });
+                }
+            }
+            // Index each fact
+            for (const fact of facts) {
+                // Check for duplicates
+                if (!skipDeduplication && existingFacts.length > 0) {
+                    const isDuplicate = existingFacts.some(existing => {
+                        const existingText = existing.fact || existing.text || '';
+                        return calculateSimilarity(existingText, fact.text) >= similarityThreshold;
+                    });
+                    if (isDuplicate) {
+                        logger?.debug('Skipping duplicate fact', {
+                            text: fact.text.substring(0, 50) + '...',
+                        });
+                        duplicates++;
+                        continue;
+                    }
+                }
+                // Build tags
+                const tags = buildTags(fact, memoryType);
+                // Save to memory with lifecycle metadata
+                try {
+                    await memory.addFactWithLifecycle(fact.text, {
+                        lifecycle: 'project',
+                        tags,
+                    });
+                    indexed++;
+                    // Add to existing facts to prevent duplicates within this batch
+                    existingFacts.push({ fact: fact.text });
+                    logger?.debug('Indexed fact', {
+                        type: fact.type,
+                        memoryType,
+                        prNumber: fact.prNumber,
+                    });
+                }
+                catch (error) {
+                    logger?.warn('Failed to index fact', {
+                        error: error.message,
+                        text: fact.text.substring(0, 50) + '...',
+                    });
+                    skipped++;
+                }
+            }
+            logger?.info('Git indexing complete', {
+                indexed,
+                skipped,
+                duplicates,
+                conventionalType,
+                memoryType,
+            });
+            return { indexed, skipped, duplicates };
+        },
+    };
+}
+//# sourceMappingURL=GitIndexingService.js.map

package/dist/lib/application/services/GitIndexingService.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"GitIndexingService.js","sourceRoot":"","sources":["../../../../src/lib/application/services/GitIndexingService.ts"],"names":[],"mappings":";AAAA;;;;;;;;;;;;GAYG;;AA8FH,4DA+GC;AAlMD,qFAAyF;AACzF,iFAAgF;AAEhF,sDAAsD;AACtD,MAAM,4BAA4B,GAAG,IAAI,CAAC;AAE1C;;;GAGG;AACH,SAAS,mBAAmB,CAAC,CAAS,EAAE,CAAS;IAC/C,MAAM,QAAQ,GAAG,CAAC,CAAS,EAAe,EAAE,CAC1C,IAAI,GAAG,CAAC,CAAC,CAAC,WAAW,EAAE,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC;IAExD,MAAM,IAAI,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC;IACzB,MAAM,IAAI,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC;IAEzB,IAAI,IAAI,CAAC,IAAI,KAAK,CAAC,IAAI,IAAI,CAAC,IAAI,KAAK,CAAC;QAAE,OAAO,CAAC,CAAC;IACjD,IAAI,IAAI,CAAC,IAAI,KAAK,CAAC,IAAI,IAAI,CAAC,IAAI,KAAK,CAAC;QAAE,OAAO,CAAC,CAAC;IAEjD,IAAI,YAAY,GAAG,CAAC,CAAC;IACrB,KAAK,MAAM,IAAI,IAAI,IAAI,EAAE,CAAC;QACxB,IAAI,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC;YAAE,YAAY,EAAE,CAAC;IACrC,CAAC;IAED,MAAM,KAAK,GAAG,IAAI,CAAC,IAAI,GAAG,IAAI,CAAC,IAAI,GAAG,YAAY,CAAC;IACnD,OAAO,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,YAAY,GAAG,KAAK,CAAC;AAChD,CAAC;AAED;;GAEG;AACH,SAAS,SAAS,CAChB,IAAoB,EACpB,UAAgC;IAEhC,MAAM,IAAI,GAAa;QACrB,eAAe;QACf,IAAA,iCAAgB,EAAC,eAAe,CAAC;QACjC,cAAc,IAAI,CAAC,UAAU,EAAE;QAC/B,mBAAmB;QAEnB,sFAAsF;QACtF,2BAA2B;QAC3B,YAAY,IAAI,CAAC,IAAI,EAAE;QACvB,cAAc,IAAI,CAAC,MAAM,EAAE;QAC3B,4BAA4B;KAC7B,CAAC;IAEF,8BAA8B;IAC9B,IAAI,UAAU,EAAE,CAAC;QACf,IAAI,CAAC,IAAI,CAAC,cAAc,UAAU,EAAE,CAAC,CAAC;IACxC,CAAC;IAED,eAAe;IACf,IAAI,IAAI,CAAC,QAAQ,KAAK,SAAS,EAAE,CAAC;QAChC,IAAI,CAAC,IAAI,CAAC,MAAM,IAAI,CAAC,QAAQ,EAAE,CAAC,CAAC;IACnC,CAAC;IACD,IAAI,IAAI,CAAC,WAAW,KAAK,SAAS,EAAE,CAAC;QACnC,IAAI,CAAC,IAAI,CAAC,SAAS,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC;IACzC,CAAC;IACD,IAAI,IAAI,CAAC,SAAS,EAAE,CAAC;QACnB,IAAI,CAAC,IAAI,CAAC,UAAU,IAAI,CAAC,SAAS,EAAE,CAAC,CAAC;IACxC,CAAC;IACD,IAAI,IAAI,CAAC,cAAc,EAAE,CAAC;QACxB,IAAI,CAAC,IAAI,CAAC,WAAW,IAAI,CAAC,cAAc,EAAE,CAAC,CAAC;IAC9C,CAAC;IAED,+DAA+D;IAC/D,KAAK,MAAM,GAAG,IAAI,IAAI,CAAC,IAAI,EAAE,CAAC;QAC5B,IAAI,CAAC,IAAI,CAAC,OAAO,GAAG,EAAE,CAAC,CAAC;IAC1B,CAAC;IAED,OAAO,IAAI,CAAC;AACd,CAAC;AAED;;;;;;GAMG;AACH,SAAgB,wBAAwB,CACtC,MAAsB,EACtB,MAAgB;IAEhB,OAAO;QACL,KAAK,CAAC,UAAU,CACd,KAAgC,EAChC,UAA+B,EAAE;YAEjC,MAAM,EACJ,gBAAgB,EAChB,iBAAiB,GAAG,KAAK,EACzB,mBAAmB,GAAG,4BAA4B,GACnD,GAAG,OAAO,CAAC;YAEZ,uCAAuC;YACvC,MAAM,UAAU,GAAyB,gBAAgB;gBACvD,CAAC,CAAC,gDAA0B,CAAC,gBAAgB,CAAC;gBAC9C,CAAC,CAAC,IAAI,CAAC;YAET,uDAAuD;YACvD,IAAI,gBAAgB,IAAI,UAAU,KAAK,IAAI,EAAE,CAAC;gBAC5C,MAAM,EAAE,KAAK,CAAC,mCAAmC,EAAE,EAAE,gBAAgB,EAAE,CAAC,CAAC;gBACzE,OAAO;oBACL,OAAO,EAAE,CAAC;oBACV,OAAO,EAAE,KAAK,CAAC,MAAM;oBACrB,UAAU,EAAE,CAAC;iBACd,CAAC;YACJ,CAAC;YAED,IAAI,OAAO,GAAG,CAAC,CAAC;YAChB,IAAI,OAAO,GAAG,CAAC,CAAC;YAChB,IAAI,UAAU,GAAG,CAAC,CAAC;YAEnB,wCAAwC;YACxC,IAAI,aAAa,GAAuC,EAAE,CAAC;YAC3D,IAAI,CAAC,iBAAiB,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBAC3C,IAAI,CAAC;oBACH,kDAAkD;oBAClD,aAAa,GAAG,MAAM,MAAM,CAAC,WAAW,CACtC,iDAAiD,EACjD,GAAG,CACJ,CAAC;oBACF,MAAM,EAAE,KAAK,CAAC,yCAAyC,EAAE;wBACvD,KAAK,EAAE,aAAa,CAAC,MAAM;qBAC5B,CAAC,CAAC;gBACL,CAAC;gBAAC,OAAO,KAAK,EAAE,CAAC;oBACf,iDAAiD;oBACjD,MAAM,EAAE,IAAI,CAAC,iDAAiD,EAAE;wBAC9D,KAAK,EAAG,KAAe,CAAC,OAAO;qBAChC,CAAC,CAAC;gBACL,CAAC;YACH,CAAC;YAED,kBAAkB;YAClB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;gBACzB,uBAAuB;gBACvB,IAAI,CAAC,iBAAiB,IAAI,aAAa,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;oBACnD,MAAM,WAAW,GAAG,aAAa,CAAC,IAAI,CAAC,QAAQ,CAAC,EAAE;wBAChD,MAAM,YAAY,GAAG,QAAQ,CAAC,IAAI,IAAI,QAAQ,CAAC,IAAI,IAAI,EAAE,CAAC;wBAC1D,OAAO,mBAAmB,CAAC,YAAY,EAAE,IAAI,CAAC,IAAI,CAAC,IAAI,mBAAmB,CAAC;oBAC7E,CAAC,CAAC,CAAC;oBAEH,IAAI,WAAW,EAAE,CAAC;wBAChB,MAAM,EAAE,KAAK,CAAC,yBAAyB,EAAE;4BACvC,IAAI,EAAE,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,EAAE,EAAE,CAAC,GAAG,KAAK;yBACzC,CAAC,CAAC;wBACH,UAAU,EAAE,CAAC;wBACb,SAAS;oBACX,CAAC;gBACH,CAAC;gBAED,aAAa;gBACb,MAAM,IAAI,GAAG,SAAS,CAAC,IAAI,EAAE,UAAU,CAAC,CAAC;gBAEzC,yCAAyC;gBACzC,IAAI,CAAC;oBACH,MAAM,MAAM,CAAC,oBAAoB,CAAC,IAAI,CAAC,IAAI,EAAE;wBAC3C,SAAS,EAAE,SAAS;wBACpB,IAAI;qBACL,CAAC,CAAC;oBACH,OAAO,EAAE,CAAC;oBAEV,gEAAgE;oBAChE,aAAa,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC;oBAExC,MAAM,EAAE,KAAK,CAAC,cAAc,EAAE;wBAC5B,IAAI,EAAE,IAAI,CAAC,IAAI;wBACf,UAAU;wBACV,QAAQ,EAAE,IAAI,CAAC,QAAQ;qBACxB,CAAC,CAAC;gBACL,CAAC;gBAAC,OAAO,KAAK,EAAE,CAAC;oBACf,MAAM,EAAE,IAAI,CAAC,sBAAsB,EAAE;wBACnC,KAAK,EAAG,KAAe,CAAC,OAAO;wBAC/B,IAAI,EAAE,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,EAAE,EAAE,CAAC,GAAG,KAAK;qBACzC,CAAC,CAAC;oBACH,OAAO,EAAE,CAAC;gBACZ,CAAC;YACH,CAAC;YAED,MAAM,EAAE,IAAI,CAAC,uBAAuB,EAAE;gBACpC,OAAO;gBACP,OAAO;gBACP,UAAU;gBACV,gBAAgB;gBAChB,UAAU;aACX,CAAC,CAAC;YAEH,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,UAAU,EAAE,CAAC;QAC1C,CAAC;KACF,CAAC;AACJ,CAAC"}

package/dist/lib/application/services/GitTriageService.d.ts

@@ -0,0 +1,45 @@
+/**
+ * GitTriageService
+ *
+ * Application service that scans git history and scores commits
+ * for "interestingness" to identify which commits warrant API enrichment.
+ *
+ * Part of the Git-Powered Memory feature (LISA-7).
+ */
+import type { IGitClient } from '../../domain/interfaces/IGitClient';
+import type { IGitTriageService, IGitCommitData, IGitCommitStats, IScoredCommit, ITriageResult, ITriageOptions } from '../../domain/interfaces/IGitTriageService';
+export declare class GitTriageService implements IGitTriageService {
+    private readonly git;
+    constructor(git: IGitClient);
+    triage(options?: ITriageOptions): Promise<ITriageResult>;
+    scoreCommit(commit: IGitCommitData, stats: IGitCommitStats | null, tagShas: ReadonlySet<string>, isTagAdjacent?: boolean): IScoredCommit;
+    /**
+     * Parse raw git log commit into domain type.
+     */
+    private parseCommit;
+    /**
+     * Parse raw stat entries into domain type.
+     */
+    private parseStats;
+    /**
+     * Detect interest signals for a commit.
+     */
+    private detectSignals;
+    /**
+     * Calculate interest score from signals.
+     */
+    private calculateScore;
+    /**
+     * Get default "since" date (3 months ago).
+     */
+    private getDefaultSince;
+    /**
+     * Check if a tag name looks like a version tag.
+     */
+    private isVersionTag;
+    /**
+     * Create an empty result (for repos with no commits in range).
+     */
+    private emptyResult;
+}
+//# sourceMappingURL=GitTriageService.d.ts.map

package/dist/lib/application/services/GitTriageService.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"GitTriageService.d.ts","sourceRoot":"","sources":["../../../../src/lib/application/services/GitTriageService.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAEH,OAAO,KAAK,EAAE,UAAU,EAAsC,MAAM,oCAAoC,CAAC;AACzG,OAAO,KAAK,EACV,iBAAiB,EACjB,cAAc,EACd,eAAe,EAEf,aAAa,EACb,aAAa,EACb,cAAc,EAGf,MAAM,2CAA2C,CAAC;AAiFnD,qBAAa,gBAAiB,YAAW,iBAAiB;IAC5C,OAAO,CAAC,QAAQ,CAAC,GAAG;gBAAH,GAAG,EAAE,UAAU;IAEtC,MAAM,CAAC,OAAO,CAAC,EAAE,cAAc,GAAG,OAAO,CAAC,aAAa,CAAC;IAyI9D,WAAW,CACT,MAAM,EAAE,cAAc,EACtB,KAAK,EAAE,eAAe,GAAG,IAAI,EAC7B,OAAO,EAAE,WAAW,CAAC,MAAM,CAAC,EAC5B,aAAa,CAAC,EAAE,OAAO,GACtB,aAAa;IAehB;;OAEG;IACH,OAAO,CAAC,WAAW;IAoBnB;;OAEG;IACH,OAAO,CAAC,UAAU;IAmClB;;OAEG;IACH,OAAO,CAAC,aAAa;IA8CrB;;OAEG;IACH,OAAO,CAAC,cAAc;IAetB;;OAEG;IACH,OAAO,CAAC,eAAe;IAMvB;;OAEG;IACH,OAAO,CAAC,YAAY;IAIpB;;OAEG;IACH,OAAO,CAAC,WAAW;CAapB"}