@wanshi-kg/wanshi 0.1.0

package/dist/core/knowledge/search/KnowledgeGraphSearch.js

@@ -0,0 +1,261 @@
+"use strict";
+var __awaiter = (this && this.__awaiter) || function (thisArg, _arguments, P, generator) {
+    function adopt(value) { return value instanceof P ? value : new P(function (resolve) { resolve(value); }); }
+    return new (P || (P = Promise))(function (resolve, reject) {
+        function fulfilled(value) { try { step(generator.next(value)); } catch (e) { reject(e); } }
+        function rejected(value) { try { step(generator["throw"](value)); } catch (e) { reject(e); } }
+        function step(result) { result.done ? resolve(result.value) : adopt(result.value).then(fulfilled, rejected); }
+        step((generator = generator.apply(thisArg, _arguments || [])).next());
+    });
+};
+var __importDefault = (this && this.__importDefault) || function (mod) {
+    return (mod && mod.__esModule) ? mod : { "default": mod };
+};
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.KnowledgeGraphSearch = void 0;
+const path_1 = __importDefault(require("path"));
+const utils_1 = require("../../../shared/utils");
+const types_1 = require("../../../types");
+// Enhanced search with multiple strategies
+class KnowledgeGraphSearch {
+    constructor(embeddingService, logger) {
+        this.embeddingService = embeddingService;
+        this.logger = logger;
+    }
+    // Strategy 1: Content-based search using file content directly
+    searchByFileContent(fileContent_1, fileName_1, graphs_1) {
+        return __awaiter(this, arguments, void 0, function* (fileContent, fileName, graphs, options = {}) {
+            this.logger.debug(`Searching knowledge graphs for context relevant to: ${fileName}`);
+            // Extract key terms from file content (simple but effective)
+            const keyTerms = this.extractKeyTerms(fileContent, fileName);
+            // Multi-strategy search
+            const results = yield Promise.all([
+                this.searchByKeyTerms(keyTerms, graphs, options),
+                this.searchByFileRelationship(fileName, graphs, options),
+                this.searchByEmbeddings(fileContent, graphs, options)
+            ]);
+            // Merge and rank results
+            return this.mergeSearchResults(results, options.limit || 10);
+        });
+    }
+    // Strategy 2: Extract meaningful terms without LLM
+    extractKeyTerms(content, fileName) {
+        const terms = new Set();
+        // Add filename-based terms
+        const fileBaseName = path_1.default.basename(fileName, path_1.default.extname(fileName));
+        terms.add(fileBaseName);
+        // Extract programming language keywords
+        const codeTerms = this.extractCodeTerms(content);
+        codeTerms.forEach(term => terms.add(term));
+        // Extract camelCase/snake_case identifiers
+        const identifiers = content.match(/[a-zA-Z_][a-zA-Z0-9_]*[A-Z][a-zA-Z0-9_]*/g) || [];
+        identifiers.slice(0, 20).forEach(term => terms.add(term)); // Limit to prevent noise
+        // Extract quoted strings (might be important concepts)
+        const quotes = content.match(/"([^"]+)"|'([^']+)'/g) || [];
+        quotes.slice(0, 10).forEach(quote => {
+            const cleaned = quote.replace(/['"]/g, '');
+            if (cleaned.length > 3 && cleaned.length < 30) {
+                terms.add(cleaned);
+            }
+        });
+        // Extract URLs/imports
+        const imports = content.match(/(?:import|from|require)\s+["']([^"']+)["']/g) || [];
+        imports.forEach(imp => {
+            var _a;
+            const module = (_a = imp.match(/["']([^"']+)["']/)) === null || _a === void 0 ? void 0 : _a[1];
+            if (module)
+                terms.add(module);
+        });
+        return Array.from(terms);
+    }
+    extractCodeTerms(content) {
+        const terms = [];
+        // Function/class names
+        const functions = content.match(/(?:function|class|def|fn)\s+([a-zA-Z_][a-zA-Z0-9_]*)/g) || [];
+        functions.forEach(match => {
+            const name = match.split(/\s+/).pop();
+            if (name)
+                terms.push(name);
+        });
+        // Variable declarations
+        const variables = content.match(/(?:const|let|var|final)\s+([a-zA-Z_][a-zA-Z0-9_]*)/g) || [];
+        variables.forEach(match => {
+            const name = match.split(/\s+/).pop();
+            if (name)
+                terms.push(name);
+        });
+        return terms;
+    }
+    // Strategy 3: Search by extracted key terms
+    searchByKeyTerms(keyTerms, graphs, options) {
+        return __awaiter(this, void 0, void 0, function* () {
+            const allEntities = [];
+            const allRelations = [];
+            const entityNames = new Set();
+            for (const graph of graphs) {
+                for (const entity of graph.entities) {
+                    let relevanceScore = 0;
+                    // Check entity name against key terms
+                    for (const term of keyTerms) {
+                        if ((0, utils_1.jaroWinklerSimilarity)(entity.name, term) > 0.7) {
+                            relevanceScore += 1;
+                        }
+                        if (entity.name.toLowerCase().includes(term.toLowerCase())) {
+                            relevanceScore += 0.5;
+                        }
+                    }
+                    // Check entity type
+                    for (const term of keyTerms) {
+                        if ((0, utils_1.jaroWinklerSimilarity)(entity.entityType, term) > 0.7) {
+                            relevanceScore += 0.8;
+                        }
+                    }
+                    // Check observations if enabled
+                    if (options.includeObservations !== false) {
+                        for (const obs of entity.observations || []) {
+                            for (const term of keyTerms) {
+                                if ((0, types_1.obsText)(obs).toLowerCase().includes(term.toLowerCase())) {
+                                    relevanceScore += 0.3;
+                                }
+                            }
+                        }
+                    }
+                    if (relevanceScore > 0.5) {
+                        allEntities.push(Object.assign(Object.assign({}, entity), { relevanceScore }));
+                        entityNames.add(entity.name);
+                    }
+                }
+                // Add relations between relevant entities
+                for (const relation of graph.relations) {
+                    if (entityNames.has(relation.from) && entityNames.has(relation.to)) {
+                        allRelations.push(relation);
+                    }
+                }
+            }
+            // Sort by relevance and limit
+            allEntities.sort((a, b) => (b.relevanceScore || 0) - (a.relevanceScore || 0));
+            return {
+                entities: allEntities.slice(0, options.limit || 10),
+                relations: allRelations
+            };
+        });
+    }
+    // Strategy 4: Search by file relationship
+    searchByFileRelationship(fileName, graphs, options) {
+        return __awaiter(this, void 0, void 0, function* () {
+            const fileBaseName = path_1.default.basename(fileName, path_1.default.extname(fileName));
+            const fileDirName = path_1.default.dirname(fileName);
+            const relatedEntities = [];
+            const relatedRelations = [];
+            const entityNames = new Set();
+            for (const graph of graphs) {
+                for (const entity of graph.entities) {
+                    let isRelated = false;
+                    // Same file
+                    if (entity.files[0] === fileName) {
+                        isRelated = true;
+                    }
+                    // Same directory
+                    if (entity.files[0] && path_1.default.dirname(entity.files[0]) === fileDirName) {
+                        isRelated = true;
+                    }
+                    // Similar file name
+                    if (entity.files[0] && (0, utils_1.jaroWinklerSimilarity)(path_1.default.basename(entity.files[0], path_1.default.extname(entity.files[0])), fileBaseName) > 0.6) {
+                        isRelated = true;
+                    }
+                    if (isRelated) {
+                        relatedEntities.push(entity);
+                        entityNames.add(entity.name);
+                    }
+                }
+                // Add relations between related entities
+                for (const relation of graph.relations) {
+                    if (entityNames.has(relation.from) && entityNames.has(relation.to)) {
+                        relatedRelations.push(relation);
+                    }
+                }
+            }
+            return {
+                entities: relatedEntities.slice(0, options.limit || 5),
+                relations: relatedRelations
+            };
+        });
+    }
+    // Strategy 5: Embedding-based search (optional, with caching)
+    searchByEmbeddings(content, graphs, options) {
+        return __awaiter(this, void 0, void 0, function* () {
+            try {
+                // Create embedding for file content (truncate if too long)
+                const truncatedContent = content.slice(0, 2000); // Prevent context overflow
+                const contentEmbedding = yield this.embeddingService.embed(truncatedContent);
+                const scoredEntities = [];
+                for (const graph of graphs) {
+                    for (const entity of graph.entities) {
+                        // Create entity text for embedding
+                        const entityText = `${entity.name} ${entity.entityType} ${(entity.observations || []).map(types_1.obsText).join(' ')}`;
+                        const entityEmbedding = yield this.embeddingService.embed(entityText);
+                        const similarity = (0, utils_1.cosineSimilarity)(contentEmbedding, entityEmbedding);
+                        if (similarity > 0.3) { // Threshold for relevance
+                            scoredEntities.push(Object.assign(Object.assign({}, entity), { similarityScore: similarity }));
+                        }
+                    }
+                }
+                // Sort by similarity and take top results
+                scoredEntities.sort((a, b) => b.similarityScore - a.similarityScore);
+                const topEntities = scoredEntities.slice(0, options.limit || 5);
+                const entityNames = new Set(topEntities.map(e => e.name));
+                const relatedRelations = [];
+                for (const graph of graphs) {
+                    for (const relation of graph.relations) {
+                        if (entityNames.has(relation.from) && entityNames.has(relation.to)) {
+                            relatedRelations.push(relation);
+                        }
+                    }
+                }
+                return {
+                    entities: topEntities,
+                    relations: relatedRelations
+                };
+            }
+            catch (error) {
+                this.logger.warn(`Embedding search failed: ${error}`);
+                return { entities: [], relations: [] };
+            }
+        });
+    }
+    // Merge and deduplicate results from multiple strategies
+    mergeSearchResults(results, limit) {
+        const entityMap = new Map();
+        const relationSet = new Set();
+        const relations = [];
+        // Merge entities, preferring higher relevance scores
+        for (const result of results) {
+            for (const entity of result.entities) {
+                const existing = entityMap.get(entity.name);
+                if (!existing || (entity.relevanceScore || 0) > (existing.relevanceScore || 0)) {
+                    entityMap.set(entity.name, entity);
+                }
+            }
+            // Merge relations
+            for (const relation of result.relations) {
+                const relationKey = `${relation.from}->${relation.to}:${JSON.stringify(relation.relationType)}`;
+                if (!relationSet.has(relationKey)) {
+                    relationSet.add(relationKey);
+                    relations.push(relation);
+                }
+            }
+        }
+        // Sort entities by relevance and limit
+        const sortedEntities = Array.from(entityMap.values())
+            .sort((a, b) => (b.relevanceScore || 0) - (a.relevanceScore || 0))
+            .slice(0, limit);
+        const finalEntityNames = new Set(sortedEntities.map(e => e.name));
+        const finalRelations = relations.filter(r => finalEntityNames.has(r.from) && finalEntityNames.has(r.to));
+        return {
+            entities: sortedEntities,
+            relations: finalRelations
+        };
+    }
+}
+exports.KnowledgeGraphSearch = KnowledgeGraphSearch;
+//# sourceMappingURL=KnowledgeGraphSearch.js.map

package/dist/core/knowledge/search/KnowledgeGraphSearch.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"KnowledgeGraphSearch.js","sourceRoot":"","sources":["../../../../src/core/knowledge/search/KnowledgeGraphSearch.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;AAAA,gDAAwB;AACxB,iDAAgF;AAChF,0CAAsH;AAGtH,2CAA2C;AAC3C,MAAa,oBAAoB;IAE/B,YACU,gBAAoC,EACpC,MAAc;QADd,qBAAgB,GAAhB,gBAAgB,CAAoB;QACpC,WAAM,GAAN,MAAM,CAAQ;IAExB,CAAC;IAED,+DAA+D;IACzD,mBAAmB;6DACvB,WAAmB,EACnB,QAAgB,EAChB,MAAwB,EACxB,UAII,EAAE;YAEN,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,uDAAuD,QAAQ,EAAE,CAAC,CAAC;YAErF,6DAA6D;YAC7D,MAAM,QAAQ,GAAG,IAAI,CAAC,eAAe,CAAC,WAAW,EAAE,QAAQ,CAAC,CAAC;YAE7D,wBAAwB;YACxB,MAAM,OAAO,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC;gBAChC,IAAI,CAAC,gBAAgB,CAAC,QAAQ,EAAE,MAAM,EAAE,OAAO,CAAC;gBAChD,IAAI,CAAC,wBAAwB,CAAC,QAAQ,EAAE,MAAM,EAAE,OAAO,CAAC;gBACxD,IAAI,CAAC,kBAAkB,CAAC,WAAW,EAAE,MAAM,EAAE,OAAO,CAAC;aACtD,CAAC,CAAC;YAEH,yBAAyB;YACzB,OAAO,IAAI,CAAC,kBAAkB,CAAC,OAAO,EAAE,OAAO,CAAC,KAAK,IAAI,EAAE,CAAC,CAAC;QAC/D,CAAC;KAAA;IAED,mDAAmD;IAC3C,eAAe,CAAC,OAAe,EAAE,QAAgB;QACvD,MAAM,KAAK,GAAG,IAAI,GAAG,EAAU,CAAC;QAEhC,2BAA2B;QAC3B,MAAM,YAAY,GAAG,cAAI,CAAC,QAAQ,CAAC,QAAQ,EAAE,cAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,CAAC;QACrE,KAAK,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC;QAExB,wCAAwC;QACxC,MAAM,SAAS,GAAG,IAAI,CAAC,gBAAgB,CAAC,OAAO,CAAC,CAAC;QACjD,SAAS,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC;QAE3C,2CAA2C;QAC3C,MAAM,WAAW,GAAG,OAAO,CAAC,KAAK,CAAC,2CAA2C,CAAC,IAAI,EAAE,CAAC;QACrF,WAAW,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,yBAAyB;QAEpF,uDAAuD;QACvD,MAAM,MAAM,GAAG,OAAO,CAAC,KAAK,CAAC,sBAAsB,CAAC,IAAI,EAAE,CAAC;QAC3D,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,KAAK,CAAC,EAAE;YAClC,MAAM,OAAO,GAAG,KAAK,CAAC,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC;YAC3C,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC,IAAI,OAAO,CAAC,MAAM,GAAG,EAAE,EAAE,CAAC;gBAC9C,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;YACrB,CAAC;QACH,CAAC,CAAC,CAAC;QAEH,uBAAuB;QACvB,MAAM,OAAO,GAAG,OAAO,CAAC,KAAK,CAAC,6CAA6C,CAAC,IAAI,EAAE,CAAC;QACnF,OAAO,CAAC,OAAO,CAAC,GAAG,CAAC,EAAE;;YACpB,MAAM,MAAM,GAAG,MAAA,GAAG,CAAC,KAAK,CAAC,kBAAkB,CAAC,0CAAG,CAAC,CAAC,CAAC;YAClD,IAAI,MAAM;gBAAE,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;QAChC,CAAC,CAAC,CAAC;QAEH,OAAO,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;IAC3B,CAAC;IAEO,gBAAgB,CAAC,OAAe;QACtC,MAAM,KAAK,GAAa,EAAE,CAAC;QAE3B,uBAAuB;QACvB,MAAM,SAAS,GAAG,OAAO,CAAC,KAAK,CAAC,uDAAuD,CAAC,IAAI,EAAE,CAAC;QAC/F,SAAS,CAAC,OAAO,CAAC,KAAK,CAAC,EAAE;YACxB,MAAM,IAAI,GAAG,KAAK,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,GAAG,EAAE,CAAC;YACtC,IAAI,IAAI;gBAAE,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAC7B,CAAC,CAAC,CAAC;QAEH,wBAAwB;QACxB,MAAM,SAAS,GAAG,OAAO,CAAC,KAAK,CAAC,qDAAqD,CAAC,IAAI,EAAE,CAAC;QAC7F,SAAS,CAAC,OAAO,CAAC,KAAK,CAAC,EAAE;YACxB,MAAM,IAAI,GAAG,KAAK,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,GAAG,EAAE,CAAC;YACtC,IAAI,IAAI;gBAAE,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAC7B,CAAC,CAAC,CAAC;QAEH,OAAO,KAAK,CAAC;IACf,CAAC;IAED,4CAA4C;IAC9B,gBAAgB,CAC5B,QAAkB,EAClB,MAAwB,EACxB,OAAY;;YAEZ,MAAM,WAAW,GAA4C,EAAE,CAAC;YAChE,MAAM,YAAY,GAAe,EAAE,CAAC;YACpC,MAAM,WAAW,GAAG,IAAI,GAAG,EAAU,CAAC;YAEtC,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;gBAC3B,KAAK,MAAM,MAAM,IAAI,KAAK,CAAC,QAAQ,EAAE,CAAC;oBACpC,IAAI,cAAc,GAAG,CAAC,CAAC;oBAEvB,sCAAsC;oBACtC,KAAK,MAAM,IAAI,IAAI,QAAQ,EAAE,CAAC;wBAC5B,IAAI,IAAA,6BAAqB,EAAC,MAAM,CAAC,IAAI,EAAE,IAAI,CAAC,GAAG,GAAG,EAAE,CAAC;4BACnD,cAAc,IAAI,CAAC,CAAC;wBACtB,CAAC;wBACD,IAAI,MAAM,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,QAAQ,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,EAAE,CAAC;4BAC3D,cAAc,IAAI,GAAG,CAAC;wBACxB,CAAC;oBACH,CAAC;oBAED,oBAAoB;oBACpB,KAAK,MAAM,IAAI,IAAI,QAAQ,EAAE,CAAC;wBAC5B,IAAI,IAAA,6BAAqB,EAAC,MAAM,CAAC,UAAU,EAAE,IAAI,CAAC,GAAG,GAAG,EAAE,CAAC;4BACzD,cAAc,IAAI,GAAG,CAAC;wBACxB,CAAC;oBACH,CAAC;oBAED,gCAAgC;oBAChC,IAAI,OAAO,CAAC,mBAAmB,KAAK,KAAK,EAAE,CAAC;wBAC1C,KAAK,MAAM,GAAG,IAAI,MAAM,CAAC,YAAY,IAAI,EAAE,EAAE,CAAC;4BAC5C,KAAK,MAAM,IAAI,IAAI,QAAQ,EAAE,CAAC;gCAC5B,IAAI,IAAA,eAAO,EAAC,GAAG,CAAC,CAAC,WAAW,EAAE,CAAC,QAAQ,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,EAAE,CAAC;oCAC5D,cAAc,IAAI,GAAG,CAAC;gCACxB,CAAC;4BACH,CAAC;wBACH,CAAC;oBACH,CAAC;oBAED,IAAI,cAAc,GAAG,GAAG,EAAE,CAAC;wBACzB,WAAW,CAAC,IAAI,iCAAM,MAAM,KAAE,cAAc,IAAG,CAAC;wBAChD,WAAW,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;oBAC/B,CAAC;gBACH,CAAC;gBAED,0CAA0C;gBAC1C,KAAK,MAAM,QAAQ,IAAI,KAAK,CAAC,SAAS,EAAE,CAAC;oBACvC,IAAI,WAAW,CAAC,GAAG,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,WAAW,CAAC,GAAG,CAAC,QAAQ,CAAC,EAAE,CAAC,EAAE,CAAC;wBACnE,YAAY,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;oBAC9B,CAAC;gBACH,CAAC;YACH,CAAC;YAED,8BAA8B;YAC9B,WAAW,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,cAAc,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,cAAc,IAAI,CAAC,CAAC,CAAC,CAAC;YAE9E,OAAO;gBACL,QAAQ,EAAE,WAAW,CAAC,KAAK,CAAC,CAAC,EAAE,OAAO,CAAC,KAAK,IAAI,EAAE,CAAC;gBACnD,SAAS,EAAE,YAAY;aACxB,CAAC;QACJ,CAAC;KAAA;IAED,0CAA0C;IAC5B,wBAAwB,CACpC,QAAgB,EAChB,MAAwB,EACxB,OAAY;;YAEZ,MAAM,YAAY,GAAG,cAAI,CAAC,QAAQ,CAAC,QAAQ,EAAE,cAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,CAAC;YACrE,MAAM,WAAW,GAAG,cAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC;YAE3C,MAAM,eAAe,GAAa,EAAE,CAAC;YACrC,MAAM,gBAAgB,GAAe,EAAE,CAAC;YACxC,MAAM,WAAW,GAAG,IAAI,GAAG,EAAU,CAAC;YAEtC,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;gBAC3B,KAAK,MAAM,MAAM,IAAI,KAAK,CAAC,QAAQ,EAAE,CAAC;oBACpC,IAAI,SAAS,GAAG,KAAK,CAAC;oBAEtB,YAAY;oBACZ,IAAI,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,KAAK,QAAQ,EAAE,CAAC;wBACjC,SAAS,GAAG,IAAI,CAAC;oBACnB,CAAC;oBAED,iBAAiB;oBACjB,IAAI,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,cAAI,CAAC,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,WAAW,EAAE,CAAC;wBACrE,SAAS,GAAG,IAAI,CAAC;oBACnB,CAAC;oBAED,oBAAoB;oBACpB,IAAI,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,IAAA,6BAAqB,EAC1C,cAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,cAAI,CAAC,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,EAC7D,YAAY,CACb,GAAG,GAAG,EAAE,CAAC;wBACR,SAAS,GAAG,IAAI,CAAC;oBACnB,CAAC;oBAED,IAAI,SAAS,EAAE,CAAC;wBACd,eAAe,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;wBAC7B,WAAW,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;oBAC/B,CAAC;gBACH,CAAC;gBAED,yCAAyC;gBACzC,KAAK,MAAM,QAAQ,IAAI,KAAK,CAAC,SAAS,EAAE,CAAC;oBACvC,IAAI,WAAW,CAAC,GAAG,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,WAAW,CAAC,GAAG,CAAC,QAAQ,CAAC,EAAE,CAAC,EAAE,CAAC;wBACnE,gBAAgB,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;oBAClC,CAAC;gBACH,CAAC;YACH,CAAC;YAED,OAAO;gBACL,QAAQ,EAAE,eAAe,CAAC,KAAK,CAAC,CAAC,EAAE,OAAO,CAAC,KAAK,IAAI,CAAC,CAAC;gBACtD,SAAS,EAAE,gBAAgB;aAC5B,CAAC;QACJ,CAAC;KAAA;IAED,8DAA8D;IAChD,kBAAkB,CAC9B,OAAe,EACf,MAAwB,EACxB,OAAY;;YAEZ,IAAI,CAAC;gBACH,2DAA2D;gBAC3D,MAAM,gBAAgB,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,CAAC,2BAA2B;gBAC5E,MAAM,gBAAgB,GAAG,MAAM,IAAI,CAAC,gBAAgB,CAAC,KAAK,CAAC,gBAAgB,CAAC,CAAC;gBAE7E,MAAM,cAAc,GAAgD,EAAE,CAAC;gBAEvE,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;oBAC3B,KAAK,MAAM,MAAM,IAAI,KAAK,CAAC,QAAQ,EAAE,CAAC;wBACpC,mCAAmC;wBACnC,MAAM,UAAU,GAAG,GAAG,MAAM,CAAC,IAAI,IAAI,MAAM,CAAC,UAAU,IAAI,CAAC,MAAM,CAAC,YAAY,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,eAAO,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;wBAC/G,MAAM,eAAe,GAAG,MAAM,IAAI,CAAC,gBAAgB,CAAC,KAAK,CAAC,UAAU,CAAC,CAAC;wBAEtE,MAAM,UAAU,GAAG,IAAA,wBAAgB,EAAC,gBAAgB,EAAE,eAAe,CAAC,CAAC;wBAEvE,IAAI,UAAU,GAAG,GAAG,EAAE,CAAC,CAAC,0BAA0B;4BAChD,cAAc,CAAC,IAAI,iCAAM,MAAM,KAAE,eAAe,EAAE,UAAU,IAAG,CAAC;wBAClE,CAAC;oBACH,CAAC;gBACH,CAAC;gBAED,0CAA0C;gBAC1C,cAAc,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,eAAe,GAAG,CAAC,CAAC,eAAe,CAAC,CAAC;gBACrE,MAAM,WAAW,GAAG,cAAc,CAAC,KAAK,CAAC,CAAC,EAAE,OAAO,CAAC,KAAK,IAAI,CAAC,CAAC,CAAC;gBAEhE,MAAM,WAAW,GAAG,IAAI,GAAG,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;gBAC1D,MAAM,gBAAgB,GAAe,EAAE,CAAC;gBAExC,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;oBAC3B,KAAK,MAAM,QAAQ,IAAI,KAAK,CAAC,SAAS,EAAE,CAAC;wBACvC,IAAI,WAAW,CAAC,GAAG,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,WAAW,CAAC,GAAG,CAAC,QAAQ,CAAC,EAAE,CAAC,EAAE,CAAC;4BACnE,gBAAgB,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;wBAClC,CAAC;oBACH,CAAC;gBACH,CAAC;gBAED,OAAO;oBACL,QAAQ,EAAE,WAAW;oBACrB,SAAS,EAAE,gBAAgB;iBAC5B,CAAC;YAEJ,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACf,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,4BAA4B,KAAK,EAAE,CAAC,CAAC;gBACtD,OAAO,EAAE,QAAQ,EAAE,EAAE,EAAE,SAAS,EAAE,EAAE,EAAE,CAAC;YACzC,CAAC;QACH,CAAC;KAAA;IAED,yDAAyD;IACjD,kBAAkB,CAAC,OAAyB,EAAE,KAAa;QACjE,MAAM,SAAS,GAAG,IAAI,GAAG,EAAkB,CAAC;QAC5C,MAAM,WAAW,GAAG,IAAI,GAAG,EAAU,CAAC;QACtC,MAAM,SAAS,GAAe,EAAE,CAAC;QAEjC,qDAAqD;QACrD,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;YAC7B,KAAK,MAAM,MAAM,IAAI,MAAM,CAAC,QAAQ,EAAE,CAAC;gBACrC,MAAM,QAAQ,GAAG,SAAS,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;gBAC5C,IAAI,CAAC,QAAQ,IAAI,CAAE,MAAc,CAAC,cAAc,IAAI,CAAC,CAAC,GAAG,CAAE,QAAgB,CAAC,cAAc,IAAI,CAAC,CAAC,EAAE,CAAC;oBACjG,SAAS,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC;gBACrC,CAAC;YACH,CAAC;YAED,kBAAkB;YAClB,KAAK,MAAM,QAAQ,IAAI,MAAM,CAAC,SAAS,EAAE,CAAC;gBACxC,MAAM,WAAW,GAAG,GAAG,QAAQ,CAAC,IAAI,KAAK,QAAQ,CAAC,EAAE,IAAI,IAAI,CAAC,SAAS,CAAC,QAAQ,CAAC,YAAY,CAAC,EAAE,CAAC;gBAChG,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,WAAW,CAAC,EAAE,CAAC;oBAClC,WAAW,CAAC,GAAG,CAAC,WAAW,CAAC,CAAC;oBAC7B,SAAS,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;gBAC3B,CAAC;YACH,CAAC;QACH,CAAC;QAED,uCAAuC;QACvC,MAAM,cAAc,GAAG,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,EAAE,CAAC;aAClD,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAE,CAAS,CAAC,cAAc,IAAI,CAAC,CAAC,GAAG,CAAE,CAAS,CAAC,cAAc,IAAI,CAAC,CAAC,CAAC;aACnF,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC;QAEnB,MAAM,gBAAgB,GAAG,IAAI,GAAG,CAAC,cAAc,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;QAClE,MAAM,cAAc,GAAG,SAAS,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAC1C,gBAAgB,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,gBAAgB,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAC3D,CAAC;QAEF,OAAO;YACL,QAAQ,EAAE,cAAc;YACxB,SAAS,EAAE,cAAc;SAC1B,CAAC;IACJ,CAAC;CACF;AA/SD,oDA+SC"}

package/dist/core/knowledge/vocabulary.js

@@ -0,0 +1,162 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.DEFAULT_MIXED_DOMAIN_THRESHOLD = exports.DEFAULT_LOW_CONFIDENCE_THRESHOLD = exports.RELATION_TYPE_ESCAPE = exports.ENTITY_TYPE_ESCAPE = exports.BASE_RELATION_TYPES = exports.BASE_ENTITY_TYPES = void 0;
+exports.configureDomainGate = configureDomainGate;
+exports.domainGateThresholds = domainGateThresholds;
+exports.resetDomainGate = resetDomainGate;
+exports.activeDomainClasses = activeDomainClasses;
+exports.domainVocabulary = domainVocabulary;
+exports.allowedEntityTypes = allowedEntityTypes;
+exports.allowedRelationTypes = allowedRelationTypes;
+const NER_DOMAIN_EXAMPLES_1 = require("../processor/classifier/NER_DOMAIN_EXAMPLES");
+/**
+ * Single source of truth for the v5 closed vocabularies (KG-05).
+ *
+ * The base sets mirror the `{{else}}` lists in `templates/v5/system.hbs`
+ * (a test in `vocabulary.test.ts` asserts they stay equal), and the same
+ * `allowed*` helpers feed *both* the Zod enum (`KnowledgeGraphBuilder`) and the
+ * prompt hints (`PromptManager.buildDomainHints`) — so the enum, the hints, and
+ * the gold examples can never drift into the three-way disagreement KG-05
+ * describes (entity enum scoped to a domain, relation enum not).
+ */
+/**
+ * Domain-agnostic entity types, always offered alongside any detected domain's
+ * vocabulary, plus an `other` escape hatch so the model is never forced to
+ * mislabel when nothing fits.
+ */
+exports.BASE_ENTITY_TYPES = [
+    "person", "organization", "location", "role", "event", "time", "metric",
+    "concept", "term", "document", "product", "technology", "standard",
+    "class", "interface", "function", "module", "service", "dependency",
+    "data_structure", "config", "file",
+];
+/** Base relation predicates, always offered alongside any detected domain. */
+exports.BASE_RELATION_TYPES = [
+    "uses", "depends_on", "calls", "implements", "extends", "contains", "part_of",
+    "produces", "consumes", "configures", "references", "defines", "targets",
+    "located_in", "works_at", "member_of", "precedes", "causes", "has_attribute",
+    "related_to",
+];
+/** Escape hatches: keep the model from being forced to invent a one-off label. */
+exports.ENTITY_TYPE_ESCAPE = "other";
+exports.RELATION_TYPE_ESCAPE = "related_to";
+/**
+ * Default minimum top-1 confidence to treat a classification as a domain signal.
+ *
+ * Calibrated for the softmax-probability confidences both classifiers now emit
+ * (S2): ~3× the 1/12 uniform baseline (≈0.083). A clearly-dominant-but-weak class
+ * (e.g. financial prose with no `$` smoking-gun lands at ~0.31, next at ~0.07)
+ * still routes; a flat/uniform distribution (garbage or empty content, p1 ≲ 0.15)
+ * abstains. The pre-S2 value (0.3) was tuned for the old *independent* tanh scores.
+ */
+exports.DEFAULT_LOW_CONFIDENCE_THRESHOLD = 0.25;
+/**
+ * Default max top1−top2 probability gap that still counts as a tie → activate both.
+ * Conservative on purpose: combined with the floor on the *second* class, multi
+ * only fires when two domains genuinely co-dominate (each ≥ floor and within this
+ * margin), which prevents the "doubling the enum" over-activation S2 flagged. The
+ * pre-S2 value (0.2) was tuned for tanh scores.
+ */
+exports.DEFAULT_MIXED_DOMAIN_THRESHOLD = 0.15;
+/**
+ * Run-global domain-gate thresholds (A1). A module singleton — like
+ * `shared/shutdown.ts` — set once per run from
+ * `classifier.{lowConfidenceThreshold,mixedDomainThreshold}` by `ContainerFactory`.
+ *
+ * Keeping it module-global (rather than threading config through every pure gate
+ * function) guarantees the Zod enum path, the prompt hints, the cascade, and the
+ * eval harness all gate on **identical** thresholds — the KG-05 single-source
+ * invariant that would otherwise be one missed caller away from divergence.
+ */
+let lowConfidenceThreshold = exports.DEFAULT_LOW_CONFIDENCE_THRESHOLD;
+let mixedDomainThreshold = exports.DEFAULT_MIXED_DOMAIN_THRESHOLD;
+/** Override the gate thresholds for this run; `undefined` values keep the current. */
+function configureDomainGate(opts) {
+    if (typeof opts.lowConfidence === "number")
+        lowConfidenceThreshold = opts.lowConfidence;
+    if (typeof opts.mixedDomain === "number")
+        mixedDomainThreshold = opts.mixedDomain;
+}
+/** The active gate thresholds (read by `activeDomainClasses` and `getTopClass`). */
+function domainGateThresholds() {
+    return { lowConfidence: lowConfidenceThreshold, mixedDomain: mixedDomainThreshold };
+}
+/** Restore the default thresholds — for tests that exercise a custom gate config. */
+function resetDomainGate() {
+    lowConfidenceThreshold = exports.DEFAULT_LOW_CONFIDENCE_THRESHOLD;
+    mixedDomainThreshold = exports.DEFAULT_MIXED_DOMAIN_THRESHOLD;
+}
+/**
+ * The deterministic confidence cascade (S2/S3): the domain class(es) a
+ * classification activates from a calibrated softmax distribution —
+ *
+ *   - **abstain** (`[]`)      when the top class is below the low-confidence floor;
+ *   - **single** (`[c1]`)     when one class clears the floor and dominates;
+ *   - **multi**  (`[c1, c2]`) when a close second also clears the floor, within
+ *                             the mixed-domain margin.
+ *
+ * Thresholds come from the run-global {@link domainGateThresholds}. This is the
+ * *one* selection both the Zod enum and the prompt hints use, so they can't
+ * disagree about which domain is active. (Phase B inserts an LLM tie-break into
+ * the "close" branch before falling through to multi.)
+ */
+function activeDomainClasses(contentClasses) {
+    if (!contentClasses || contentClasses.length === 0)
+        return [];
+    const sorted = [...contentClasses].sort((a, b) => b.confidence - a.confidence);
+    const top = sorted[0];
+    if (top.confidence < lowConfidenceThreshold)
+        return [];
+    const active = [top.class];
+    if (sorted.length > 1 &&
+        sorted[1].confidence >= lowConfidenceThreshold &&
+        top.confidence - sorted[1].confidence <= mixedDomainThreshold) {
+        active.push(sorted[1].class);
+    }
+    return active;
+}
+/**
+ * The union of primary entity/relation types across the active domain class(es),
+ * in active-class order. Empty when no class clears the threshold.
+ */
+function domainVocabulary(contentClasses) {
+    const entityTypes = [];
+    const relationTypes = [];
+    for (const cls of activeDomainClasses(contentClasses)) {
+        const ner = NER_DOMAIN_EXAMPLES_1.NER_DOMAIN_EXAMPLES[cls];
+        if (!ner)
+            continue;
+        entityTypes.push(...ner.primaryEntityTypes);
+        relationTypes.push(...ner.primaryRelationTypes);
+    }
+    return { entityTypes, relationTypes };
+}
+/**
+ * Closed entity-type set for the Zod enum: active-domain primary types ∪ corpus
+ * glossary types ∪ base set ∪ `other`. Always non-empty (the base set is the
+ * floor), so `entityType` is an enforced enum even with no class and no glossary.
+ */
+function allowedEntityTypes(contentClasses, glossaryTypes = []) {
+    return Array.from(new Set([
+        ...domainVocabulary(contentClasses).entityTypes,
+        ...glossaryTypes,
+        ...exports.BASE_ENTITY_TYPES,
+        exports.ENTITY_TYPE_ESCAPE,
+    ]));
+}
+/**
+ * Closed relation-predicate set for the Zod enum: active-domain primary
+ * predicates ∪ corpus glossary predicates ∪ base set ∪ `related_to`. Unlike the
+ * pre-Phase-2 resolver this DOES include the domain predicates, closing the
+ * KG-05 gap where the relation enum excluded exactly the predicates the hints
+ * and gold examples taught.
+ */
+function allowedRelationTypes(contentClasses, glossaryTypes = []) {
+    return Array.from(new Set([
+        ...domainVocabulary(contentClasses).relationTypes,
+        ...glossaryTypes,
+        ...exports.BASE_RELATION_TYPES,
+        exports.RELATION_TYPE_ESCAPE,
+    ]));
+}
+//# sourceMappingURL=vocabulary.js.map

package/dist/core/knowledge/vocabulary.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"vocabulary.js","sourceRoot":"","sources":["../../../src/core/knowledge/vocabulary.ts"],"names":[],"mappings":";;;AAwEA,kDAMC;AAGD,oDAKC;AAGD,0CAGC;AAgBD,kDAgBC;AAMD,4CAYC;AAOD,gDAYC;AASD,oDAYC;AArLD,qFAAkF;AAElF;;;;;;;;;GASG;AAEH;;;;GAIG;AACU,QAAA,iBAAiB,GAAG;IAC/B,QAAQ,EAAE,cAAc,EAAE,UAAU,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,EAAE,QAAQ;IACvE,SAAS,EAAE,MAAM,EAAE,UAAU,EAAE,SAAS,EAAE,YAAY,EAAE,UAAU;IAClE,OAAO,EAAE,WAAW,EAAE,UAAU,EAAE,QAAQ,EAAE,SAAS,EAAE,YAAY;IACnE,gBAAgB,EAAE,QAAQ,EAAE,MAAM;CACnC,CAAC;AAEF,8EAA8E;AACjE,QAAA,mBAAmB,GAAG;IACjC,MAAM,EAAE,YAAY,EAAE,OAAO,EAAE,YAAY,EAAE,SAAS,EAAE,UAAU,EAAE,SAAS;IAC7E,UAAU,EAAE,UAAU,EAAE,YAAY,EAAE,YAAY,EAAE,SAAS,EAAE,SAAS;IACxE,YAAY,EAAE,UAAU,EAAE,WAAW,EAAE,UAAU,EAAE,QAAQ,EAAE,eAAe;IAC5E,YAAY;CACb,CAAC;AAEF,kFAAkF;AACrE,QAAA,kBAAkB,GAAG,OAAO,CAAC;AAC7B,QAAA,oBAAoB,GAAG,YAAY,CAAC;AAEjD;;;;;;;;GAQG;AACU,QAAA,gCAAgC,GAAG,IAAI,CAAC;AAErD;;;;;;GAMG;AACU,QAAA,8BAA8B,GAAG,IAAI,CAAC;AAEnD;;;;;;;;;GASG;AACH,IAAI,sBAAsB,GAAG,wCAAgC,CAAC;AAC9D,IAAI,oBAAoB,GAAG,sCAA8B,CAAC;AAE1D,sFAAsF;AACtF,SAAgB,mBAAmB,CAAC,IAGnC;IACC,IAAI,OAAO,IAAI,CAAC,aAAa,KAAK,QAAQ;QAAE,sBAAsB,GAAG,IAAI,CAAC,aAAa,CAAC;IACxF,IAAI,OAAO,IAAI,CAAC,WAAW,KAAK,QAAQ;QAAE,oBAAoB,GAAG,IAAI,CAAC,WAAW,CAAC;AACpF,CAAC;AAED,oFAAoF;AACpF,SAAgB,oBAAoB;IAIlC,OAAO,EAAE,aAAa,EAAE,sBAAsB,EAAE,WAAW,EAAE,oBAAoB,EAAE,CAAC;AACtF,CAAC;AAED,qFAAqF;AACrF,SAAgB,eAAe;IAC7B,sBAAsB,GAAG,wCAAgC,CAAC;IAC1D,oBAAoB,GAAG,sCAA8B,CAAC;AACxD,CAAC;AAED;;;;;;;;;;;;;GAaG;AACH,SAAgB,mBAAmB,CACjC,cAAuC;IAEvC,IAAI,CAAC,cAAc,IAAI,cAAc,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAC9D,MAAM,MAAM,GAAG,CAAC,GAAG,cAAc,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,UAAU,GAAG,CAAC,CAAC,UAAU,CAAC,CAAC;IAC/E,MAAM,GAAG,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC;IACtB,IAAI,GAAG,CAAC,UAAU,GAAG,sBAAsB;QAAE,OAAO,EAAE,CAAC;IACvD,MAAM,MAAM,GAAmB,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;IAC3C,IACE,MAAM,CAAC,MAAM,GAAG,CAAC;QACjB,MAAM,CAAC,CAAC,CAAC,CAAC,UAAU,IAAI,sBAAsB;QAC9C,GAAG,CAAC,UAAU,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,UAAU,IAAI,oBAAoB,EAC7D,CAAC;QACD,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC;IAC/B,CAAC;IACD,OAAO,MAAM,CAAC;AAChB,CAAC;AAED;;;GAGG;AACH,SAAgB,gBAAgB,CAC9B,cAAuC;IAEvC,MAAM,WAAW,GAAa,EAAE,CAAC;IACjC,MAAM,aAAa,GAAa,EAAE,CAAC;IACnC,KAAK,MAAM,GAAG,IAAI,mBAAmB,CAAC,cAAc,CAAC,EAAE,CAAC;QACtD,MAAM,GAAG,GAAG,yCAAmB,CAAC,GAAG,CAAC,CAAC;QACrC,IAAI,CAAC,GAAG;YAAE,SAAS;QACnB,WAAW,CAAC,IAAI,CAAC,GAAG,GAAG,CAAC,kBAAkB,CAAC,CAAC;QAC5C,aAAa,CAAC,IAAI,CAAC,GAAG,GAAG,CAAC,oBAAoB,CAAC,CAAC;IAClD,CAAC;IACD,OAAO,EAAE,WAAW,EAAE,aAAa,EAAE,CAAC;AACxC,CAAC;AAED;;;;GAIG;AACH,SAAgB,kBAAkB,CAChC,cAAuC,EACvC,gBAA0B,EAAE;IAE5B,OAAO,KAAK,CAAC,IAAI,CACf,IAAI,GAAG,CAAC;QACN,GAAG,gBAAgB,CAAC,cAAc,CAAC,CAAC,WAAW;QAC/C,GAAG,aAAa;QAChB,GAAG,yBAAiB;QACpB,0BAAkB;KACnB,CAAC,CACH,CAAC;AACJ,CAAC;AAED;;;;;;GAMG;AACH,SAAgB,oBAAoB,CAClC,cAAuC,EACvC,gBAA0B,EAAE;IAE5B,OAAO,KAAK,CAAC,IAAI,CACf,IAAI,GAAG,CAAC;QACN,GAAG,gBAAgB,CAAC,cAAc,CAAC,CAAC,aAAa;QACjD,GAAG,aAAa;QAChB,GAAG,2BAAmB;QACtB,4BAAoB;KACrB,CAAC,CACH,CAAC;AACJ,CAAC"}

package/dist/core/llm/EmbeddingService.js

@@ -0,0 +1,113 @@
+"use strict";
+var __awaiter = (this && this.__awaiter) || function (thisArg, _arguments, P, generator) {
+    function adopt(value) { return value instanceof P ? value : new P(function (resolve) { resolve(value); }); }
+    return new (P || (P = Promise))(function (resolve, reject) {
+        function fulfilled(value) { try { step(generator.next(value)); } catch (e) { reject(e); } }
+        function rejected(value) { try { step(generator["throw"](value)); } catch (e) { reject(e); } }
+        function step(result) { result.done ? resolve(result.value) : adopt(result.value).then(fulfilled, rejected); }
+        step((generator = generator.apply(thisArg, _arguments || [])).next());
+    });
+};
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.EmbeddingService = void 0;
+const ollama_1 = require("ollama");
+const embeddingUtils_1 = require("./embeddingUtils");
+/**
+ * Service for generating text embeddings
+ */
+class EmbeddingService {
+    constructor(options, logger) {
+        this.options = Object.assign({ maxInputChars: embeddingUtils_1.DEFAULT_MAX_EMBED_CHARS }, options);
+        this.cache = new Map();
+        this.logger = logger;
+        this.ollama = new ollama_1.Ollama({ host: options.host });
+    }
+    /** Trim input to the configured char budget so long observations/entities
+     * don't exceed the embedding model's context window. */
+    truncate(text) {
+        const max = this.options.maxInputChars;
+        if (max && max > 0 && text.length > max) {
+            this.logger.debug(`Truncating embedding input from ${text.length} to ${max} chars`);
+            return text.slice(0, max);
+        }
+        return text;
+    }
+    /**
+     * Generate embeddings for a single text
+     */
+    embed(text) {
+        return __awaiter(this, void 0, void 0, function* () {
+            // Check cache first
+            const cached = this.cache.get(text);
+            if (cached) {
+                this.logger.debug(`Using cached embedding for text: ${text.substring(0, 50)}...`);
+                return cached;
+            }
+            this.logger.debug(`Generating embedding for text: ${text.substring(0, 50)}...`);
+            const embedding = yield this.embedWithFallback(this.truncate(text));
+            this.cache.set(text, embedding);
+            return embedding;
+        });
+    }
+    /**
+     * Embed `input`, halving it and retrying if the model rejects it for being
+     * too long. The char cap can't know any given model's exact token limit, so
+     * this self-corrects for dense content (JSON/code) on small-context models.
+     */
+    embedWithFallback(input) {
+        return __awaiter(this, void 0, void 0, function* () {
+            let current = input;
+            while (true) {
+                try {
+                    const response = yield this.ollama.embeddings({
+                        model: this.options.model,
+                        prompt: current,
+                    });
+                    return response.embedding;
+                }
+                catch (error) {
+                    if ((0, embeddingUtils_1.isContextLengthError)(error) && current.length > embeddingUtils_1.MIN_EMBED_CHARS) {
+                        const next = Math.max(embeddingUtils_1.MIN_EMBED_CHARS, Math.floor(current.length / 2));
+                        this.logger.warn(`Embedding input too long (${current.length} chars); retrying with ${next}`);
+                        current = current.slice(0, next);
+                        continue;
+                    }
+                    this.logger.error(`Failed to generate embedding: ${error}`);
+                    throw new Error(`Failed to generate embedding: ${error}`);
+                }
+            }
+        });
+    }
+    /**
+     * Generate embeddings for multiple texts
+     */
+    embedBatch(texts) {
+        return __awaiter(this, void 0, void 0, function* () {
+            this.logger.debug(`Generating embeddings for ${texts.length} texts`);
+            const embeddings = [];
+            // Process in batches to avoid overwhelming the model
+            const batchSize = 10;
+            for (let i = 0; i < texts.length; i += batchSize) {
+                const batch = texts.slice(i, i + batchSize);
+                const batchEmbeddings = yield Promise.all(batch.map(text => this.embed(text)));
+                embeddings.push(...batchEmbeddings);
+            }
+            return embeddings;
+        });
+    }
+    /**
+     * Clear the embedding cache
+     */
+    clearCache() {
+        this.cache.clear();
+        this.logger.debug('Embedding cache cleared');
+    }
+    /**
+     * Get cache size
+     */
+    getCacheSize() {
+        return this.cache.size;
+    }
+}
+exports.EmbeddingService = EmbeddingService;
+//# sourceMappingURL=EmbeddingService.js.map

package/dist/core/llm/EmbeddingService.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"EmbeddingService.js","sourceRoot":"","sources":["../../../src/core/llm/EmbeddingService.ts"],"names":[],"mappings":";;;;;;;;;;;;AAAA,mCAAgC;AAGhC,qDAAkG;AAUlG;;GAEG;AACH,MAAa,gBAAgB;IAM3B,YAAY,OAAyB,EAAE,MAAc;QACnD,IAAI,CAAC,OAAO,mBAAK,aAAa,EAAE,wCAAuB,IAAK,OAAO,CAAE,CAAC;QACtE,IAAI,CAAC,KAAK,GAAG,IAAI,GAAG,EAAE,CAAC;QACvB,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC;QACrB,IAAI,CAAC,MAAM,GAAG,IAAI,eAAM,CAAC,EAAE,IAAI,EAAE,OAAO,CAAC,IAAI,EAAE,CAAC,CAAC;IACnD,CAAC;IAED;4DACwD;IAChD,QAAQ,CAAC,IAAY;QAC3B,MAAM,GAAG,GAAG,IAAI,CAAC,OAAO,CAAC,aAAa,CAAC;QACvC,IAAI,GAAG,IAAI,GAAG,GAAG,CAAC,IAAI,IAAI,CAAC,MAAM,GAAG,GAAG,EAAE,CAAC;YACxC,IAAI,CAAC,MAAM,CAAC,KAAK,CACf,mCAAmC,IAAI,CAAC,MAAM,OAAO,GAAG,QAAQ,CACjE,CAAC;YACF,OAAO,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC;QAC5B,CAAC;QACD,OAAO,IAAI,CAAC;IACd,CAAC;IAED;;OAEG;IACG,KAAK,CAAC,IAAY;;YACtB,oBAAoB;YACpB,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;YACpC,IAAI,MAAM,EAAE,CAAC;gBACX,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,oCAAoC,IAAI,CAAC,SAAS,CAAC,CAAC,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC;gBAClF,OAAO,MAAM,CAAC;YAChB,CAAC;YAED,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,kCAAkC,IAAI,CAAC,SAAS,CAAC,CAAC,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC;YAEhF,MAAM,SAAS,GAAG,MAAM,IAAI,CAAC,iBAAiB,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC;YACpE,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,EAAE,SAAS,CAAC,CAAC;YAChC,OAAO,SAAS,CAAC;QACnB,CAAC;KAAA;IAED;;;;OAIG;IACW,iBAAiB,CAAC,KAAa;;YAC3C,IAAI,OAAO,GAAG,KAAK,CAAC;YACpB,OAAO,IAAI,EAAE,CAAC;gBACZ,IAAI,CAAC;oBACH,MAAM,QAAQ,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,UAAU,CAAC;wBAC5C,KAAK,EAAE,IAAI,CAAC,OAAO,CAAC,KAAK;wBACzB,MAAM,EAAE,OAAO;qBAChB,CAAC,CAAC;oBACH,OAAO,QAAQ,CAAC,SAAS,CAAC;gBAC5B,CAAC;gBAAC,OAAO,KAAK,EAAE,CAAC;oBACf,IAAI,IAAA,qCAAoB,EAAC,KAAK,CAAC,IAAI,OAAO,CAAC,MAAM,GAAG,gCAAe,EAAE,CAAC;wBACpE,MAAM,IAAI,GAAG,IAAI,CAAC,GAAG,CAAC,gCAAe,EAAE,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC;wBACvE,IAAI,CAAC,MAAM,CAAC,IAAI,CACd,6BAA6B,OAAO,CAAC,MAAM,0BAA0B,IAAI,EAAE,CAC5E,CAAC;wBACF,OAAO,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC;wBACjC,SAAS;oBACX,CAAC;oBACD,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,iCAAiC,KAAK,EAAE,CAAC,CAAC;oBAC5D,MAAM,IAAI,KAAK,CAAC,iCAAiC,KAAK,EAAE,CAAC,CAAC;gBAC5D,CAAC;YACH,CAAC;QACH,CAAC;KAAA;IAED;;OAEG;IACG,UAAU,CAAC,KAAe;;YAC9B,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,6BAA6B,KAAK,CAAC,MAAM,QAAQ,CAAC,CAAC;YAErE,MAAM,UAAU,GAAe,EAAE,CAAC;YAElC,qDAAqD;YACrD,MAAM,SAAS,GAAG,EAAE,CAAC;YACrB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,IAAI,SAAS,EAAE,CAAC;gBACjD,MAAM,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,SAAS,CAAC,CAAC;gBAC5C,MAAM,eAAe,GAAG,MAAM,OAAO,CAAC,GAAG,CACvC,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CACpC,CAAC;gBACF,UAAU,CAAC,IAAI,CAAC,GAAG,eAAe,CAAC,CAAC;YACtC,CAAC;YAED,OAAO,UAAU,CAAC;QACpB,CAAC;KAAA;IAED;;OAEG;IACH,UAAU;QACR,IAAI,CAAC,KAAK,CAAC,KAAK,EAAE,CAAC;QACnB,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,yBAAyB,CAAC,CAAC;IAC/C,CAAC;IAED;;OAEG;IACH,YAAY;QACV,OAAO,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC;IACzB,CAAC;CACF;AA5GD,4CA4GC"}