src-mcp 1.0.1

package/dist/core/embeddings/callgraph.js

@@ -0,0 +1,415 @@
+/**
+ * Call graph extraction and storage
+ *
+ * Extracts function call relationships from code using tree-sitter
+ * to build a graph showing which functions call which.
+ *
+ * Features:
+ * - Persistent caching in .src-index/call-graph.json
+ * - Hash-based invalidation for changed files
+ */
+import { Query } from "web-tree-sitter";
+import * as fs from "node:fs";
+import * as path from "node:path";
+import * as crypto from "node:crypto";
+import { parseCode } from "@core/parser";
+import { extractSymbols } from "@core/symbols";
+import { registerCache } from "@core/utils";
+import { logger } from "@utils";
+/**
+ * Compute SHA-256 hash of content
+ */
+function computeHash(content) {
+    return crypto.createHash("sha256").update(content).digest("hex").slice(0, 16);
+}
+/**
+ * Get call graph cache path for a directory
+ */
+function getCachePath(directory) {
+    return path.join(directory, ".src-index", "call-graph.json");
+}
+/**
+ * Save call graph to persistent cache
+ */
+function saveCallGraphCache(directory, graph, fileHashes) {
+    try {
+        const cachePath = getCachePath(directory);
+        const cacheDir = path.dirname(cachePath);
+        // Ensure cache directory exists
+        if (!fs.existsSync(cacheDir)) {
+            fs.mkdirSync(cacheDir, { recursive: true });
+        }
+        const serialized = {
+            nodes: Object.fromEntries(graph.nodes),
+            files: graph.files,
+            edgeCount: graph.edgeCount,
+            fileHashes,
+            timestamp: Date.now(),
+        };
+        fs.writeFileSync(cachePath, JSON.stringify(serialized), "utf-8");
+        logger.debug(`Call graph cache saved: ${String(graph.nodes.size)} nodes`);
+    }
+    catch {
+        // Silently ignore cache save errors (directory not writable, etc.)
+        logger.debug("Call graph cache save skipped: directory not writable");
+    }
+}
+/**
+ * Load call graph from persistent cache if valid
+ */
+function loadCallGraphCache(directory, currentHashes) {
+    const cachePath = getCachePath(directory);
+    if (!fs.existsSync(cachePath)) {
+        return null;
+    }
+    try {
+        const content = fs.readFileSync(cachePath, "utf-8");
+        const cached = JSON.parse(content);
+        // Validate hashes - check if any file has changed
+        const cachedFiles = new Set(Object.keys(cached.fileHashes));
+        const currentFiles = new Set(Object.keys(currentHashes));
+        // Check for added or removed files
+        if (cachedFiles.size !== currentFiles.size) {
+            logger.debug("Call graph cache invalid: file count changed");
+            return null;
+        }
+        // Check for modified files
+        for (const [filePath, hash] of Object.entries(currentHashes)) {
+            if (cached.fileHashes[filePath] !== hash) {
+                logger.debug(`Call graph cache invalid: ${filePath} changed`);
+                return null;
+            }
+        }
+        // Cache is valid - restore the Map
+        const nodes = new Map(Object.entries(cached.nodes));
+        logger.debug(`Call graph cache loaded: ${String(nodes.size)} nodes`);
+        return {
+            nodes,
+            files: cached.files,
+            edgeCount: cached.edgeCount,
+        };
+    }
+    catch (error) {
+        logger.debug(`Failed to load call graph cache: ${error instanceof Error ? error.message : String(error)}`);
+        return null;
+    }
+}
+const callGraphCache = new Map();
+/**
+ * Clear the call graph cache
+ */
+export function clearCallGraphCache() {
+    callGraphCache.clear();
+}
+// Register cache for centralized clearing
+registerCache("embeddings:callGraphCache", clearCallGraphCache);
+/**
+ * Extract function calls from a tree-sitter node
+ */
+function extractCallsFromTree(tree, languageInstance, language) {
+    const callsBySymbol = new Map();
+    // Query patterns for function calls in different languages
+    const callPatterns = {
+        typescript: `
+      (call_expression
+        function: [(identifier) @callee
+                   (member_expression property: (property_identifier) @callee)]
+        arguments: (arguments) @args)
+    `,
+        javascript: `
+      (call_expression
+        function: [(identifier) @callee
+                   (member_expression property: (property_identifier) @callee)]
+        arguments: (arguments) @args)
+    `,
+        python: `
+      (call
+        function: [(identifier) @callee
+                   (attribute attribute: (identifier) @callee)]
+        arguments: (argument_list) @args)
+    `,
+        go: `
+      (call_expression
+        function: [(identifier) @callee
+                   (selector_expression field: (field_identifier) @callee)]
+        arguments: (argument_list) @args)
+    `,
+    };
+    const pattern = callPatterns[language];
+    if (!pattern) {
+        return callsBySymbol;
+    }
+    try {
+        const query = new Query(languageInstance, pattern);
+        const matches = query.matches(tree.rootNode);
+        // Extract callee names from matches
+        const callCaptures = [];
+        for (const match of matches) {
+            for (const capture of match.captures) {
+                if (capture.name === "callee") {
+                    callCaptures.push({
+                        callee: capture.node.text,
+                        position: {
+                            line: capture.node.startPosition.row + 1,
+                            column: capture.node.startPosition.column,
+                            offset: capture.node.startIndex,
+                        },
+                    });
+                }
+            }
+        }
+        // For now, store all calls without symbol association
+        // A more sophisticated approach would track which symbol contains each call
+        if (callCaptures.length > 0) {
+            callsBySymbol.set("__global__", callCaptures.map((c) => ({
+                callee: c.callee,
+                position: c.position,
+            })));
+        }
+    }
+    catch (error) {
+        logger.debug(`Failed to extract calls for ${language}: ${error instanceof Error ? error.message : String(error)}`);
+    }
+    return callsBySymbol;
+}
+/**
+ * Associate calls with their containing symbols
+ */
+function associateCallsWithSymbols(symbols, allCalls) {
+    const callsBySymbol = new Map();
+    // Sort symbols by start offset for efficient lookup
+    const sortedSymbols = [...symbols]
+        .filter((s) => s.type === "function" || s.type === "method")
+        .sort((a, b) => a.start.offset - b.start.offset);
+    for (const call of allCalls) {
+        // Find the symbol that contains this call
+        let containingSymbol = null;
+        for (const symbol of sortedSymbols) {
+            if (call.position.offset >= symbol.start.offset &&
+                call.position.offset <= symbol.end.offset) {
+                containingSymbol = symbol;
+            }
+            else if (call.position.offset < symbol.start.offset) {
+                // Calls are sorted by position, so we can break early
+                break;
+            }
+        }
+        const symbolName = containingSymbol?.name ?? "__global__";
+        const existing = callsBySymbol.get(symbolName) ?? [];
+        existing.push(call);
+        callsBySymbol.set(symbolName, existing);
+    }
+    return callsBySymbol;
+}
+/**
+ * Analyze a file and extract call graph data
+ */
+export async function analyzeFileForCallGraph(filePath, content) {
+    // Check cache
+    const cached = callGraphCache.get(filePath);
+    if (cached) {
+        return cached;
+    }
+    try {
+        const parseResult = await parseCode(content, { filePath });
+        const { symbols } = extractSymbols(parseResult.tree, parseResult.languageInstance, parseResult.language);
+        const callsMap = extractCallsFromTree(parseResult.tree, parseResult.languageInstance, parseResult.language);
+        // Get all calls and associate with symbols
+        const allCalls = callsMap.get("__global__") ?? [];
+        const callsBySymbol = associateCallsWithSymbols(symbols, allCalls);
+        const data = {
+            symbols,
+            calls: callsBySymbol,
+        };
+        callGraphCache.set(filePath, data);
+        return data;
+    }
+    catch (error) {
+        logger.debug(`Failed to analyze ${filePath} for call graph: ${error instanceof Error ? error.message : String(error)}`);
+        return null;
+    }
+}
+/**
+ * Build a call graph from multiple files
+ *
+ * Uses persistent caching with hash-based invalidation for performance.
+ */
+export async function buildCallGraph(files) {
+    if (files.length === 0) {
+        return { nodes: new Map(), files: [], edgeCount: 0 };
+    }
+    // Compute hashes for all files
+    const fileHashes = {};
+    for (const file of files) {
+        fileHashes[file.path] = computeHash(file.content);
+    }
+    // Determine base directory from common path prefix
+    const baseDir = findCommonDirectory(files.map((f) => f.path));
+    // Try to load from persistent cache
+    const cached = loadCallGraphCache(baseDir, fileHashes);
+    if (cached) {
+        return cached;
+    }
+    // Build the call graph
+    const nodes = new Map();
+    const filePaths = [];
+    let edgeCount = 0;
+    // First pass: collect all symbols
+    for (const file of files) {
+        filePaths.push(file.path);
+        const data = await analyzeFileForCallGraph(file.path, file.content);
+        if (!data) {
+            continue;
+        }
+        // Create nodes for all functions/methods
+        for (const symbol of data.symbols) {
+            if (symbol.type === "function" || symbol.type === "method") {
+                const qualifiedName = `${file.path}:${symbol.name}`;
+                nodes.set(qualifiedName, {
+                    name: symbol.name,
+                    qualifiedName,
+                    filePath: file.path,
+                    type: symbol.type,
+                    start: symbol.start,
+                    end: symbol.end,
+                    calls: [],
+                    calledBy: [],
+                });
+            }
+        }
+    }
+    // Second pass: build edges
+    for (const file of files) {
+        const data = callGraphCache.get(file.path);
+        if (!data) {
+            continue;
+        }
+        for (const [symbolName, calls] of data.calls) {
+            const callerKey = `${file.path}:${symbolName}`;
+            const callerNode = nodes.get(callerKey);
+            if (!callerNode && symbolName !== "__global__") {
+                continue;
+            }
+            for (const call of calls) {
+                // Try to find the callee in our nodes
+                // This is a simplified approach - in reality we'd need to resolve imports
+                for (const [nodeKey, node] of nodes) {
+                    if (node.name === call.callee) {
+                        // Add edge
+                        if (callerNode) {
+                            callerNode.calls.push(nodeKey);
+                        }
+                        node.calledBy.push(callerKey);
+                        edgeCount++;
+                    }
+                }
+            }
+        }
+    }
+    const graph = {
+        nodes,
+        files: filePaths,
+        edgeCount,
+    };
+    // Save to persistent cache
+    saveCallGraphCache(baseDir, graph, fileHashes);
+    return graph;
+}
+/**
+ * Find common directory from a list of file paths
+ */
+function findCommonDirectory(paths) {
+    if (paths.length === 0) {
+        return ".";
+    }
+    const firstPathStr = paths[0];
+    if (!firstPathStr) {
+        return ".";
+    }
+    if (paths.length === 1) {
+        return path.dirname(firstPathStr);
+    }
+    // Normalize paths and split into segments
+    const segments = paths.map((p) => path.normalize(p).split(path.sep));
+    const firstPath = segments[0];
+    if (!firstPath) {
+        return ".";
+    }
+    // Find common prefix
+    let commonLength = 0;
+    for (let i = 0; i < firstPath.length; i++) {
+        const segment = firstPath[i];
+        if (segment && segments.every((s) => s[i] === segment)) {
+            commonLength = i + 1;
+        }
+        else {
+            break;
+        }
+    }
+    // Build common directory path
+    const commonSegments = firstPath.slice(0, commonLength);
+    const commonDir = commonSegments.join(path.sep);
+    // If the common path is a file, return its directory
+    if (commonDir &&
+        fs.existsSync(commonDir) &&
+        fs.statSync(commonDir).isFile()) {
+        return path.dirname(commonDir);
+    }
+    return commonDir || ".";
+}
+/**
+ * Get callers and callees for a specific function
+ */
+export function getCallContext(graph, filePath, functionName) {
+    const qualifiedName = `${filePath}:${functionName}`;
+    const node = graph.nodes.get(qualifiedName);
+    if (!node) {
+        return null;
+    }
+    const callers = [];
+    const callees = [];
+    for (const callerKey of node.calledBy) {
+        const caller = graph.nodes.get(callerKey);
+        if (caller) {
+            callers.push(caller);
+        }
+    }
+    for (const calleeKey of node.calls) {
+        const callee = graph.nodes.get(calleeKey);
+        if (callee) {
+            callees.push(callee);
+        }
+    }
+    return { callers, callees };
+}
+/**
+ * Format call context as a string for enrichment
+ */
+export function formatCallContext(callers, callees, maxItems = 5) {
+    const lines = [];
+    if (callers.length > 0) {
+        const callerNames = callers
+            .slice(0, maxItems)
+            .map((c) => c.name)
+            .join(", ");
+        lines.push(`Called by: ${callerNames}`);
+    }
+    if (callees.length > 0) {
+        const calleeNames = callees
+            .slice(0, maxItems)
+            .map((c) => c.name)
+            .join(", ");
+        lines.push(`Calls: ${calleeNames}`);
+    }
+    return lines.join("\n");
+}
+/**
+ * Get call graph cache statistics
+ */
+export function getCallGraphCacheStats() {
+    return {
+        files: callGraphCache.size,
+        entries: Array.from(callGraphCache.keys()),
+    };
+}
+//# sourceMappingURL=callgraph.js.map

package/dist/core/embeddings/callgraph.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"callgraph.js","sourceRoot":"","sources":["../../../src/core/embeddings/callgraph.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAEH,OAAO,EAAE,KAAK,EAAE,MAAM,iBAAiB,CAAC;AACxC,OAAO,KAAK,EAAE,MAAM,SAAS,CAAC;AAC9B,OAAO,KAAK,IAAI,MAAM,WAAW,CAAC;AAClC,OAAO,KAAK,MAAM,MAAM,aAAa,CAAC;AAEtC,OAAO,EAAE,SAAS,EAAoB,MAAM,cAAc,CAAC;AAC3D,OAAO,EAAE,cAAc,EAAE,MAAM,eAAe,CAAC;AAC/C,OAAO,EAAE,aAAa,EAAE,MAAM,aAAa,CAAC;AAC5C,OAAO,EAAE,MAAM,EAAE,MAAM,QAAQ,CAAC;AA2DhC;;GAEG;AACH,SAAS,WAAW,CAAC,OAAe;IAClC,OAAO,MAAM,CAAC,UAAU,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;AAChF,CAAC;AAED;;GAEG;AACH,SAAS,YAAY,CAAC,SAAiB;IACrC,OAAO,IAAI,CAAC,IAAI,CAAC,SAAS,EAAE,YAAY,EAAE,iBAAiB,CAAC,CAAC;AAC/D,CAAC;AAED;;GAEG;AACH,SAAS,kBAAkB,CACzB,SAAiB,EACjB,KAAgB,EAChB,UAAkC;IAElC,IAAI,CAAC;QACH,MAAM,SAAS,GAAG,YAAY,CAAC,SAAS,CAAC,CAAC;QAC1C,MAAM,QAAQ,GAAG,IAAI,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC;QAEzC,gCAAgC;QAChC,IAAI,CAAC,EAAE,CAAC,UAAU,CAAC,QAAQ,CAAC,EAAE,CAAC;YAC7B,EAAE,CAAC,SAAS,CAAC,QAAQ,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;QAC9C,CAAC;QAED,MAAM,UAAU,GAAwB;YACtC,KAAK,EAAE,MAAM,CAAC,WAAW,CAAC,KAAK,CAAC,KAAK,CAAC;YACtC,KAAK,EAAE,KAAK,CAAC,KAAK;YAClB,SAAS,EAAE,KAAK,CAAC,SAAS;YAC1B,UAAU;YACV,SAAS,EAAE,IAAI,CAAC,GAAG,EAAE;SACtB,CAAC;QAEF,EAAE,CAAC,aAAa,CAAC,SAAS,EAAE,IAAI,CAAC,SAAS,CAAC,UAAU,CAAC,EAAE,OAAO,CAAC,CAAC;QACjE,MAAM,CAAC,KAAK,CAAC,2BAA2B,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;IAC5E,CAAC;IAAC,MAAM,CAAC;QACP,mEAAmE;QACnE,MAAM,CAAC,KAAK,CAAC,uDAAuD,CAAC,CAAC;IACxE,CAAC;AACH,CAAC;AAED;;GAEG;AACH,SAAS,kBAAkB,CACzB,SAAiB,EACjB,aAAqC;IAErC,MAAM,SAAS,GAAG,YAAY,CAAC,SAAS,CAAC,CAAC;IAE1C,IAAI,CAAC,EAAE,CAAC,UAAU,CAAC,SAAS,CAAC,EAAE,CAAC;QAC9B,OAAO,IAAI,CAAC;IACd,CAAC;IAED,IAAI,CAAC;QACH,MAAM,OAAO,GAAG,EAAE,CAAC,YAAY,CAAC,SAAS,EAAE,OAAO,CAAC,CAAC;QACpD,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,OAAO,CAAwB,CAAC;QAE1D,kDAAkD;QAClD,MAAM,WAAW,GAAG,IAAI,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,UAAU,CAAC,CAAC,CAAC;QAC5D,MAAM,YAAY,GAAG,IAAI,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC,CAAC;QAEzD,mCAAmC;QACnC,IAAI,WAAW,CAAC,IAAI,KAAK,YAAY,CAAC,IAAI,EAAE,CAAC;YAC3C,MAAM,CAAC,KAAK,CAAC,8CAA8C,CAAC,CAAC;YAC7D,OAAO,IAAI,CAAC;QACd,CAAC;QAED,2BAA2B;QAC3B,KAAK,MAAM,CAAC,QAAQ,EAAE,IAAI,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,aAAa,CAAC,EAAE,CAAC;YAC7D,IAAI,MAAM,CAAC,UAAU,CAAC,QAAQ,CAAC,KAAK,IAAI,EAAE,CAAC;gBACzC,MAAM,CAAC,KAAK,CAAC,6BAA6B,QAAQ,UAAU,CAAC,CAAC;gBAC9D,OAAO,IAAI,CAAC;YACd,CAAC;QACH,CAAC;QAED,mCAAmC;QACnC,MAAM,KAAK,GAAG,IAAI,GAAG,CAAwB,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC;QAE3E,MAAM,CAAC,KAAK,CAAC,4BAA4B,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;QAErE,OAAO;YACL,KAAK;YACL,KAAK,EAAE,MAAM,CAAC,KAAK;YACnB,SAAS,EAAE,MAAM,CAAC,SAAS;SAC5B,CAAC;IACJ,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QACf,MAAM,CAAC,KAAK,CACV,oCAAoC,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,CAC7F,CAAC;QACF,OAAO,IAAI,CAAC;IACd,CAAC;AACH,CAAC;AAUD,MAAM,cAAc,GAAG,IAAI,GAAG,EAAwB,CAAC;AAEvD;;GAEG;AACH,MAAM,UAAU,mBAAmB;IACjC,cAAc,CAAC,KAAK,EAAE,CAAC;AACzB,CAAC;AAED,0CAA0C;AAC1C,aAAa,CAAC,2BAA2B,EAAE,mBAAmB,CAAC,CAAC;AAEhE;;GAEG;AACH,SAAS,oBAAoB,CAC3B,IAAyB,EACzB,gBAAiD,EACjD,QAAgB;IAEhB,MAAM,aAAa,GAAG,IAAI,GAAG,EAA0B,CAAC;IAExD,2DAA2D;IAC3D,MAAM,YAAY,GAA2B;QAC3C,UAAU,EAAE;;;;;KAKX;QACD,UAAU,EAAE;;;;;KAKX;QACD,MAAM,EAAE;;;;;KAKP;QACD,EAAE,EAAE;;;;;KAKH;KACF,CAAC;IAEF,MAAM,OAAO,GAAG,YAAY,CAAC,QAAQ,CAAC,CAAC;IACvC,IAAI,CAAC,OAAO,EAAE,CAAC;QACb,OAAO,aAAa,CAAC;IACvB,CAAC;IAED,IAAI,CAAC;QACH,MAAM,KAAK,GAAG,IAAI,KAAK,CAAC,gBAAgB,EAAE,OAAO,CAAC,CAAC;QACnD,MAAM,OAAO,GAAG,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;QAE7C,oCAAoC;QACpC,MAAM,YAAY,GAA6C,EAAE,CAAC;QAElE,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;YAC5B,KAAK,MAAM,OAAO,IAAI,KAAK,CAAC,QAAQ,EAAE,CAAC;gBACrC,IAAI,OAAO,CAAC,IAAI,KAAK,QAAQ,EAAE,CAAC;oBAC9B,YAAY,CAAC,IAAI,CAAC;wBAChB,MAAM,EAAE,OAAO,CAAC,IAAI,CAAC,IAAI;wBACzB,QAAQ,EAAE;4BACR,IAAI,EAAE,OAAO,CAAC,IAAI,CAAC,aAAa,CAAC,GAAG,GAAG,CAAC;4BACxC,MAAM,EAAE,OAAO,CAAC,IAAI,CAAC,aAAa,CAAC,MAAM;4BACzC,MAAM,EAAE,OAAO,CAAC,IAAI,CAAC,UAAU;yBAChC;qBACF,CAAC,CAAC;gBACL,CAAC;YACH,CAAC;QACH,CAAC;QAED,sDAAsD;QACtD,4EAA4E;QAC5E,IAAI,YAAY,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAC5B,aAAa,CAAC,GAAG,CACf,YAAY,EACZ,YAAY,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;gBACvB,MAAM,EAAE,CAAC,CAAC,MAAM;gBAChB,QAAQ,EAAE,CAAC,CAAC,QAAQ;aACrB,CAAC,CAAC,CACJ,CAAC;QACJ,CAAC;IACH,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QACf,MAAM,CAAC,KAAK,CACV,+BAA+B,QAAQ,KAAK,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,CACrG,CAAC;IACJ,CAAC;IAED,OAAO,aAAa,CAAC;AACvB,CAAC;AAED;;GAEG;AACH,SAAS,yBAAyB,CAChC,OAAiB,EACjB,QAAwB;IAExB,MAAM,aAAa,GAAG,IAAI,GAAG,EAA0B,CAAC;IAExD,oDAAoD;IACpD,MAAM,aAAa,GAAG,CAAC,GAAG,OAAO,CAAC;SAC/B,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,UAAU,IAAI,CAAC,CAAC,IAAI,KAAK,QAAQ,CAAC;SAC3D,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;IAEnD,KAAK,MAAM,IAAI,IAAI,QAAQ,EAAE,CAAC;QAC5B,0CAA0C;QAC1C,IAAI,gBAAgB,GAAkB,IAAI,CAAC;QAE3C,KAAK,MAAM,MAAM,IAAI,aAAa,EAAE,CAAC;YACnC,IACE,IAAI,CAAC,QAAQ,CAAC,MAAM,IAAI,MAAM,CAAC,KAAK,CAAC,MAAM;gBAC3C,IAAI,CAAC,QAAQ,CAAC,MAAM,IAAI,MAAM,CAAC,GAAG,CAAC,MAAM,EACzC,CAAC;gBACD,gBAAgB,GAAG,MAAM,CAAC;YAC5B,CAAC;iBAAM,IAAI,IAAI,CAAC,QAAQ,CAAC,MAAM,GAAG,MAAM,CAAC,KAAK,CAAC,MAAM,EAAE,CAAC;gBACtD,sDAAsD;gBACtD,MAAM;YACR,CAAC;QACH,CAAC;QAED,MAAM,UAAU,GAAG,gBAAgB,EAAE,IAAI,IAAI,YAAY,CAAC;QAC1D,MAAM,QAAQ,GAAG,aAAa,CAAC,GAAG,CAAC,UAAU,CAAC,IAAI,EAAE,CAAC;QACrD,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACpB,aAAa,CAAC,GAAG,CAAC,UAAU,EAAE,QAAQ,CAAC,CAAC;IAC1C,CAAC;IAED,OAAO,aAAa,CAAC;AACvB,CAAC;AAED;;GAEG;AACH,MAAM,CAAC,KAAK,UAAU,uBAAuB,CAC3C,QAAgB,EAChB,OAAe;IAEf,cAAc;IACd,MAAM,MAAM,GAAG,cAAc,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;IAC5C,IAAI,MAAM,EAAE,CAAC;QACX,OAAO,MAAM,CAAC;IAChB,CAAC;IAED,IAAI,CAAC;QACH,MAAM,WAAW,GAAG,MAAM,SAAS,CAAC,OAAO,EAAE,EAAE,QAAQ,EAAE,CAAC,CAAC;QAE3D,MAAM,EAAE,OAAO,EAAE,GAAG,cAAc,CAChC,WAAW,CAAC,IAAI,EAChB,WAAW,CAAC,gBAAgB,EAC5B,WAAW,CAAC,QAAQ,CACrB,CAAC;QAEF,MAAM,QAAQ,GAAG,oBAAoB,CACnC,WAAW,CAAC,IAAI,EAChB,WAAW,CAAC,gBAAgB,EAC5B,WAAW,CAAC,QAAQ,CACrB,CAAC;QAEF,2CAA2C;QAC3C,MAAM,QAAQ,GAAG,QAAQ,CAAC,GAAG,CAAC,YAAY,CAAC,IAAI,EAAE,CAAC;QAClD,MAAM,aAAa,GAAG,yBAAyB,CAAC,OAAO,EAAE,QAAQ,CAAC,CAAC;QAEnE,MAAM,IAAI,GAAiB;YACzB,OAAO;YACP,KAAK,EAAE,aAAa;SACrB,CAAC;QAEF,cAAc,CAAC,GAAG,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC;QACnC,OAAO,IAAI,CAAC;IACd,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QACf,MAAM,CAAC,KAAK,CACV,qBAAqB,QAAQ,oBAAoB,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,CAC1G,CAAC;QACF,OAAO,IAAI,CAAC;IACd,CAAC;AACH,CAAC;AAED;;;;GAIG;AACH,MAAM,CAAC,KAAK,UAAU,cAAc,CAClC,KAA0C;IAE1C,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACvB,OAAO,EAAE,KAAK,EAAE,IAAI,GAAG,EAAE,EAAE,KAAK,EAAE,EAAE,EAAE,SAAS,EAAE,CAAC,EAAE,CAAC;IACvD,CAAC;IAED,+BAA+B;IAC/B,MAAM,UAAU,GAA2B,EAAE,CAAC;IAC9C,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,WAAW,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;IACpD,CAAC;IAED,mDAAmD;IACnD,MAAM,OAAO,GAAG,mBAAmB,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;IAE9D,oCAAoC;IACpC,MAAM,MAAM,GAAG,kBAAkB,CAAC,OAAO,EAAE,UAAU,CAAC,CAAC;IACvD,IAAI,MAAM,EAAE,CAAC;QACX,OAAO,MAAM,CAAC;IAChB,CAAC;IAED,uBAAuB;IACvB,MAAM,KAAK,GAAG,IAAI,GAAG,EAAyB,CAAC;IAC/C,MAAM,SAAS,GAAa,EAAE,CAAC;IAC/B,IAAI,SAAS,GAAG,CAAC,CAAC;IAElB,kCAAkC;IAClC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAC1B,MAAM,IAAI,GAAG,MAAM,uBAAuB,CAAC,IAAI,CAAC,IAAI,EAAE,IAAI,CAAC,OAAO,CAAC,CAAC;QAEpE,IAAI,CAAC,IAAI,EAAE,CAAC;YACV,SAAS;QACX,CAAC;QAED,yCAAyC;QACzC,KAAK,MAAM,MAAM,IAAI,IAAI,CAAC,OAAO,EAAE,CAAC;YAClC,IAAI,MAAM,CAAC,IAAI,KAAK,UAAU,IAAI,MAAM,CAAC,IAAI,KAAK,QAAQ,EAAE,CAAC;gBAC3D,MAAM,aAAa,GAAG,GAAG,IAAI,CAAC,IAAI,IAAI,MAAM,CAAC,IAAI,EAAE,CAAC;gBACpD,KAAK,CAAC,GAAG,CAAC,aAAa,EAAE;oBACvB,IAAI,EAAE,MAAM,CAAC,IAAI;oBACjB,aAAa;oBACb,QAAQ,EAAE,IAAI,CAAC,IAAI;oBACnB,IAAI,EAAE,MAAM,CAAC,IAAI;oBACjB,KAAK,EAAE,MAAM,CAAC,KAAK;oBACnB,GAAG,EAAE,MAAM,CAAC,GAAG;oBACf,KAAK,EAAE,EAAE;oBACT,QAAQ,EAAE,EAAE;iBACb,CAAC,CAAC;YACL,CAAC;QACH,CAAC;IACH,CAAC;IAED,2BAA2B;IAC3B,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,IAAI,GAAG,cAAc,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAC3C,IAAI,CAAC,IAAI,EAAE,CAAC;YACV,SAAS;QACX,CAAC;QAED,KAAK,MAAM,CAAC,UAAU,EAAE,KAAK,CAAC,IAAI,IAAI,CAAC,KAAK,EAAE,CAAC;YAC7C,MAAM,SAAS,GAAG,GAAG,IAAI,CAAC,IAAI,IAAI,UAAU,EAAE,CAAC;YAC/C,MAAM,UAAU,GAAG,KAAK,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;YAExC,IAAI,CAAC,UAAU,IAAI,UAAU,KAAK,YAAY,EAAE,CAAC;gBAC/C,SAAS;YACX,CAAC;YAED,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;gBACzB,sCAAsC;gBACtC,0EAA0E;gBAC1E,KAAK,MAAM,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,KAAK,EAAE,CAAC;oBACpC,IAAI,IAAI,CAAC,IAAI,KAAK,IAAI,CAAC,MAAM,EAAE,CAAC;wBAC9B,WAAW;wBACX,IAAI,UAAU,EAAE,CAAC;4BACf,UAAU,CAAC,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;wBACjC,CAAC;wBACD,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;wBAC9B,SAAS,EAAE,CAAC;oBACd,CAAC;gBACH,CAAC;YACH,CAAC;QACH,CAAC;IACH,CAAC;IAED,MAAM,KAAK,GAAc;QACvB,KAAK;QACL,KAAK,EAAE,SAAS;QAChB,SAAS;KACV,CAAC;IAEF,2BAA2B;IAC3B,kBAAkB,CAAC,OAAO,EAAE,KAAK,EAAE,UAAU,CAAC,CAAC;IAE/C,OAAO,KAAK,CAAC;AACf,CAAC;AAED;;GAEG;AACH,SAAS,mBAAmB,CAAC,KAAe;IAC1C,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACvB,OAAO,GAAG,CAAC;IACb,CAAC;IAED,MAAM,YAAY,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;IAC9B,IAAI,CAAC,YAAY,EAAE,CAAC;QAClB,OAAO,GAAG,CAAC;IACb,CAAC;IAED,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACvB,OAAO,IAAI,CAAC,OAAO,CAAC,YAAY,CAAC,CAAC;IACpC,CAAC;IAED,0CAA0C;IAC1C,MAAM,QAAQ,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;IACrE,MAAM,SAAS,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC;IAE9B,IAAI,CAAC,SAAS,EAAE,CAAC;QACf,OAAO,GAAG,CAAC;IACb,CAAC;IAED,qBAAqB;IACrB,IAAI,YAAY,GAAG,CAAC,CAAC;IAErB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,SAAS,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAC1C,MAAM,OAAO,GAAG,SAAS,CAAC,CAAC,CAAC,CAAC;QAC7B,IAAI,OAAO,IAAI,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,OAAO,CAAC,EAAE,CAAC;YACvD,YAAY,GAAG,CAAC,GAAG,CAAC,CAAC;QACvB,CAAC;aAAM,CAAC;YACN,MAAM;QACR,CAAC;IACH,CAAC;IAED,8BAA8B;IAC9B,MAAM,cAAc,GAAG,SAAS,CAAC,KAAK,CAAC,CAAC,EAAE,YAAY,CAAC,CAAC;IACxD,MAAM,SAAS,GAAG,cAAc,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IAEhD,qDAAqD;IACrD,IACE,SAAS;QACT,EAAE,CAAC,UAAU,CAAC,SAAS,CAAC;QACxB,EAAE,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,MAAM,EAAE,EAC/B,CAAC;QACD,OAAO,IAAI,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC;IACjC,CAAC;IAED,OAAO,SAAS,IAAI,GAAG,CAAC;AAC1B,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,cAAc,CAC5B,KAAgB,EAChB,QAAgB,EAChB,YAAoB;IAKpB,MAAM,aAAa,GAAG,GAAG,QAAQ,IAAI,YAAY,EAAE,CAAC;IACpD,MAAM,IAAI,GAAG,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,aAAa,CAAC,CAAC;IAE5C,IAAI,CAAC,IAAI,EAAE,CAAC;QACV,OAAO,IAAI,CAAC;IACd,CAAC;IAED,MAAM,OAAO,GAAoB,EAAE,CAAC;IACpC,MAAM,OAAO,GAAoB,EAAE,CAAC;IAEpC,KAAK,MAAM,SAAS,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;QACtC,MAAM,MAAM,GAAG,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;QAC1C,IAAI,MAAM,EAAE,CAAC;YACX,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QACvB,CAAC;IACH,CAAC;IAED,KAAK,MAAM,SAAS,IAAI,IAAI,CAAC,KAAK,EAAE,CAAC;QACnC,MAAM,MAAM,GAAG,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;QAC1C,IAAI,MAAM,EAAE,CAAC;YACX,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QACvB,CAAC;IACH,CAAC;IAED,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,CAAC;AAC9B,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,iBAAiB,CAC/B,OAAwB,EACxB,OAAwB,EACxB,QAAQ,GAAG,CAAC;IAEZ,MAAM,KAAK,GAAa,EAAE,CAAC;IAE3B,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACvB,MAAM,WAAW,GAAG,OAAO;aACxB,KAAK,CAAC,CAAC,EAAE,QAAQ,CAAC;aAClB,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC;aAClB,IAAI,CAAC,IAAI,CAAC,CAAC;QACd,KAAK,CAAC,IAAI,CAAC,cAAc,WAAW,EAAE,CAAC,CAAC;IAC1C,CAAC;IAED,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACvB,MAAM,WAAW,GAAG,OAAO;aACxB,KAAK,CAAC,CAAC,EAAE,QAAQ,CAAC;aAClB,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC;aAClB,IAAI,CAAC,IAAI,CAAC,CAAC;QACd,KAAK,CAAC,IAAI,CAAC,UAAU,WAAW,EAAE,CAAC,CAAC;IACtC,CAAC;IAED,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AAC1B,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,sBAAsB;IAIpC,OAAO;QACL,KAAK,EAAE,cAAc,CAAC,IAAI;QAC1B,OAAO,EAAE,KAAK,CAAC,IAAI,CAAC,cAAc,CAAC,IAAI,EAAE,CAAC;KAC3C,CAAC;AACJ,CAAC"}

package/dist/core/embeddings/chunker.d.ts

@@ -0,0 +1,37 @@
+/**
+ * Code chunker for splitting source files into embeddable chunks
+ *
+ * Uses tree-sitter for semantic chunking based on symbols (functions, classes, etc.)
+ * This produces much better embeddings than character-based splitting.
+ */
+import type { CodeChunk, EmbeddingConfig } from "./types";
+/**
+ * Detect language from file extension
+ */
+export declare function detectLanguage(filePath: string): string;
+/**
+ * Chunk a source file using tree-sitter for semantic boundaries
+ *
+ * Strategy:
+ * 1. Parse file with tree-sitter and extract symbols
+ * 2. Each function/class/method becomes its own chunk
+ * 3. Code between symbols (imports, top-level code) is grouped together
+ * 4. Large symbols are split at line boundaries if they exceed maxSize
+ */
+export declare function chunkFile(filePath: string, content: string, config: Pick<EmbeddingConfig, "defaultChunkSize" | "defaultChunkOverlap">): Promise<CodeChunk[]>;
+/**
+ * Chunk multiple files
+ */
+export declare function chunkFiles(files: {
+    path: string;
+    content: string;
+}[], config: Pick<EmbeddingConfig, "defaultChunkSize" | "defaultChunkOverlap">): Promise<CodeChunk[]>;
+/**
+ * Supported file extensions for indexing
+ */
+export declare const SUPPORTED_EXTENSIONS: string[];
+/**
+ * Check if a file should be indexed
+ */
+export declare function shouldIndexFile(filePath: string): boolean;
+//# sourceMappingURL=chunker.d.ts.map

package/dist/core/embeddings/chunker.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"chunker.d.ts","sourceRoot":"","sources":["../../../src/core/embeddings/chunker.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAQH,OAAO,KAAK,EAAE,SAAS,EAAE,eAAe,EAAE,MAAM,SAAS,CAAC;AAkB1D;;GAEG;AACH,wBAAgB,cAAc,CAAC,QAAQ,EAAE,MAAM,GAAG,MAAM,CA2CvD;AA6GD;;;;;;;;GAQG;AACH,wBAAsB,SAAS,CAC7B,QAAQ,EAAE,MAAM,EAChB,OAAO,EAAE,MAAM,EACf,MAAM,EAAE,IAAI,CAAC,eAAe,EAAE,kBAAkB,GAAG,qBAAqB,CAAC,GACxE,OAAO,CAAC,SAAS,EAAE,CAAC,CAmItB;AAuCD;;GAEG;AACH,wBAAsB,UAAU,CAC9B,KAAK,EAAE;IAAE,IAAI,EAAE,MAAM,CAAC;IAAC,OAAO,EAAE,MAAM,CAAA;CAAE,EAAE,EAC1C,MAAM,EAAE,IAAI,CAAC,eAAe,EAAE,kBAAkB,GAAG,qBAAqB,CAAC,GACxE,OAAO,CAAC,SAAS,EAAE,CAAC,CAStB;AAED;;GAEG;AACH,eAAO,MAAM,oBAAoB,UAwBhC,CAAC;AAEF;;GAEG;AACH,wBAAgB,eAAe,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAGzD"}