@know-graph/core 0.4.0

package/dist/parsers/python-parser.js

@@ -0,0 +1,226 @@
+import { extractMetadata } from './metadata-extractor.js';
+const PYTHON_EXTENSIONS = ['.py', '.pyi'];
+/**
+ * Regex to match Python triple-quoted docstrings.
+ * Captures the quote style and the content between quotes.
+ */
+const DOCSTRING_REGEX = /("""[\s\S]*?"""|'''[\s\S]*?''')/g;
+/**
+ * Regex to match class definitions.
+ */
+const CLASS_DEF_REGEX = /^(\s*)class\s+(\w+)(?:\(([^)]*)\))?:/;
+/**
+ * Regex to match function/method definitions.
+ */
+const FUNC_DEF_REGEX = /^(\s*)(?:async\s+)?def\s+(\w+)\(([^)]*)\)(?:\s*->\s*([^:]+))?:/;
+/**
+ * Regex to match decorator lines.
+ */
+const DECORATOR_REGEX = /^\s*@(\S+)/;
+function getLineNumber(source, charIndex) {
+    let line = 1;
+    for (let i = 0; i < charIndex && i < source.length; i++) {
+        if (source[i] === '\n') {
+            line++;
+        }
+    }
+    return line;
+}
+function stripDocstringQuotes(raw) {
+    const quoteChar = raw[0];
+    const tripleQuote = quoteChar.repeat(3);
+    if (raw.startsWith(tripleQuote) && raw.endsWith(tripleQuote)) {
+        return raw.slice(3, -3);
+    }
+    return raw;
+}
+function findAllDocstrings(content) {
+    const results = [];
+    let match;
+    const regex = new RegExp(DOCSTRING_REGEX.source, 'g');
+    while ((match = regex.exec(content)) !== null) {
+        const startLine = getLineNumber(content, match.index);
+        const endLine = getLineNumber(content, match.index + match[0].length - 1);
+        results.push({
+            content: stripDocstringQuotes(match[0]),
+            startLine,
+            endLine,
+            startIndex: match.index,
+        });
+    }
+    return results;
+}
+function getLineAt(content, lineNumber) {
+    const lines = content.split('\n');
+    return lines[lineNumber - 1] ?? '';
+}
+function collectDecorators(content, defLineNumber) {
+    const lines = content.split('\n');
+    const decorators = [];
+    let lineIdx = defLineNumber - 2; // 0-indexed, line before def
+    while (lineIdx >= 0) {
+        const line = lines[lineIdx];
+        if (line === undefined)
+            break;
+        const decoratorMatch = line.match(DECORATOR_REGEX);
+        if (decoratorMatch) {
+            decorators.unshift(decoratorMatch[1]);
+            lineIdx--;
+        }
+        else if (line.trim() === '' || line.trim().startsWith('#')) {
+            lineIdx--;
+        }
+        else {
+            break;
+        }
+    }
+    return decorators;
+}
+function findDefinitionBefore(content, docstringStartLine) {
+    const lineAbove = getLineAt(content, docstringStartLine - 1);
+    // Check if it's a function/method definition
+    const funcMatch = lineAbove.match(FUNC_DEF_REGEX);
+    if (funcMatch) {
+        const indent = funcMatch[1];
+        const name = funcMatch[2];
+        const params = funcMatch[3];
+        const returnType = funcMatch[4]?.trim();
+        const defLine = docstringStartLine - 1;
+        const decorators = collectDecorators(content, defLine);
+        const signatureParts = [`def ${name}(${params})`];
+        if (returnType) {
+            signatureParts.push(` -> ${returnType}`);
+        }
+        // Determine if method (has indentation and a class context)
+        const isMethod = (indent?.length ?? 0) > 0;
+        let parent;
+        if (isMethod) {
+            parent = findEnclosingClass(content, defLine);
+        }
+        return {
+            name,
+            kind: isMethod && parent ? 'method' : 'function',
+            signature: signatureParts.join(''),
+            parent,
+            line: defLine,
+            decorators,
+        };
+    }
+    // Check if it's a class definition
+    const classMatch = lineAbove.match(CLASS_DEF_REGEX);
+    if (classMatch) {
+        const name = classMatch[2];
+        const defLine = docstringStartLine - 1;
+        const decorators = collectDecorators(content, defLine);
+        return {
+            name,
+            kind: 'class',
+            line: defLine,
+            decorators,
+        };
+    }
+    return null;
+}
+function findEnclosingClass(content, lineNumber) {
+    const lines = content.split('\n');
+    const targetLine = lines[lineNumber - 1];
+    if (!targetLine)
+        return undefined;
+    const targetIndent = targetLine.search(/\S/);
+    for (let i = lineNumber - 2; i >= 0; i--) {
+        const line = lines[i];
+        if (line === undefined)
+            continue;
+        const currentIndent = line.search(/\S/);
+        if (currentIndent === -1)
+            continue; // blank line
+        if (currentIndent < targetIndent) {
+            const classMatch = line.match(CLASS_DEF_REGEX);
+            if (classMatch) {
+                return classMatch[2];
+            }
+            // If we hit a non-class at a lower indent, stop
+            break;
+        }
+    }
+    return undefined;
+}
+function isModuleLevelDocstring(content, docstringStartLine) {
+    // A module-level docstring is at the beginning of the file,
+    // possibly preceded by comments, blank lines, or shebang
+    const lines = content.split('\n');
+    for (let i = 0; i < docstringStartLine - 1; i++) {
+        const line = lines[i]?.trim() ?? '';
+        if (line === '' || line.startsWith('#') || line.startsWith('!')) {
+            continue;
+        }
+        // Found a non-comment, non-blank line before the docstring
+        return false;
+    }
+    return true;
+}
+function getModuleName(filePath) {
+    const parts = filePath.split('/');
+    const fileName = parts[parts.length - 1] ?? '';
+    return fileName.replace(/\.(py|pyi)$/, '');
+}
+export function createPythonParser() {
+    return {
+        name: 'python',
+        supportedExtensions: PYTHON_EXTENSIONS,
+        parse(content, filePath) {
+            const docstrings = findAllDocstrings(content);
+            const results = [];
+            for (const docstring of docstrings) {
+                const extraction = extractMetadata(docstring.content, docstring.startLine);
+                if (!extraction.metadata) {
+                    continue;
+                }
+                const metadata = extraction.metadata;
+                // Determine context
+                const defContext = findDefinitionBefore(content, docstring.startLine);
+                if (defContext) {
+                    results.push({
+                        name: defContext.name,
+                        filePath,
+                        line: defContext.line,
+                        column: 1,
+                        language: 'python',
+                        entityType: metadata.type,
+                        metadata,
+                        rawDocstring: docstring.content,
+                        signature: defContext.signature,
+                        parent: defContext.parent,
+                    });
+                }
+                else if (isModuleLevelDocstring(content, docstring.startLine)) {
+                    results.push({
+                        name: getModuleName(filePath),
+                        filePath,
+                        line: docstring.startLine,
+                        column: 1,
+                        language: 'python',
+                        entityType: metadata.type,
+                        metadata,
+                        rawDocstring: docstring.content,
+                    });
+                }
+                else {
+                    // Standalone docstring not at module level - still include it
+                    results.push({
+                        name: 'unknown',
+                        filePath,
+                        line: docstring.startLine,
+                        column: 1,
+                        language: 'python',
+                        entityType: metadata.type,
+                        metadata,
+                        rawDocstring: docstring.content,
+                    });
+                }
+            }
+            return results;
+        },
+    };
+}
+//# sourceMappingURL=python-parser.js.map

package/dist/parsers/python-parser.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"python-parser.js","sourceRoot":"","sources":["../../src/parsers/python-parser.ts"],"names":[],"mappings":"AAaA,OAAO,EAAE,eAAe,EAAE,MAAM,yBAAyB,CAAC;AAE1D,MAAM,iBAAiB,GAAG,CAAC,KAAK,EAAE,MAAM,CAAU,CAAC;AAEnD;;;GAGG;AACH,MAAM,eAAe,GAAG,kCAAkC,CAAC;AAE3D;;GAEG;AACH,MAAM,eAAe,GAAG,sCAAsC,CAAC;AAE/D;;GAEG;AACH,MAAM,cAAc,GAClB,gEAAgE,CAAC;AAEnE;;GAEG;AACH,MAAM,eAAe,GAAG,YAAY,CAAC;AAkBrC,SAAS,aAAa,CAAC,MAAc,EAAE,SAAiB;IACtD,IAAI,IAAI,GAAG,CAAC,CAAC;IACb,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,SAAS,IAAI,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACxD,IAAI,MAAM,CAAC,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;YACvB,IAAI,EAAE,CAAC;QACT,CAAC;IACH,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED,SAAS,oBAAoB,CAAC,GAAW;IACvC,MAAM,SAAS,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC;IACzB,MAAM,WAAW,GAAG,SAAS,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC;IACxC,IAAI,GAAG,CAAC,UAAU,CAAC,WAAW,CAAC,IAAI,GAAG,CAAC,QAAQ,CAAC,WAAW,CAAC,EAAE,CAAC;QAC7D,OAAO,GAAG,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;IAC1B,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAED,SAAS,iBAAiB,CAAC,OAAe;IACxC,MAAM,OAAO,GAAqB,EAAE,CAAC;IACrC,IAAI,KAA6B,CAAC;IAClC,MAAM,KAAK,GAAG,IAAI,MAAM,CAAC,eAAe,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC;IAEtD,OAAO,CAAC,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;QAC9C,MAAM,SAAS,GAAG,aAAa,CAAC,OAAO,EAAE,KAAK,CAAC,KAAK,CAAC,CAAC;QACtD,MAAM,OAAO,GAAG,aAAa,CAAC,OAAO,EAAE,KAAK,CAAC,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;QAC1E,OAAO,CAAC,IAAI,CAAC;YACX,OAAO,EAAE,oBAAoB,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;YACvC,SAAS;YACT,OAAO;YACP,UAAU,EAAE,KAAK,CAAC,KAAK;SACxB,CAAC,CAAC;IACL,CAAC;IAED,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,SAAS,SAAS,CAAC,OAAe,EAAE,UAAkB;IACpD,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAClC,OAAO,KAAK,CAAC,UAAU,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;AACrC,CAAC;AAED,SAAS,iBAAiB,CACxB,OAAe,EACf,aAAqB;IAErB,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAClC,MAAM,UAAU,GAAa,EAAE,CAAC;IAChC,IAAI,OAAO,GAAG,aAAa,GAAG,CAAC,CAAC,CAAC,6BAA6B;IAE9D,OAAO,OAAO,IAAI,CAAC,EAAE,CAAC;QACpB,MAAM,IAAI,GAAG,KAAK,CAAC,OAAO,CAAC,CAAC;QAC5B,IAAI,IAAI,KAAK,SAAS;YAAE,MAAM;QAC9B,MAAM,cAAc,GAAG,IAAI,CAAC,KAAK,CAAC,eAAe,CAAC,CAAC;QACnD,IAAI,cAAc,EAAE,CAAC;YACnB,UAAU,CAAC,OAAO,CAAC,cAAc,CAAC,CAAC,CAAC,CAAC,CAAC;YACtC,OAAO,EAAE,CAAC;QACZ,CAAC;aAAM,IAAI,IAAI,CAAC,IAAI,EAAE,KAAK,EAAE,IAAI,IAAI,CAAC,IAAI,EAAE,CAAC,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;YAC7D,OAAO,EAAE,CAAC;QACZ,CAAC;aAAM,CAAC;YACN,MAAM;QACR,CAAC;IACH,CAAC;IAED,OAAO,UAAU,CAAC;AACpB,CAAC;AAED,SAAS,oBAAoB,CAC3B,OAAe,EACf,kBAA0B;IAE1B,MAAM,SAAS,GAAG,SAAS,CAAC,OAAO,EAAE,kBAAkB,GAAG,CAAC,CAAC,CAAC;IAE7D,6CAA6C;IAC7C,MAAM,SAAS,GAAG,SAAS,CAAC,KAAK,CAAC,cAAc,CAAC,CAAC;IAClD,IAAI,SAAS,EAAE,CAAC;QACd,MAAM,MAAM,GAAG,SAAS,CAAC,CAAC,CAAC,CAAC;QAC5B,MAAM,IAAI,GAAG,SAAS,CAAC,CAAC,CAAC,CAAC;QAC1B,MAAM,MAAM,GAAG,SAAS,CAAC,CAAC,CAAC,CAAC;QAC5B,MAAM,UAAU,GAAG,SAAS,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC;QACxC,MAAM,OAAO,GAAG,kBAAkB,GAAG,CAAC,CAAC;QACvC,MAAM,UAAU,GAAG,iBAAiB,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;QAEvD,MAAM,cAAc,GAAG,CAAC,OAAO,IAAI,IAAI,MAAM,GAAG,CAAC,CAAC;QAClD,IAAI,UAAU,EAAE,CAAC;YACf,cAAc,CAAC,IAAI,CAAC,OAAO,UAAU,EAAE,CAAC,CAAC;QAC3C,CAAC;QAED,4DAA4D;QAC5D,MAAM,QAAQ,GAAG,CAAC,MAAM,EAAE,MAAM,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC;QAC3C,IAAI,MAA0B,CAAC;QAE/B,IAAI,QAAQ,EAAE,CAAC;YACb,MAAM,GAAG,kBAAkB,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;QAChD,CAAC;QAED,OAAO;YACL,IAAI;YACJ,IAAI,EAAE,QAAQ,IAAI,MAAM,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,UAAU;YAChD,SAAS,EAAE,cAAc,CAAC,IAAI,CAAC,EAAE,CAAC;YAClC,MAAM;YACN,IAAI,EAAE,OAAO;YACb,UAAU;SACX,CAAC;IACJ,CAAC;IAED,mCAAmC;IACnC,MAAM,UAAU,GAAG,SAAS,CAAC,KAAK,CAAC,eAAe,CAAC,CAAC;IACpD,IAAI,UAAU,EAAE,CAAC;QACf,MAAM,IAAI,GAAG,UAAU,CAAC,CAAC,CAAC,CAAC;QAC3B,MAAM,OAAO,GAAG,kBAAkB,GAAG,CAAC,CAAC;QACvC,MAAM,UAAU,GAAG,iBAAiB,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;QAEvD,OAAO;YACL,IAAI;YACJ,IAAI,EAAE,OAAO;YACb,IAAI,EAAE,OAAO;YACb,UAAU;SACX,CAAC;IACJ,CAAC;IAED,OAAO,IAAI,CAAC;AACd,CAAC;AAED,SAAS,kBAAkB,CACzB,OAAe,EACf,UAAkB;IAElB,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAClC,MAAM,UAAU,GAAG,KAAK,CAAC,UAAU,GAAG,CAAC,CAAC,CAAC;IACzC,IAAI,CAAC,UAAU;QAAE,OAAO,SAAS,CAAC;IAElC,MAAM,YAAY,GAAG,UAAU,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;IAE7C,KAAK,IAAI,CAAC,GAAG,UAAU,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;QACzC,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;QACtB,IAAI,IAAI,KAAK,SAAS;YAAE,SAAS;QACjC,MAAM,aAAa,GAAG,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;QACxC,IAAI,aAAa,KAAK,CAAC,CAAC;YAAE,SAAS,CAAC,aAAa;QAEjD,IAAI,aAAa,GAAG,YAAY,EAAE,CAAC;YACjC,MAAM,UAAU,GAAG,IAAI,CAAC,KAAK,CAAC,eAAe,CAAC,CAAC;YAC/C,IAAI,UAAU,EAAE,CAAC;gBACf,OAAO,UAAU,CAAC,CAAC,CAAC,CAAC;YACvB,CAAC;YACD,gDAAgD;YAChD,MAAM;QACR,CAAC;IACH,CAAC;IAED,OAAO,SAAS,CAAC;AACnB,CAAC;AAED,SAAS,sBAAsB,CAC7B,OAAe,EACf,kBAA0B;IAE1B,4DAA4D;IAC5D,yDAAyD;IACzD,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAClC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,kBAAkB,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;QAChD,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC;QACpC,IAAI,IAAI,KAAK,EAAE,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;YAChE,SAAS;QACX,CAAC;QACD,2DAA2D;QAC3D,OAAO,KAAK,CAAC;IACf,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED,SAAS,aAAa,CAAC,QAAgB;IACrC,MAAM,KAAK,GAAG,QAAQ,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;IAClC,MAAM,QAAQ,GAAG,KAAK,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;IAC/C,OAAO,QAAQ,CAAC,OAAO,CAAC,aAAa,EAAE,EAAE,CAAC,CAAC;AAC7C,CAAC;AAED,MAAM,UAAU,kBAAkB;IAChC,OAAO;QACL,IAAI,EAAE,QAAQ;QACd,mBAAmB,EAAE,iBAAiB;QAEtC,KAAK,CAAC,OAAe,EAAE,QAAgB;YACrC,MAAM,UAAU,GAAG,iBAAiB,CAAC,OAAO,CAAC,CAAC;YAC9C,MAAM,OAAO,GAAkB,EAAE,CAAC;YAElC,KAAK,MAAM,SAAS,IAAI,UAAU,EAAE,CAAC;gBACnC,MAAM,UAAU,GAAG,eAAe,CAChC,SAAS,CAAC,OAAO,EACjB,SAAS,CAAC,SAAS,CACpB,CAAC;gBAEF,IAAI,CAAC,UAAU,CAAC,QAAQ,EAAE,CAAC;oBACzB,SAAS;gBACX,CAAC;gBAED,MAAM,QAAQ,GAAG,UAAU,CAAC,QAAQ,CAAC;gBAErC,oBAAoB;gBACpB,MAAM,UAAU,GAAG,oBAAoB,CAAC,OAAO,EAAE,SAAS,CAAC,SAAS,CAAC,CAAC;gBAEtE,IAAI,UAAU,EAAE,CAAC;oBACf,OAAO,CAAC,IAAI,CAAC;wBACX,IAAI,EAAE,UAAU,CAAC,IAAI;wBACrB,QAAQ;wBACR,IAAI,EAAE,UAAU,CAAC,IAAI;wBACrB,MAAM,EAAE,CAAC;wBACT,QAAQ,EAAE,QAAQ;wBAClB,UAAU,EAAE,QAAQ,CAAC,IAAI;wBACzB,QAAQ;wBACR,YAAY,EAAE,SAAS,CAAC,OAAO;wBAC/B,SAAS,EAAE,UAAU,CAAC,SAAS;wBAC/B,MAAM,EAAE,UAAU,CAAC,MAAM;qBAC1B,CAAC,CAAC;gBACL,CAAC;qBAAM,IAAI,sBAAsB,CAAC,OAAO,EAAE,SAAS,CAAC,SAAS,CAAC,EAAE,CAAC;oBAChE,OAAO,CAAC,IAAI,CAAC;wBACX,IAAI,EAAE,aAAa,CAAC,QAAQ,CAAC;wBAC7B,QAAQ;wBACR,IAAI,EAAE,SAAS,CAAC,SAAS;wBACzB,MAAM,EAAE,CAAC;wBACT,QAAQ,EAAE,QAAQ;wBAClB,UAAU,EAAE,QAAQ,CAAC,IAAI;wBACzB,QAAQ;wBACR,YAAY,EAAE,SAAS,CAAC,OAAO;qBAChC,CAAC,CAAC;gBACL,CAAC;qBAAM,CAAC;oBACN,8DAA8D;oBAC9D,OAAO,CAAC,IAAI,CAAC;wBACX,IAAI,EAAE,SAAS;wBACf,QAAQ;wBACR,IAAI,EAAE,SAAS,CAAC,SAAS;wBACzB,MAAM,EAAE,CAAC;wBACT,QAAQ,EAAE,QAAQ;wBAClB,UAAU,EAAE,QAAQ,CAAC,IAAI;wBACzB,QAAQ;wBACR,YAAY,EAAE,SAAS,CAAC,OAAO;qBAChC,CAAC,CAAC;gBACL,CAAC;YACH,CAAC;YAED,OAAO,OAAO,CAAC;QACjB,CAAC;KACF,CAAC;AACJ,CAAC"}

package/dist/parsers/registry.d.ts

@@ -0,0 +1,3 @@
+import type { ParserRegistry } from './types.js';
+export declare function createDefaultRegistry(): ParserRegistry;
+//# sourceMappingURL=registry.d.ts.map

package/dist/parsers/registry.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"registry.d.ts","sourceRoot":"","sources":["../../src/parsers/registry.ts"],"names":[],"mappings":"AAYA,OAAO,KAAK,EAAU,cAAc,EAAE,MAAM,YAAY,CAAC;AAsCzD,wBAAgB,qBAAqB,IAAI,cAAc,CAOtD"}

package/dist/parsers/registry.js

@@ -0,0 +1,41 @@
+import { createPythonParser } from './python-parser.js';
+import { createTypescriptParser } from './typescript-parser.js';
+import { createGenericParser } from './generic-parser.js';
+import { createGoParser } from './go-parser.js';
+import { createJavaParser } from './java-parser.js';
+function getExtension(filePath) {
+    const lastDot = filePath.lastIndexOf('.');
+    if (lastDot === -1)
+        return '';
+    return filePath.slice(lastDot);
+}
+function createRegistry() {
+    const parsers = [];
+    const genericParser = createGenericParser();
+    return {
+        register(parser) {
+            parsers.push(parser);
+        },
+        getParser(filePath) {
+            const ext = getExtension(filePath);
+            const specific = parsers.find((p) => p.supportedExtensions.includes(ext));
+            return specific ?? genericParser;
+        },
+        parseFile(content, filePath) {
+            const parser = this.getParser(filePath);
+            if (!parser) {
+                return [];
+            }
+            return parser.parse(content, filePath);
+        },
+    };
+}
+export function createDefaultRegistry() {
+    const registry = createRegistry();
+    registry.register(createPythonParser());
+    registry.register(createTypescriptParser());
+    registry.register(createGoParser());
+    registry.register(createJavaParser());
+    return registry;
+}
+//# sourceMappingURL=registry.js.map

package/dist/parsers/registry.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"registry.js","sourceRoot":"","sources":["../../src/parsers/registry.ts"],"names":[],"mappings":"AAaA,OAAO,EAAE,kBAAkB,EAAE,MAAM,oBAAoB,CAAC;AACxD,OAAO,EAAE,sBAAsB,EAAE,MAAM,wBAAwB,CAAC;AAChE,OAAO,EAAE,mBAAmB,EAAE,MAAM,qBAAqB,CAAC;AAC1D,OAAO,EAAE,cAAc,EAAE,MAAM,gBAAgB,CAAC;AAChD,OAAO,EAAE,gBAAgB,EAAE,MAAM,kBAAkB,CAAC;AAEpD,SAAS,YAAY,CAAC,QAAgB;IACpC,MAAM,OAAO,GAAG,QAAQ,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;IAC1C,IAAI,OAAO,KAAK,CAAC,CAAC;QAAE,OAAO,EAAE,CAAC;IAC9B,OAAO,QAAQ,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;AACjC,CAAC;AAED,SAAS,cAAc;IACrB,MAAM,OAAO,GAAa,EAAE,CAAC;IAC7B,MAAM,aAAa,GAAG,mBAAmB,EAAE,CAAC;IAE5C,OAAO;QACL,QAAQ,CAAC,MAAc;YACrB,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QACvB,CAAC;QAED,SAAS,CAAC,QAAgB;YACxB,MAAM,GAAG,GAAG,YAAY,CAAC,QAAQ,CAAC,CAAC;YACnC,MAAM,QAAQ,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,mBAAmB,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC;YAC1E,OAAO,QAAQ,IAAI,aAAa,CAAC;QACnC,CAAC;QAED,SAAS,CAAC,OAAe,EAAE,QAAgB;YACzC,MAAM,MAAM,GAAG,IAAI,CAAC,SAAS,CAAC,QAAQ,CAAC,CAAC;YACxC,IAAI,CAAC,MAAM,EAAE,CAAC;gBACZ,OAAO,EAAE,CAAC;YACZ,CAAC;YACD,OAAO,MAAM,CAAC,KAAK,CAAC,OAAO,EAAE,QAAQ,CAAC,CAAC;QACzC,CAAC;KACF,CAAC;AACJ,CAAC;AAED,MAAM,UAAU,qBAAqB;IACnC,MAAM,QAAQ,GAAG,cAAc,EAAE,CAAC;IAClC,QAAQ,CAAC,QAAQ,CAAC,kBAAkB,EAAE,CAAC,CAAC;IACxC,QAAQ,CAAC,QAAQ,CAAC,sBAAsB,EAAE,CAAC,CAAC;IAC5C,QAAQ,CAAC,QAAQ,CAAC,cAAc,EAAE,CAAC,CAAC;IACpC,QAAQ,CAAC,QAAQ,CAAC,gBAAgB,EAAE,CAAC,CAAC;IACtC,OAAO,QAAQ,CAAC;AAClB,CAAC"}

package/dist/parsers/types.d.ts

@@ -0,0 +1,23 @@
+/**
+ * @knowgraph
+ * type: interface
+ * description: Parser and ParserRegistry interfaces for the language parser system
+ * owner: knowgraph-core
+ * status: stable
+ * tags: [parser, types, interface, registry]
+ * context:
+ *   business_goal: Define contracts for pluggable language parser implementations
+ *   domain: parser-engine
+ */
+import type { ParseResult } from '../types/parse-result.js';
+export interface Parser {
+    readonly name: string;
+    readonly supportedExtensions: readonly string[];
+    parse(content: string, filePath: string): readonly ParseResult[];
+}
+export interface ParserRegistry {
+    register(parser: Parser): void;
+    getParser(filePath: string): Parser | undefined;
+    parseFile(content: string, filePath: string): readonly ParseResult[];
+}
+//# sourceMappingURL=types.d.ts.map

package/dist/parsers/types.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"types.d.ts","sourceRoot":"","sources":["../../src/parsers/types.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;GAUG;AACH,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,0BAA0B,CAAC;AAE5D,MAAM,WAAW,MAAM;IACrB,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB,QAAQ,CAAC,mBAAmB,EAAE,SAAS,MAAM,EAAE,CAAC;IAChD,KAAK,CAAC,OAAO,EAAE,MAAM,EAAE,QAAQ,EAAE,MAAM,GAAG,SAAS,WAAW,EAAE,CAAC;CAClE;AAED,MAAM,WAAW,cAAc;IAC7B,QAAQ,CAAC,MAAM,EAAE,MAAM,GAAG,IAAI,CAAC;IAC/B,SAAS,CAAC,QAAQ,EAAE,MAAM,GAAG,MAAM,GAAG,SAAS,CAAC;IAChD,SAAS,CAAC,OAAO,EAAE,MAAM,EAAE,QAAQ,EAAE,MAAM,GAAG,SAAS,WAAW,EAAE,CAAC;CACtE"}

package/dist/parsers/types.js

@@ -0,0 +1,2 @@
+export {};
+//# sourceMappingURL=types.js.map

package/dist/parsers/types.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"types.js","sourceRoot":"","sources":["../../src/parsers/types.ts"],"names":[],"mappings":""}

package/dist/parsers/typescript-parser.d.ts

@@ -0,0 +1,3 @@
+import type { Parser } from './types.js';
+export declare function createTypescriptParser(): Parser;
+//# sourceMappingURL=typescript-parser.d.ts.map

package/dist/parsers/typescript-parser.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"typescript-parser.d.ts","sourceRoot":"","sources":["../../src/parsers/typescript-parser.ts"],"names":[],"mappings":"AAYA,OAAO,KAAK,EAAE,MAAM,EAAE,MAAM,YAAY,CAAC;AAyMzC,wBAAgB,sBAAsB,IAAI,MAAM,CAyL/C"}

package/dist/parsers/typescript-parser.js

@@ -0,0 +1,321 @@
+import { extractMetadata } from './metadata-extractor.js';
+const TS_EXTENSIONS = ['.ts', '.tsx', '.js', '.jsx', '.mts', '.cts'];
+/**
+ * Regex to match JSDoc comment blocks: /** ... * /
+ */
+const JSDOC_REGEX = /\/\*\*[\s\S]*?\*\//g;
+/**
+ * Regex to match class declarations after a JSDoc block.
+ */
+const CLASS_DECL_REGEX = /^(?:export\s+)?(?:abstract\s+)?class\s+(\w+)(?:<[^>]*>)?(?:\s+extends\s+\S+)?(?:\s+implements\s+[^{]+)?\s*\{/;
+/**
+ * Regex to match function declarations after a JSDoc block.
+ */
+const FUNC_DECL_REGEX = /^(?:export\s+)?(?:async\s+)?function\s+(\w+)(?:<[^>]*>)?\s*\(([^)]*)\)(?:\s*:\s*([^{;]+))?/;
+/**
+ * Regex to match arrow function / const assigned function.
+ */
+const ARROW_FUNC_REGEX = /^(?:export\s+)?(?:const|let|var)\s+(\w+)(?:\s*:\s*[^=]+)?\s*=\s*(?:async\s+)?(?:\([^)]*\)|[^=]+)\s*=>/;
+/**
+ * Regex to match interface declarations.
+ */
+const INTERFACE_DECL_REGEX = /^(?:export\s+)?interface\s+(\w+)(?:<[^>]*>)?(?:\s+extends\s+[^{]+)?\s*\{/;
+/**
+ * Regex to match type alias declarations.
+ */
+const TYPE_DECL_REGEX = /^(?:export\s+)?type\s+(\w+)(?:<[^>]*>)?\s*=/;
+/**
+ * Regex to match method declarations inside a class.
+ */
+const METHOD_DECL_REGEX = /^\s+(?:(?:public|private|protected|static|async|readonly|override|abstract)\s+)*(\w+)(?:<[^>]*>)?\s*\(([^)]*)\)(?:\s*:\s*([^{;]+))?/;
+/**
+ * Regex to match enum declarations.
+ */
+const ENUM_DECL_REGEX = /^(?:export\s+)?(?:const\s+)?enum\s+(\w+)\s*\{/;
+function getLineNumber(source, charIndex) {
+    let line = 1;
+    for (let i = 0; i < charIndex && i < source.length; i++) {
+        if (source[i] === '\n') {
+            line++;
+        }
+    }
+    return line;
+}
+function stripJsdoc(raw) {
+    // Remove /** and */
+    const inner = raw.slice(3, -2);
+    // Remove leading * on each line
+    const lines = inner.split('\n').map((line) => line.replace(/^\s*\*\s?/, ''));
+    return lines.join('\n').trim();
+}
+function findAllJsdocBlocks(content) {
+    const results = [];
+    let match;
+    const regex = new RegExp(JSDOC_REGEX.source, 'g');
+    while ((match = regex.exec(content)) !== null) {
+        const startLine = getLineNumber(content, match.index);
+        const endIndex = match.index + match[0].length;
+        const endLine = getLineNumber(content, endIndex - 1);
+        results.push({
+            content: stripJsdoc(match[0]),
+            startLine,
+            endLine,
+            endIndex,
+        });
+    }
+    return results;
+}
+function getNextNonEmptyLine(content, afterIndex) {
+    const rest = content.slice(afterIndex);
+    const lines = rest.split('\n');
+    for (const line of lines) {
+        const trimmed = line.trim();
+        if (trimmed !== '') {
+            return trimmed;
+        }
+    }
+    return '';
+}
+function getNextNonEmptyLineRaw(content, afterIndex) {
+    const rest = content.slice(afterIndex);
+    const lines = rest.split('\n');
+    for (const line of lines) {
+        if (line.trim() !== '') {
+            return line;
+        }
+    }
+    return '';
+}
+function getNextNonEmptyLineNumber(content, afterIndex) {
+    const baseLine = getLineNumber(content, afterIndex);
+    const rest = content.slice(afterIndex);
+    const lines = rest.split('\n');
+    for (let i = 0; i < lines.length; i++) {
+        if (lines[i]?.trim() !== '') {
+            return baseLine + i;
+        }
+    }
+    return baseLine;
+}
+function isModuleLevelJsdoc(content, startLine) {
+    const lines = content.split('\n');
+    for (let i = 0; i < startLine - 1; i++) {
+        const line = lines[i]?.trim() ?? '';
+        if (line === '' ||
+            line.startsWith('//') ||
+            line.startsWith('/*') ||
+            line.startsWith('*')) {
+            continue;
+        }
+        // Check for imports/requires - those are fine before module-level jsdoc
+        if (line.startsWith('import ') ||
+            line.startsWith('require(') ||
+            line.startsWith("'use strict'") ||
+            line.startsWith('"use strict"')) {
+            continue;
+        }
+        return false;
+    }
+    return true;
+}
+function findEnclosingClassName(content, lineNumber) {
+    const lines = content.split('\n');
+    const targetLine = lines[lineNumber - 1];
+    if (!targetLine)
+        return undefined;
+    const targetIndent = targetLine.search(/\S/);
+    if (targetIndent <= 0)
+        return undefined;
+    for (let i = lineNumber - 2; i >= 0; i--) {
+        const line = lines[i];
+        if (line === undefined)
+            continue;
+        const currentIndent = line.search(/\S/);
+        if (currentIndent === -1)
+            continue;
+        if (currentIndent < targetIndent) {
+            const trimmed = line.trim();
+            const classMatch = trimmed.match(CLASS_DECL_REGEX);
+            if (classMatch) {
+                return classMatch[1];
+            }
+            break;
+        }
+    }
+    return undefined;
+}
+function getModuleName(filePath) {
+    const parts = filePath.split('/');
+    const fileName = parts[parts.length - 1] ?? '';
+    return fileName.replace(/\.(ts|tsx|js|jsx|mts|cts)$/, '');
+}
+export function createTypescriptParser() {
+    return {
+        name: 'typescript',
+        supportedExtensions: TS_EXTENSIONS,
+        parse(content, filePath) {
+            const jsdocBlocks = findAllJsdocBlocks(content);
+            const results = [];
+            for (const jsdoc of jsdocBlocks) {
+                const extraction = extractMetadata(jsdoc.content, jsdoc.startLine);
+                if (!extraction.metadata) {
+                    continue;
+                }
+                const metadata = extraction.metadata;
+                const nextLine = getNextNonEmptyLine(content, jsdoc.endIndex);
+                const nextLineRaw = getNextNonEmptyLineRaw(content, jsdoc.endIndex);
+                const nextLineNumber = getNextNonEmptyLineNumber(content, jsdoc.endIndex);
+                // Try to match class declaration
+                const classMatch = nextLine.match(CLASS_DECL_REGEX);
+                if (classMatch) {
+                    results.push({
+                        name: classMatch[1],
+                        filePath,
+                        line: nextLineNumber,
+                        column: 1,
+                        language: 'typescript',
+                        entityType: metadata.type,
+                        metadata,
+                        rawDocstring: jsdoc.content,
+                    });
+                    continue;
+                }
+                // Try to match function declaration
+                const funcMatch = nextLine.match(FUNC_DECL_REGEX);
+                if (funcMatch) {
+                    const name = funcMatch[1];
+                    const params = funcMatch[2];
+                    const returnType = funcMatch[3]?.trim();
+                    const sigParts = [`function ${name}(${params})`];
+                    if (returnType) {
+                        sigParts.push(`: ${returnType}`);
+                    }
+                    results.push({
+                        name,
+                        filePath,
+                        line: nextLineNumber,
+                        column: 1,
+                        language: 'typescript',
+                        entityType: metadata.type,
+                        metadata,
+                        rawDocstring: jsdoc.content,
+                        signature: sigParts.join(''),
+                    });
+                    continue;
+                }
+                // Try to match arrow function
+                const arrowMatch = nextLine.match(ARROW_FUNC_REGEX);
+                if (arrowMatch) {
+                    results.push({
+                        name: arrowMatch[1],
+                        filePath,
+                        line: nextLineNumber,
+                        column: 1,
+                        language: 'typescript',
+                        entityType: metadata.type,
+                        metadata,
+                        rawDocstring: jsdoc.content,
+                    });
+                    continue;
+                }
+                // Try to match interface declaration
+                const ifaceMatch = nextLine.match(INTERFACE_DECL_REGEX);
+                if (ifaceMatch) {
+                    results.push({
+                        name: ifaceMatch[1],
+                        filePath,
+                        line: nextLineNumber,
+                        column: 1,
+                        language: 'typescript',
+                        entityType: metadata.type,
+                        metadata,
+                        rawDocstring: jsdoc.content,
+                    });
+                    continue;
+                }
+                // Try to match type alias
+                const typeMatch = nextLine.match(TYPE_DECL_REGEX);
+                if (typeMatch) {
+                    results.push({
+                        name: typeMatch[1],
+                        filePath,
+                        line: nextLineNumber,
+                        column: 1,
+                        language: 'typescript',
+                        entityType: metadata.type,
+                        metadata,
+                        rawDocstring: jsdoc.content,
+                    });
+                    continue;
+                }
+                // Try to match enum declaration
+                const enumMatch = nextLine.match(ENUM_DECL_REGEX);
+                if (enumMatch) {
+                    results.push({
+                        name: enumMatch[1],
+                        filePath,
+                        line: nextLineNumber,
+                        column: 1,
+                        language: 'typescript',
+                        entityType: metadata.type,
+                        metadata,
+                        rawDocstring: jsdoc.content,
+                    });
+                    continue;
+                }
+                // Try to match method declaration (inside a class)
+                const methodMatch = nextLineRaw.match(METHOD_DECL_REGEX);
+                if (methodMatch) {
+                    const name = methodMatch[1];
+                    const params = methodMatch[2];
+                    const returnType = methodMatch[3]?.trim();
+                    const parent = findEnclosingClassName(content, nextLineNumber);
+                    const sigParts = [`${name}(${params})`];
+                    if (returnType) {
+                        sigParts.push(`: ${returnType}`);
+                    }
+                    results.push({
+                        name,
+                        filePath,
+                        line: nextLineNumber,
+                        column: 1,
+                        language: 'typescript',
+                        entityType: metadata.type,
+                        metadata,
+                        rawDocstring: jsdoc.content,
+                        signature: sigParts.join(''),
+                        parent,
+                    });
+                    continue;
+                }
+                // Module-level JSDoc or unrecognized target
+                if (isModuleLevelJsdoc(content, jsdoc.startLine)) {
+                    results.push({
+                        name: getModuleName(filePath),
+                        filePath,
+                        line: jsdoc.startLine,
+                        column: 1,
+                        language: 'typescript',
+                        entityType: metadata.type,
+                        metadata,
+                        rawDocstring: jsdoc.content,
+                    });
+                }
+                else {
+                    results.push({
+                        name: 'unknown',
+                        filePath,
+                        line: jsdoc.startLine,
+                        column: 1,
+                        language: 'typescript',
+                        entityType: metadata.type,
+                        metadata,
+                        rawDocstring: jsdoc.content,
+                    });
+                }
+            }
+            return results;
+        },
+    };
+}
+//# sourceMappingURL=typescript-parser.js.map