@opencodehub/ingestion 0.1.0

package/dist/extract/tool-detector.js

@@ -0,0 +1,364 @@
+/**
+ * MCP / JSON-RPC tool detector.
+ *
+ * The heuristic is intentionally coarse: many agent frameworks (MCP, LangChain
+ * tools, OpenAI function schemas) converge on a `{ name, description, ... }`
+ * object literal for each tool. At MVP we only claim a match when the file
+ * path *also* signals "tool-ness", which eliminates the majority of false
+ * positives coming from generic config blocks.
+ *
+ * Pattern authored fresh — no external regex reused.
+ */
+/** Path-based gate: file must look like a tool definition. */
+function filePathLooksLikeTool(filePath) {
+    const normalized = filePath.replace(/\\/g, "/").toLowerCase();
+    if (normalized.includes("/tool"))
+        return true;
+    const basename = normalized.slice(normalized.lastIndexOf("/") + 1);
+    // `*-tool.ts`, `*_tool.ts`, or `*tool*.ts` (and JS/MJS siblings).
+    if (/(^|[-_])tool([-_.]|s?\.)/.test(basename))
+        return true;
+    if (/tool/.test(basename))
+        return true;
+    return false;
+}
+/**
+ * Match `name: "..."`, `"name": '...'`, `name: \`...\`` (no interpolation).
+ * The key itself is captured so the same regex can harvest `description:`.
+ */
+const STRING_PROP_RE = /(?:^|[\s,{(])(?:"([a-zA-Z_][\w]*)"|'([a-zA-Z_][\w]*)'|([a-zA-Z_][\w]*))\s*:\s*(?:"([^"\\]*(?:\\.[^"\\]*)*)"|'([^'\\]*(?:\\.[^'\\]*)*)'|`([^`$]*)`)/g;
+function collectStringProps(content, keyName) {
+    const hits = [];
+    STRING_PROP_RE.lastIndex = 0;
+    let match = STRING_PROP_RE.exec(content);
+    while (match !== null) {
+        const key = match[1] ?? match[2] ?? match[3];
+        const value = match[4] ?? match[5] ?? match[6];
+        const idx = match.index;
+        match = STRING_PROP_RE.exec(content);
+        if (key !== keyName || value === undefined)
+            continue;
+        // Count newlines in prefix to convert byte-offset -> 1-based line number.
+        let line = 1;
+        for (let i = 0; i < idx; i++) {
+            if (content.charCodeAt(i) === 10)
+                line += 1;
+        }
+        hits.push({ line, value });
+    }
+    return hits;
+}
+/**
+ * Match `inputSchema:` followed by a `{ ... }` object literal. The capture
+ * begins at the opening brace; the closing brace is resolved by
+ * {@link findBalancedObjectEnd} so nested objects and strings don't confuse
+ * the match.
+ */
+const INPUT_SCHEMA_RE = /(?:^|[\s,{(])(?:"inputSchema"|'inputSchema'|inputSchema)\s*:\s*\{/g;
+function collectInputSchemaLiterals(content) {
+    const hits = [];
+    INPUT_SCHEMA_RE.lastIndex = 0;
+    let match = INPUT_SCHEMA_RE.exec(content);
+    while (match !== null) {
+        const openBrace = match.index + match[0].length - 1;
+        const close = findBalancedObjectEnd(content, openBrace);
+        const idx = match.index;
+        // Resume scanning after the open brace so a malformed literal cannot
+        // wedge the loop.
+        INPUT_SCHEMA_RE.lastIndex = openBrace + 1;
+        match = INPUT_SCHEMA_RE.exec(content);
+        if (close === -1)
+            continue;
+        const rawObjectLiteral = content.slice(openBrace, close + 1);
+        let line = 1;
+        for (let i = 0; i < idx; i++) {
+            if (content.charCodeAt(i) === 10)
+                line += 1;
+        }
+        hits.push({ line, rawObjectLiteral });
+    }
+    return hits;
+}
+/**
+ * Walk forward from `start` (pointing at `{`) and return the index of the
+ * matching `}`, or -1 when the brace is unbalanced / the source runs out.
+ * Single-, double-, and backtick-string literals are skipped so a brace
+ * inside a string doesn't fool depth counting.
+ */
+function findBalancedObjectEnd(source, start) {
+    if (source.charCodeAt(start) !== 0x7b /* { */)
+        return -1;
+    let depth = 0;
+    let i = start;
+    const n = source.length;
+    while (i < n) {
+        const ch = source.charCodeAt(i);
+        if (ch === 0x22 /* " */ || ch === 0x27 /* ' */ || ch === 0x60 /* ` */) {
+            const next = skipString(source, i, ch);
+            if (next === -1)
+                return -1;
+            i = next;
+            continue;
+        }
+        if (ch === 0x7b /* { */) {
+            depth += 1;
+        }
+        else if (ch === 0x7d /* } */) {
+            depth -= 1;
+            if (depth === 0)
+                return i;
+        }
+        i += 1;
+    }
+    return -1;
+}
+function skipString(source, start, quote) {
+    const n = source.length;
+    let i = start + 1;
+    while (i < n) {
+        const ch = source.charCodeAt(i);
+        if (ch === 0x5c /* \ */) {
+            // Skip escape sequence (conservatively consume one trailing char).
+            i += 2;
+            continue;
+        }
+        if (ch === quote)
+            return i + 1;
+        i += 1;
+    }
+    return -1;
+}
+/**
+ * Translate an object literal written in relaxed JS syntax (single quotes,
+ * unquoted keys, trailing commas) into strict canonical JSON sorted by key.
+ * Returns `undefined` when the literal is too exotic to normalize safely
+ * (template strings, bareword values, computed keys).
+ */
+export function canonicalizeObjectLiteral(literal) {
+    const jsonLike = relaxedToJson(literal);
+    if (jsonLike === undefined)
+        return undefined;
+    let parsed;
+    try {
+        parsed = JSON.parse(jsonLike);
+    }
+    catch {
+        return undefined;
+    }
+    return stableStringify(parsed);
+}
+function stableStringify(value) {
+    if (value === null || typeof value !== "object") {
+        return JSON.stringify(value);
+    }
+    if (Array.isArray(value)) {
+        return `[${value.map((v) => stableStringify(v)).join(",")}]`;
+    }
+    const entries = Object.entries(value)
+        .filter(([, v]) => v !== undefined)
+        .sort(([a], [b]) => (a < b ? -1 : a > b ? 1 : 0));
+    return `{${entries.map(([k, v]) => `${JSON.stringify(k)}:${stableStringify(v)}`).join(",")}}`;
+}
+/**
+ * Rewrite a JS-ish object/array literal so `JSON.parse` accepts it:
+ *   - single-quoted strings to double-quoted.
+ *   - unquoted identifier keys to double-quoted keys.
+ *   - trailing commas stripped.
+ * Template strings and bareword values produce `undefined` so the caller
+ * falls back to "no schema captured".
+ */
+function relaxedToJson(literal) {
+    let out = "";
+    let i = 0;
+    const n = literal.length;
+    while (i < n) {
+        const ch = literal[i];
+        if (ch === '"') {
+            const end = findStringEnd(literal, i, 0x22);
+            if (end === -1)
+                return undefined;
+            out += literal.slice(i, end + 1);
+            i = end + 1;
+            continue;
+        }
+        if (ch === "'") {
+            const end = findStringEnd(literal, i, 0x27);
+            if (end === -1)
+                return undefined;
+            out += `"${jsSingleQuotedToJsonInner(literal.slice(i + 1, end))}"`;
+            i = end + 1;
+            continue;
+        }
+        if (ch === "`")
+            return undefined;
+        if (ch !== undefined && /[A-Za-z_$]/.test(ch)) {
+            let j = i;
+            while (j < n) {
+                const cj = literal[j];
+                if (cj === undefined || !/[A-Za-z0-9_$]/.test(cj))
+                    break;
+                j += 1;
+            }
+            const ident = literal.slice(i, j);
+            if (ident === "true" || ident === "false" || ident === "null") {
+                out += ident;
+                i = j;
+                continue;
+            }
+            let k = j;
+            while (k < n && /\s/.test(literal[k] ?? ""))
+                k += 1;
+            if (literal[k] === ":") {
+                out += `"${ident}"`;
+                i = j;
+                continue;
+            }
+            return undefined;
+        }
+        if (ch === ",") {
+            let k = i + 1;
+            while (k < n && /\s/.test(literal[k] ?? ""))
+                k += 1;
+            const next = literal[k];
+            if (next === "}" || next === "]") {
+                i = k;
+                continue;
+            }
+        }
+        out += ch ?? "";
+        i += 1;
+    }
+    return out;
+}
+/**
+ * Translate the *inside* of a JS single-quoted string literal into the
+ * inside of a JSON double-quoted string literal, character by character:
+ *
+ *   - `\'` (a JS-only escape) becomes `'` — not legal inside a JSON
+ *     double-quoted string.
+ *   - JSON-recognized escapes (`\"`, `\\`, `\/`, `\b`, `\f`, `\n`, `\r`,
+ *     `\t`, `\uXXXX`) pass through unchanged.
+ *   - Any other `\X` JS escape that JSON does not understand has its
+ *     leading backslash doubled so the parser sees the literal characters.
+ *   - A bare `"` is escaped to `\"`.
+ *
+ * The character-by-character pass replaces a chained `replace()` sequence
+ * that doubled every `\` and broke valid escapes like `\n`. Without the
+ * pass, an input containing `\"` would have produced malformed JSON —
+ * the js/incomplete-sanitization defect.
+ */
+function jsSingleQuotedToJsonInner(inner) {
+    const JSON_SIMPLE_ESCAPE = /^["\\/bfnrt]$/;
+    const HEX = /^[0-9a-fA-F]$/;
+    let out = "";
+    for (let i = 0; i < inner.length; i += 1) {
+        const ch = inner[i];
+        if (ch === "\\") {
+            const next = inner[i + 1] ?? "";
+            if (next === "'") {
+                // JS-only escape — drop the backslash, keep the quote.
+                out += "'";
+                i += 1;
+                continue;
+            }
+            if (JSON_SIMPLE_ESCAPE.test(next)) {
+                // Pass `\\`, `\"`, `\/`, `\b`, `\f`, `\n`, `\r`, `\t` through.
+                out += `\\${next}`;
+                i += 1;
+                continue;
+            }
+            if (next === "u" &&
+                HEX.test(inner[i + 2] ?? "") &&
+                HEX.test(inner[i + 3] ?? "") &&
+                HEX.test(inner[i + 4] ?? "") &&
+                HEX.test(inner[i + 5] ?? "")) {
+                out += inner.slice(i, i + 6);
+                i += 5;
+                continue;
+            }
+            // Unknown JS escape (e.g. `\x41`, `\0`) or a stray backslash —
+            // double it so the literal `\` survives the JSON parser.
+            out += "\\\\";
+            continue;
+        }
+        if (ch === '"') {
+            out += '\\"';
+            continue;
+        }
+        out += ch;
+    }
+    return out;
+}
+function findStringEnd(src, start, quote) {
+    let i = start + 1;
+    const n = src.length;
+    while (i < n) {
+        const ch = src.charCodeAt(i);
+        if (ch === 0x5c /* \ */) {
+            i += 2;
+            continue;
+        }
+        if (ch === quote)
+            return i;
+        i += 1;
+    }
+    return -1;
+}
+/**
+ * Emit one `ExtractedTool` per `{ name, description }` literal pair found in
+ * the same 5-line window. Duplicates (same `toolName` + `description`) are
+ * de-duplicated per file; cross-file de-duplication is a later phase concern.
+ */
+export function detectMcpTools(input) {
+    const { filePath, content } = input;
+    if (!filePathLooksLikeTool(filePath))
+        return [];
+    if (!/\bname\b/.test(content) || !/\bdescription\b/.test(content))
+        return [];
+    const names = collectStringProps(content, "name");
+    const descriptions = collectStringProps(content, "description");
+    if (names.length === 0 || descriptions.length === 0)
+        return [];
+    const schemas = collectInputSchemaLiterals(content);
+    const seen = new Set();
+    const out = [];
+    for (const name of names) {
+        // Pick the nearest description within 5 lines in either direction.
+        let best;
+        let bestDistance = Number.POSITIVE_INFINITY;
+        for (const desc of descriptions) {
+            const distance = Math.abs(desc.line - name.line);
+            if (distance <= 5 && distance < bestDistance) {
+                best = desc;
+                bestDistance = distance;
+            }
+        }
+        if (best === undefined)
+            continue;
+        const key = `${name.value}\u0000${best.value}`;
+        if (seen.has(key))
+            continue;
+        seen.add(key);
+        // Pair the nearest inputSchema literal within a 10-line window. Schemas
+        // often live a few lines below the name/description pair so we allow a
+        // wider span than description pairing.
+        let bestSchema;
+        let bestSchemaDistance = Number.POSITIVE_INFINITY;
+        for (const s of schemas) {
+            const distance = Math.abs(s.line - name.line);
+            if (distance <= 10 && distance < bestSchemaDistance) {
+                bestSchema = s;
+                bestSchemaDistance = distance;
+            }
+        }
+        const inputSchemaJson = bestSchema !== undefined ? canonicalizeObjectLiteral(bestSchema.rawObjectLiteral) : undefined;
+        out.push({
+            toolName: name.value,
+            handlerFile: filePath,
+            description: best.value,
+            ...(inputSchemaJson !== undefined ? { inputSchemaJson } : {}),
+        });
+    }
+    return out;
+}
+//# sourceMappingURL=tool-detector.js.map

package/dist/extract/tool-detector.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"tool-detector.js","sourceRoot":"","sources":["../../src/extract/tool-detector.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;GAUG;AAIH,8DAA8D;AAC9D,SAAS,qBAAqB,CAAC,QAAgB;IAC7C,MAAM,UAAU,GAAG,QAAQ,CAAC,OAAO,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC,WAAW,EAAE,CAAC;IAC9D,IAAI,UAAU,CAAC,QAAQ,CAAC,OAAO,CAAC;QAAE,OAAO,IAAI,CAAC;IAC9C,MAAM,QAAQ,GAAG,UAAU,CAAC,KAAK,CAAC,UAAU,CAAC,WAAW,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC;IACnE,kEAAkE;IAClE,IAAI,0BAA0B,CAAC,IAAI,CAAC,QAAQ,CAAC;QAAE,OAAO,IAAI,CAAC;IAC3D,IAAI,MAAM,CAAC,IAAI,CAAC,QAAQ,CAAC;QAAE,OAAO,IAAI,CAAC;IACvC,OAAO,KAAK,CAAC;AACf,CAAC;AAQD;;;GAGG;AACH,MAAM,cAAc,GAClB,qJAAqJ,CAAC;AAExJ,SAAS,kBAAkB,CAAC,OAAe,EAAE,OAAe;IAC1D,MAAM,IAAI,GAAmB,EAAE,CAAC;IAChC,cAAc,CAAC,SAAS,GAAG,CAAC,CAAC;IAC7B,IAAI,KAAK,GAA2B,cAAc,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;IACjE,OAAO,KAAK,KAAK,IAAI,EAAE,CAAC;QACtB,MAAM,GAAG,GAAG,KAAK,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC;QAC7C,MAAM,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC;QAC/C,MAAM,GAAG,GAAG,KAAK,CAAC,KAAK,CAAC;QACxB,KAAK,GAAG,cAAc,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;QACrC,IAAI,GAAG,KAAK,OAAO,IAAI,KAAK,KAAK,SAAS;YAAE,SAAS;QACrD,0EAA0E;QAC1E,IAAI,IAAI,GAAG,CAAC,CAAC;QACb,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,GAAG,EAAE,CAAC,EAAE,EAAE,CAAC;YAC7B,IAAI,OAAO,CAAC,UAAU,CAAC,CAAC,CAAC,KAAK,EAAE;gBAAE,IAAI,IAAI,CAAC,CAAC;QAC9C,CAAC;QACD,IAAI,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,KAAK,EAAE,CAAC,CAAC;IAC7B,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED;;;;;GAKG;AACH,MAAM,eAAe,GAAG,oEAAoE,CAAC;AAO7F,SAAS,0BAA0B,CAAC,OAAe;IACjD,MAAM,IAAI,GAAqB,EAAE,CAAC;IAClC,eAAe,CAAC,SAAS,GAAG,CAAC,CAAC;IAC9B,IAAI,KAAK,GAA2B,eAAe,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;IAClE,OAAO,KAAK,KAAK,IAAI,EAAE,CAAC;QACtB,MAAM,SAAS,GAAG,KAAK,CAAC,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC;QACpD,MAAM,KAAK,GAAG,qBAAqB,CAAC,OAAO,EAAE,SAAS,CAAC,CAAC;QACxD,MAAM,GAAG,GAAG,KAAK,CAAC,KAAK,CAAC;QACxB,qEAAqE;QACrE,kBAAkB;QAClB,eAAe,CAAC,SAAS,GAAG,SAAS,GAAG,CAAC,CAAC;QAC1C,KAAK,GAAG,eAAe,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;QACtC,IAAI,KAAK,KAAK,CAAC,CAAC;YAAE,SAAS;QAC3B,MAAM,gBAAgB,GAAG,OAAO,CAAC,KAAK,CAAC,SAAS,EAAE,KAAK,GAAG,CAAC,CAAC,CAAC;QAC7D,IAAI,IAAI,GAAG,CAAC,CAAC;QACb,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,GAAG,EAAE,CAAC,EAAE,EAAE,CAAC;YAC7B,IAAI,OAAO,CAAC,UAAU,CAAC,CAAC,CAAC,KAAK,EAAE;gBAAE,IAAI,IAAI,CAAC,CAAC;QAC9C,CAAC;QACD,IAAI,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,gBAAgB,EAAE,CAAC,CAAC;IACxC,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED;;;;;GAKG;AACH,SAAS,qBAAqB,CAAC,MAAc,EAAE,KAAa;IAC1D,IAAI,MAAM,CAAC,UAAU,CAAC,KAAK,CAAC,KAAK,IAAI,CAAC,OAAO;QAAE,OAAO,CAAC,CAAC,CAAC;IACzD,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,IAAI,CAAC,GAAG,KAAK,CAAC;IACd,MAAM,CAAC,GAAG,MAAM,CAAC,MAAM,CAAC;IACxB,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC;QACb,MAAM,EAAE,GAAG,MAAM,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC;QAChC,IAAI,EAAE,KAAK,IAAI,CAAC,OAAO,IAAI,EAAE,KAAK,IAAI,CAAC,OAAO,IAAI,EAAE,KAAK,IAAI,CAAC,OAAO,EAAE,CAAC;YACtE,MAAM,IAAI,GAAG,UAAU,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC;YACvC,IAAI,IAAI,KAAK,CAAC,CAAC;gBAAE,OAAO,CAAC,CAAC,CAAC;YAC3B,CAAC,GAAG,IAAI,CAAC;YACT,SAAS;QACX,CAAC;QACD,IAAI,EAAE,KAAK,IAAI,CAAC,OAAO,EAAE,CAAC;YACxB,KAAK,IAAI,CAAC,CAAC;QACb,CAAC;aAAM,IAAI,EAAE,KAAK,IAAI,CAAC,OAAO,EAAE,CAAC;YAC/B,KAAK,IAAI,CAAC,CAAC;YACX,IAAI,KAAK,KAAK,CAAC;gBAAE,OAAO,CAAC,CAAC;QAC5B,CAAC;QACD,CAAC,IAAI,CAAC,CAAC;IACT,CAAC;IACD,OAAO,CAAC,CAAC,CAAC;AACZ,CAAC;AAED,SAAS,UAAU,CAAC,MAAc,EAAE,KAAa,EAAE,KAAa;IAC9D,MAAM,CAAC,GAAG,MAAM,CAAC,MAAM,CAAC;IACxB,IAAI,CAAC,GAAG,KAAK,GAAG,CAAC,CAAC;IAClB,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC;QACb,MAAM,EAAE,GAAG,MAAM,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC;QAChC,IAAI,EAAE,KAAK,IAAI,CAAC,OAAO,EAAE,CAAC;YACxB,mEAAmE;YACnE,CAAC,IAAI,CAAC,CAAC;YACP,SAAS;QACX,CAAC;QACD,IAAI,EAAE,KAAK,KAAK;YAAE,OAAO,CAAC,GAAG,CAAC,CAAC;QAC/B,CAAC,IAAI,CAAC,CAAC;IACT,CAAC;IACD,OAAO,CAAC,CAAC,CAAC;AACZ,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,yBAAyB,CAAC,OAAe;IACvD,MAAM,QAAQ,GAAG,aAAa,CAAC,OAAO,CAAC,CAAC;IACxC,IAAI,QAAQ,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IAC7C,IAAI,MAAe,CAAC;IACpB,IAAI,CAAC;QACH,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC;IAChC,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,SAAS,CAAC;IACnB,CAAC;IACD,OAAO,eAAe,CAAC,MAAM,CAAC,CAAC;AACjC,CAAC;AAED,SAAS,eAAe,CAAC,KAAc;IACrC,IAAI,KAAK,KAAK,IAAI,IAAI,OAAO,KAAK,KAAK,QAAQ,EAAE,CAAC;QAChD,OAAO,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC;IAC/B,CAAC;IACD,IAAI,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC;QACzB,OAAO,IAAI,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,eAAe,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC;IAC/D,CAAC;IACD,MAAM,OAAO,GAAG,MAAM,CAAC,OAAO,CAAC,KAAgC,CAAC;SAC7D,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,KAAK,SAAS,CAAC;SAClC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IACpD,OAAO,IAAI,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,GAAG,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,eAAe,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC;AAChG,CAAC;AAED;;;;;;;GAOG;AACH,SAAS,aAAa,CAAC,OAAe;IACpC,IAAI,GAAG,GAAG,EAAE,CAAC;IACb,IAAI,CAAC,GAAG,CAAC,CAAC;IACV,MAAM,CAAC,GAAG,OAAO,CAAC,MAAM,CAAC;IACzB,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC;QACb,MAAM,EAAE,GAAG,OAAO,CAAC,CAAC,CAAC,CAAC;QACtB,IAAI,EAAE,KAAK,GAAG,EAAE,CAAC;YACf,MAAM,GAAG,GAAG,aAAa,CAAC,OAAO,EAAE,CAAC,EAAE,IAAI,CAAC,CAAC;YAC5C,IAAI,GAAG,KAAK,CAAC,CAAC;gBAAE,OAAO,SAAS,CAAC;YACjC,GAAG,IAAI,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,GAAG,CAAC,CAAC,CAAC;YACjC,CAAC,GAAG,GAAG,GAAG,CAAC,CAAC;YACZ,SAAS;QACX,CAAC;QACD,IAAI,EAAE,KAAK,GAAG,EAAE,CAAC;YACf,MAAM,GAAG,GAAG,aAAa,CAAC,OAAO,EAAE,CAAC,EAAE,IAAI,CAAC,CAAC;YAC5C,IAAI,GAAG,KAAK,CAAC,CAAC;gBAAE,OAAO,SAAS,CAAC;YACjC,GAAG,IAAI,IAAI,yBAAyB,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,EAAE,GAAG,CAAC,CAAC,GAAG,CAAC;YACnE,CAAC,GAAG,GAAG,GAAG,CAAC,CAAC;YACZ,SAAS;QACX,CAAC;QACD,IAAI,EAAE,KAAK,GAAG;YAAE,OAAO,SAAS,CAAC;QACjC,IAAI,EAAE,KAAK,SAAS,IAAI,YAAY,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,CAAC;YAC9C,IAAI,CAAC,GAAG,CAAC,CAAC;YACV,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC;gBACb,MAAM,EAAE,GAAG,OAAO,CAAC,CAAC,CAAC,CAAC;gBACtB,IAAI,EAAE,KAAK,SAAS,IAAI,CAAC,eAAe,CAAC,IAAI,CAAC,EAAE,CAAC;oBAAE,MAAM;gBACzD,CAAC,IAAI,CAAC,CAAC;YACT,CAAC;YACD,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;YAClC,IAAI,KAAK,KAAK,MAAM,IAAI,KAAK,KAAK,OAAO,IAAI,KAAK,KAAK,MAAM,EAAE,CAAC;gBAC9D,GAAG,IAAI,KAAK,CAAC;gBACb,CAAC,GAAG,CAAC,CAAC;gBACN,SAAS;YACX,CAAC;YACD,IAAI,CAAC,GAAG,CAAC,CAAC;YACV,OAAO,CAAC,GAAG,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;gBAAE,CAAC,IAAI,CAAC,CAAC;YACpD,IAAI,OAAO,CAAC,CAAC,CAAC,KAAK,GAAG,EAAE,CAAC;gBACvB,GAAG,IAAI,IAAI,KAAK,GAAG,CAAC;gBACpB,CAAC,GAAG,CAAC,CAAC;gBACN,SAAS;YACX,CAAC;YACD,OAAO,SAAS,CAAC;QACnB,CAAC;QACD,IAAI,EAAE,KAAK,GAAG,EAAE,CAAC;YACf,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;YACd,OAAO,CAAC,GAAG,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;gBAAE,CAAC,IAAI,CAAC,CAAC;YACpD,MAAM,IAAI,GAAG,OAAO,CAAC,CAAC,CAAC,CAAC;YACxB,IAAI,IAAI,KAAK,GAAG,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;gBACjC,CAAC,GAAG,CAAC,CAAC;gBACN,SAAS;YACX,CAAC;QACH,CAAC;QACD,GAAG,IAAI,EAAE,IAAI,EAAE,CAAC;QAChB,CAAC,IAAI,CAAC,CAAC;IACT,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAED;;;;;;;;;;;;;;;;GAgBG;AACH,SAAS,yBAAyB,CAAC,KAAa;IAC9C,MAAM,kBAAkB,GAAG,eAAe,CAAC;IAC3C,MAAM,GAAG,GAAG,eAAe,CAAC;IAC5B,IAAI,GAAG,GAAG,EAAE,CAAC;IACb,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;QACzC,MAAM,EAAE,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;QACpB,IAAI,EAAE,KAAK,IAAI,EAAE,CAAC;YAChB,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;YAChC,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;gBACjB,uDAAuD;gBACvD,GAAG,IAAI,GAAG,CAAC;gBACX,CAAC,IAAI,CAAC,CAAC;gBACP,SAAS;YACX,CAAC;YACD,IAAI,kBAAkB,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;gBAClC,+DAA+D;gBAC/D,GAAG,IAAI,KAAK,IAAI,EAAE,CAAC;gBACnB,CAAC,IAAI,CAAC,CAAC;gBACP,SAAS;YACX,CAAC;YACD,IACE,IAAI,KAAK,GAAG;gBACZ,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;gBAC5B,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;gBAC5B,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;gBAC5B,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,EAC5B,CAAC;gBACD,GAAG,IAAI,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;gBAC7B,CAAC,IAAI,CAAC,CAAC;gBACP,SAAS;YACX,CAAC;YACD,+DAA+D;YAC/D,yDAAyD;YACzD,GAAG,IAAI,MAAM,CAAC;YACd,SAAS;QACX,CAAC;QACD,IAAI,EAAE,KAAK,GAAG,EAAE,CAAC;YACf,GAAG,IAAI,KAAK,CAAC;YACb,SAAS;QACX,CAAC;QACD,GAAG,IAAI,EAAE,CAAC;IACZ,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAED,SAAS,aAAa,CAAC,GAAW,EAAE,KAAa,EAAE,KAAa;IAC9D,IAAI,CAAC,GAAG,KAAK,GAAG,CAAC,CAAC;IAClB,MAAM,CAAC,GAAG,GAAG,CAAC,MAAM,CAAC;IACrB,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC;QACb,MAAM,EAAE,GAAG,GAAG,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC;QAC7B,IAAI,EAAE,KAAK,IAAI,CAAC,OAAO,EAAE,CAAC;YACxB,CAAC,IAAI,CAAC,CAAC;YACP,SAAS;QACX,CAAC;QACD,IAAI,EAAE,KAAK,KAAK;YAAE,OAAO,CAAC,CAAC;QAC3B,CAAC,IAAI,CAAC,CAAC;IACT,CAAC;IACD,OAAO,CAAC,CAAC,CAAC;AACZ,CAAC;AAED;;;;GAIG;AACH,MAAM,UAAU,cAAc,CAAC,KAAmB;IAChD,MAAM,EAAE,QAAQ,EAAE,OAAO,EAAE,GAAG,KAAK,CAAC;IACpC,IAAI,CAAC,qBAAqB,CAAC,QAAQ,CAAC;QAAE,OAAO,EAAE,CAAC;IAChD,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,iBAAiB,CAAC,IAAI,CAAC,OAAO,CAAC;QAAE,OAAO,EAAE,CAAC;IAE7E,MAAM,KAAK,GAAG,kBAAkB,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;IAClD,MAAM,YAAY,GAAG,kBAAkB,CAAC,OAAO,EAAE,aAAa,CAAC,CAAC;IAChE,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,IAAI,YAAY,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAE/D,MAAM,OAAO,GAAG,0BAA0B,CAAC,OAAO,CAAC,CAAC;IAEpD,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAC;IAC/B,MAAM,GAAG,GAAoB,EAAE,CAAC;IAChC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,mEAAmE;QACnE,IAAI,IAA8B,CAAC;QACnC,IAAI,YAAY,GAAG,MAAM,CAAC,iBAAiB,CAAC;QAC5C,KAAK,MAAM,IAAI,IAAI,YAAY,EAAE,CAAC;YAChC,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC;YACjD,IAAI,QAAQ,IAAI,CAAC,IAAI,QAAQ,GAAG,YAAY,EAAE,CAAC;gBAC7C,IAAI,GAAG,IAAI,CAAC;gBACZ,YAAY,GAAG,QAAQ,CAAC;YAC1B,CAAC;QACH,CAAC;QACD,IAAI,IAAI,KAAK,SAAS;YAAE,SAAS;QACjC,MAAM,GAAG,GAAG,GAAG,IAAI,CAAC,KAAK,SAAS,IAAI,CAAC,KAAK,EAAE,CAAC;QAC/C,IAAI,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC;YAAE,SAAS;QAC5B,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QACd,wEAAwE;QACxE,uEAAuE;QACvE,uCAAuC;QACvC,IAAI,UAAsC,CAAC;QAC3C,IAAI,kBAAkB,GAAG,MAAM,CAAC,iBAAiB,CAAC;QAClD,KAAK,MAAM,CAAC,IAAI,OAAO,EAAE,CAAC;YACxB,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC;YAC9C,IAAI,QAAQ,IAAI,EAAE,IAAI,QAAQ,GAAG,kBAAkB,EAAE,CAAC;gBACpD,UAAU,GAAG,CAAC,CAAC;gBACf,kBAAkB,GAAG,QAAQ,CAAC;YAChC,CAAC;QACH,CAAC;QACD,MAAM,eAAe,GACnB,UAAU,KAAK,SAAS,CAAC,CAAC,CAAC,yBAAyB,CAAC,UAAU,CAAC,gBAAgB,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;QAChG,GAAG,CAAC,IAAI,CAAC;YACP,QAAQ,EAAE,IAAI,CAAC,KAAK;YACpB,WAAW,EAAE,QAAQ;YACrB,WAAW,EAAE,IAAI,CAAC,KAAK;YACvB,GAAG,CAAC,eAAe,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,eAAe,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;SAC9D,CAAC,CAAC;IACL,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC"}

package/dist/extract/types.d.ts

@@ -0,0 +1,89 @@
+/**
+ * Shared extraction result shapes for the route / tool / ORM detectors.
+ *
+ * These types are deliberately minimal: they describe the evidence a detector
+ * has discovered in a single file (or, for filesystem-routed frameworks, a
+ * group of files) without taking any dependency on the graph schema. The
+ * routes / tools / orm pipeline phases turn these records into graph nodes
+ * and edges.
+ */
+/** A handler-to-URL mapping discovered by a route detector. */
+export interface ExtractedRoute {
+    /** Canonical URL path with `{param}` / `{+catchAll}` placeholders. */
+    readonly url: string;
+    /** HTTP method in uppercase, or `undefined` for filesystem pages. */
+    readonly method?: string;
+    /** Repo-relative path to the file that declared the route. */
+    readonly handlerFile: string;
+    /** Middleware identifiers observed on the route chain, in source order. */
+    readonly middleware?: readonly string[];
+    /** Response payload keys scraped from the handler body (best effort). */
+    readonly responseKeys?: readonly string[];
+    /** Framework the detector is claiming. */
+    readonly framework: "nextjs" | "express" | "fastapi" | "spring" | "nestjs" | "rails" | "unknown";
+}
+/** A single tool definition (MCP / JSON-RPC) discovered by heuristic scan. */
+export interface ExtractedTool {
+    /** Name string from the tool definition literal. */
+    readonly toolName: string;
+    /** Repo-relative path to the file that declared the tool. */
+    readonly handlerFile: string;
+    /** Description string from the tool definition literal, if present. */
+    readonly description?: string;
+    /**
+     * Canonical (key-sorted) JSON-encoded `inputSchema` literal harvested
+     * from the tool definition object, when present. Absent when the
+     * detector can't find a parseable literal nearby.
+     */
+    readonly inputSchemaJson?: string;
+}
+/** A file -> ORM model call edge discovered by regex scan. */
+export interface ExtractedOrmEdge {
+    /** Repo-relative path to the caller. */
+    readonly callerFile: string;
+    /** Model / table identifier as written in source. */
+    readonly modelName: string;
+    /** Operation identifier (e.g. "findMany", "create", "select"). */
+    readonly operation: string;
+    /** ORM family this edge belongs to. */
+    readonly orm: "prisma" | "supabase";
+    /** Detector confidence in [0, 1]. */
+    readonly confidence: number;
+    /**
+     * Provenance tag: `"receiver-confirmed"` when the receiver type was
+     * confirmed via import graph or ts-morph, `"heuristic"` when the emit
+     * was based on regex alone (legacy behavior, used only when
+     * `strictDetectors` is false and no import map is available).
+     */
+    readonly reason?: "receiver-confirmed" | "heuristic";
+}
+/**
+ * Input bundle passed to file-scoped detectors. The `captures` field is
+ * reserved for a future AST-driven path and is unused at MVP; detectors rely
+ * on the file path and raw UTF-8 content.
+ *
+ * `importsByFile` + `tsMorphProject` are the P06 plumbing used by the
+ * receiver resolver to confirm that a call's receiver truly originates
+ * from the framework the detector claims. Both are optional — when
+ * absent, detectors fall back to the pre-P06 regex-only behavior unless
+ * `strictDetectors` is set.
+ */
+export interface ExtractInput {
+    readonly filePath: string;
+    readonly content: string;
+    readonly captures?: readonly unknown[];
+    /**
+     * Per-file import declarations the parse phase already extracted.
+     * Used by the receiver resolver's fast path.
+     */
+    readonly importsByFile?: ReadonlyMap<string, readonly import("./receiver-resolver.js").ImportedSymbol[]>;
+    /** Optional ts-morph project for the receiver resolver's type-check path. */
+    readonly tsMorphProject?: import("./receiver-resolver.js").TsMorphProject;
+    /**
+     * When `true`, drop heuristic-only matches entirely — detectors emit
+     * only records whose receiver resolved successfully. Exposed through
+     * the `codehub analyze --strict-detectors` flag.
+     */
+    readonly strictDetectors?: boolean;
+}
+//# sourceMappingURL=types.d.ts.map

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

@@ -0,0 +1 @@
+{"version":3,"file":"types.d.ts","sourceRoot":"","sources":["../../src/extract/types.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAEH,+DAA+D;AAC/D,MAAM,WAAW,cAAc;IAC7B,sEAAsE;IACtE,QAAQ,CAAC,GAAG,EAAE,MAAM,CAAC;IACrB,qEAAqE;IACrE,QAAQ,CAAC,MAAM,CAAC,EAAE,MAAM,CAAC;IACzB,8DAA8D;IAC9D,QAAQ,CAAC,WAAW,EAAE,MAAM,CAAC;IAC7B,2EAA2E;IAC3E,QAAQ,CAAC,UAAU,CAAC,EAAE,SAAS,MAAM,EAAE,CAAC;IACxC,yEAAyE;IACzE,QAAQ,CAAC,YAAY,CAAC,EAAE,SAAS,MAAM,EAAE,CAAC;IAC1C,0CAA0C;IAC1C,QAAQ,CAAC,SAAS,EAAE,QAAQ,GAAG,SAAS,GAAG,SAAS,GAAG,QAAQ,GAAG,QAAQ,GAAG,OAAO,GAAG,SAAS,CAAC;CAClG;AAED,8EAA8E;AAC9E,MAAM,WAAW,aAAa;IAC5B,oDAAoD;IACpD,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAC;IAC1B,6DAA6D;IAC7D,QAAQ,CAAC,WAAW,EAAE,MAAM,CAAC;IAC7B,uEAAuE;IACvE,QAAQ,CAAC,WAAW,CAAC,EAAE,MAAM,CAAC;IAC9B;;;;OAIG;IACH,QAAQ,CAAC,eAAe,CAAC,EAAE,MAAM,CAAC;CACnC;AAED,8DAA8D;AAC9D,MAAM,WAAW,gBAAgB;IAC/B,wCAAwC;IACxC,QAAQ,CAAC,UAAU,EAAE,MAAM,CAAC;IAC5B,qDAAqD;IACrD,QAAQ,CAAC,SAAS,EAAE,MAAM,CAAC;IAC3B,kEAAkE;IAClE,QAAQ,CAAC,SAAS,EAAE,MAAM,CAAC;IAC3B,uCAAuC;IACvC,QAAQ,CAAC,GAAG,EAAE,QAAQ,GAAG,UAAU,CAAC;IACpC,qCAAqC;IACrC,QAAQ,CAAC,UAAU,EAAE,MAAM,CAAC;IAC5B;;;;;OAKG;IACH,QAAQ,CAAC,MAAM,CAAC,EAAE,oBAAoB,GAAG,WAAW,CAAC;CACtD;AAED;;;;;;;;;;GAUG;AACH,MAAM,WAAW,YAAY;IAC3B,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAC;IAC1B,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAC;IACzB,QAAQ,CAAC,QAAQ,CAAC,EAAE,SAAS,OAAO,EAAE,CAAC;IACvC;;;OAGG;IACH,QAAQ,CAAC,aAAa,CAAC,EAAE,WAAW,CAClC,MAAM,EACN,SAAS,OAAO,wBAAwB,EAAE,cAAc,EAAE,CAC3D,CAAC;IACF,6EAA6E;IAC7E,QAAQ,CAAC,cAAc,CAAC,EAAE,OAAO,wBAAwB,EAAE,cAAc,CAAC;IAC1E;;;;OAIG;IACH,QAAQ,CAAC,eAAe,CAAC,EAAE,OAAO,CAAC;CACpC"}

package/dist/extract/types.js

@@ -0,0 +1,11 @@
+/**
+ * Shared extraction result shapes for the route / tool / ORM detectors.
+ *
+ * These types are deliberately minimal: they describe the evidence a detector
+ * has discovered in a single file (or, for filesystem-routed frameworks, a
+ * group of files) without taking any dependency on the graph schema. The
+ * routes / tools / orm pipeline phases turn these records into graph nodes
+ * and edges.
+ */
+export {};
+//# sourceMappingURL=types.js.map

package/dist/extract/types.js.map

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

package/dist/index.d.ts

@@ -0,0 +1,10 @@
+/**
+ * @opencodehub/ingestion — indexing pipeline root barrel.
+ *
+ * Exports the `parse`, `providers`, and `pipeline` subsystems. Extending
+ * the default phase set does not change the exported shape.
+ */
+export * as parse from "./parse/index.js";
+export * as pipeline from "./pipeline/index.js";
+export * as providers from "./providers/index.js";
+//# sourceMappingURL=index.d.ts.map

package/dist/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAEH,OAAO,KAAK,KAAK,MAAM,kBAAkB,CAAC;AAC1C,OAAO,KAAK,QAAQ,MAAM,qBAAqB,CAAC;AAChD,OAAO,KAAK,SAAS,MAAM,sBAAsB,CAAC"}

package/dist/index.js

@@ -0,0 +1,10 @@
+/**
+ * @opencodehub/ingestion — indexing pipeline root barrel.
+ *
+ * Exports the `parse`, `providers`, and `pipeline` subsystems. Extending
+ * the default phase set does not change the exported shape.
+ */
+export * as parse from "./parse/index.js";
+export * as pipeline from "./pipeline/index.js";
+export * as providers from "./providers/index.js";
+//# sourceMappingURL=index.js.map

package/dist/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAEH,OAAO,KAAK,KAAK,MAAM,kBAAkB,CAAC;AAC1C,OAAO,KAAK,QAAQ,MAAM,qBAAqB,CAAC;AAChD,OAAO,KAAK,SAAS,MAAM,sBAAsB,CAAC"}

package/dist/parse/cobol-regex.d.ts

@@ -0,0 +1,85 @@
+/**
+ * COBOL regex hot path.
+ *
+ * Pure-function extractor for fixed-format COBOL files (`.cbl`, `.cob`,
+ * `.cpy`). Emits {@link CobolElement} records for the five targets that a
+ * human reader would use to navigate a legacy mainframe program:
+ *
+ *   - `program-id`   — `PROGRAM-ID. <name>.`, one per file
+ *   - `paragraph`    — labels in Area A: `^[ ]{7}[A-Z0-9][A-Z0-9-]*\.`
+ *   - `perform`      — `PERFORM <identifier>`, each occurrence (heuristic
+ *                      CALL-like reference; the enclosing paragraph is the
+ *                      caller)
+ *   - `copy`         — `COPY <name>`, each occurrence (copybook inclusion)
+ *   - `cics`         — `EXEC CICS ... END-EXEC` spans (multi-line aware)
+ *
+ * ## Fixed-format COBOL refresher
+ *
+ *   Columns 1-6   sequence numbers (ignored)
+ *   Column  7     indicator area: `*` or `/` = comment line, `-` =
+ *                 continuation, `D` = debugging aid, ` ` = normal
+ *   Columns 8-11  Area A: divisions, sections, paragraphs
+ *   Columns 12-72 Area B: statements
+ *   Columns 73-80 identification (ignored)
+ *
+ * The default parse path runs at ≤ 1 ms on 1000-line fixtures; a p50
+ * regression in that number is a graph-ingestion regression.
+ *
+ * ## Anti-goals
+ *
+ *   - NOT a full parse: `PERFORM ... THRU ... VARYING`, `COPY ... REPLACING
+ *     ==tag== BY ==value==`, and nested `EXEC SQL` blocks are all resolved
+ *     heuristically. The deep-parse path (ProLeap, when wired in) owns the
+ *     precise AST.
+ *   - NOT free-format aware: the 99% legacy estate is fixed-format;
+ *     free-format COBOL (column-0 start) lands with the ProLeap backend.
+ *   - NO filesystem I/O, NO subprocesses, NO external deps. The function
+ *     is pure over `(path, content)`.
+ *
+ * ## Author's note
+ *
+ * The regex vocabulary here (PROGRAM-ID, PARAGRAPH, PERFORM, COPY, CICS) is
+ * explicitly allow-listed in `scripts/check-banned-strings.sh` (U2 in spec
+ * 004) because it's the standard public COBOL surface.
+ */
+import type { LanguageId } from "./types.js";
+/** Tag for the kind of construct a {@link CobolElement} describes. */
+export type CobolElementKind = "program-id" | "paragraph" | "perform" | "copy" | "cics";
+/**
+ * One element extracted from a COBOL file. The pipeline maps these to
+ * `CodeElement` graph nodes downstream (see `pipeline/phases/parse.ts`).
+ *
+ * Line numbers are 1-indexed. `endLine` equals `startLine` for the
+ * single-line PROGRAM-ID, paragraph, PERFORM, and COPY markers; CICS
+ * spans cover the `EXEC CICS` → `END-EXEC` range.
+ */
+export interface CobolElement {
+    readonly kind: CobolElementKind;
+    /** Program name, paragraph label, target identifier, or copybook name. */
+    readonly name: string;
+    readonly filePath: string;
+    readonly startLine: number;
+    readonly endLine: number;
+    readonly language: LanguageId;
+    /** Regex extraction is not a parse; the confidence tier says so. */
+    readonly confidence: "heuristic";
+    /**
+     * Optional human-readable snippet — the matched line (or first line of a
+     * multi-line CICS block), whitespace-trimmed. Kept short so graph-node
+     * payloads stay deterministic and compact.
+     */
+    readonly snippet?: string;
+}
+export interface CobolRegexResult {
+    readonly elements: readonly CobolElement[];
+    /** Every `COPY <name>` target referenced by this file, deduped + sorted. */
+    readonly copybookRefs: readonly string[];
+    /** Non-fatal notes (e.g. malformed CICS block). Empty on happy path. */
+    readonly diagnostics: readonly string[];
+}
+/**
+ * Parse a COBOL file and return the extracted element set. Pure function;
+ * safe to call from any thread / worker.
+ */
+export declare function parseCobolFile(path: string, content: string): CobolRegexResult;
+//# sourceMappingURL=cobol-regex.d.ts.map

package/dist/parse/cobol-regex.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"cobol-regex.d.ts","sourceRoot":"","sources":["../../src/parse/cobol-regex.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2CG;AAEH,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,YAAY,CAAC;AAE7C,sEAAsE;AACtE,MAAM,MAAM,gBAAgB,GAAG,YAAY,GAAG,WAAW,GAAG,SAAS,GAAG,MAAM,GAAG,MAAM,CAAC;AAExF;;;;;;;GAOG;AACH,MAAM,WAAW,YAAY;IAC3B,QAAQ,CAAC,IAAI,EAAE,gBAAgB,CAAC;IAChC,0EAA0E;IAC1E,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAC;IAC1B,QAAQ,CAAC,SAAS,EAAE,MAAM,CAAC;IAC3B,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAC;IACzB,QAAQ,CAAC,QAAQ,EAAE,UAAU,CAAC;IAC9B,oEAAoE;IACpE,QAAQ,CAAC,UAAU,EAAE,WAAW,CAAC;IACjC;;;;OAIG;IACH,QAAQ,CAAC,OAAO,CAAC,EAAE,MAAM,CAAC;CAC3B;AAED,MAAM,WAAW,gBAAgB;IAC/B,QAAQ,CAAC,QAAQ,EAAE,SAAS,YAAY,EAAE,CAAC;IAC3C,4EAA4E;IAC5E,QAAQ,CAAC,YAAY,EAAE,SAAS,MAAM,EAAE,CAAC;IACzC,wEAAwE;IACxE,QAAQ,CAAC,WAAW,EAAE,SAAS,MAAM,EAAE,CAAC;CACzC;AAgED;;;GAGG;AACH,wBAAgB,cAAc,CAAC,IAAI,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,GAAG,gBAAgB,CA8K9E"}