narai-primitives 2.0.0-rc.1

package/dist/toolkit/extract_binary.js

@@ -0,0 +1,342 @@
+#!/usr/bin/env node
+/**
+ * Binary file extraction: PDF, DOCX, PPTX to markdown.
+ *
+ * TypeScript port of the doc-wiki extract_binary.ts, with additions:
+ *   - `maxBytes` size cap (default 50 MiB) enforced before read, raising
+ *     `ExtractCapExceeded` so callers can surface the cap in an envelope.
+ *   - `extract()` returns a structured `ExtractResult` instead of a bare
+ *     string, carrying `{ format, text, sizeBytes }`. The per-format
+ *     helpers (`extractPdf` / `extractDocx` / `extractPptx`) continue to
+ *     return `string`; the wrapping happens in `extract()`.
+ *   - CLI emits JSON instead of plain text so it stays pipe-parseable.
+ *
+ * Usage as a library:
+ *     import { extract, extractPdf } from "./extract_binary.js";
+ *     const { text, format, sizeBytes } = await extract("report.pdf");
+ *
+ * Usage as a script:
+ *     node extract_binary.js report.pdf
+ *     node extract_binary.js slides.pptx --format pptx
+ */
+import * as fs from "node:fs";
+import * as path from "node:path";
+import { fileURLToPath } from "node:url";
+import { importOptional } from "./_optional.js";
+// ── Format registry ────────────────────────────────────────────────
+export const FORMAT_MAP = {
+    ".pdf": "pdf",
+    ".docx": "docx",
+    ".pptx": "pptx",
+};
+export { importOptional };
+// ── Size cap ───────────────────────────────────────────────────────
+const DEFAULT_MAX_BYTES = 50 * 1024 * 1024;
+export class ExtractCapExceeded extends Error {
+    sizeBytes;
+    maxBytes;
+    constructor(sizeBytes, maxBytes) {
+        super(`File size ${sizeBytes} exceeds cap ${maxBytes}`);
+        this.sizeBytes = sizeBytes;
+        this.maxBytes = maxBytes;
+        this.name = "ExtractCapExceeded";
+    }
+}
+// ── Whitespace normalization ───────────────────────────────────────
+/**
+ * Normalize extracted plain text so PDF / DOCX / PPTX outputs share the
+ * same shape as the Python pipeline:
+ *   - strip per-line trailing whitespace
+ *   - collapse 3+ consecutive newlines down to exactly 2
+ *   - trim leading/trailing whitespace overall
+ */
+export function normalizeExtracted(text) {
+    const stripped = text
+        .split("\n")
+        .map((line) => line.replace(/[ \t\r\f\v]+$/, ""))
+        .join("\n");
+    const collapsed = stripped.replace(/\n{3,}/g, "\n\n");
+    return collapsed.trim();
+}
+/**
+ * Extract text from a PDF file using pdfjs-dist's legacy build (Node-safe).
+ * Per-format helpers return raw strings; `extract()` wraps into ExtractResult.
+ */
+export async function extractPdf(filePath) {
+    const pdfjs = await importOptional("pdfjs-dist/legacy/build/pdf.mjs");
+    const data = fs.readFileSync(filePath);
+    const bytes = new Uint8Array(data.byteLength);
+    bytes.set(data);
+    const doc = await pdfjs.getDocument({ data: bytes, verbosity: 0 }).promise;
+    try {
+        const pages = [];
+        for (let i = 1; i <= doc.numPages; i++) {
+            const page = await doc.getPage(i);
+            const tc = await page.getTextContent();
+            const parts = [];
+            for (const item of tc.items) {
+                if ("str" in item) {
+                    parts.push(item.str);
+                    if (item.hasEOL)
+                        parts.push("\n");
+                }
+            }
+            const joined = parts.join("").trim();
+            if (joined.length > 0)
+                pages.push(joined);
+        }
+        return normalizeExtracted(pages.join("\n\n"));
+    }
+    finally {
+        await doc.destroy();
+    }
+}
+/** Extract text from a DOCX file using mammoth's undocumented `convertToMarkdown`. */
+export async function extractDocx(filePath) {
+    const mod = await importOptional("mammoth");
+    const convertToMarkdown = mod.convertToMarkdown ?? mod.default?.convertToMarkdown;
+    if (typeof convertToMarkdown !== "function") {
+        throw new Error("Missing optional dependency 'mammoth'. Install with: npm install mammoth");
+    }
+    const buffer = fs.readFileSync(filePath);
+    const result = await convertToMarkdown({ buffer });
+    const unescaped = result.value
+        .replace(/\\([`*_{}\[\]()#+\-.!])/g, "$1")
+        .replace(/\\\\/g, "\\");
+    return normalizeExtracted(unescaped);
+}
+function collectPptxParagraphs(node) {
+    const paragraphs = [];
+    function visit(n) {
+        if (n === null || n === undefined)
+            return;
+        if (Array.isArray(n)) {
+            for (const item of n)
+                visit(item);
+            return;
+        }
+        if (typeof n !== "object")
+            return;
+        const obj = n;
+        for (const [key, value] of Object.entries(obj)) {
+            if (key === "a:p") {
+                const paras = Array.isArray(value) ? value : [value];
+                for (const p of paras) {
+                    paragraphs.push(collectRuns(p));
+                }
+            }
+            else {
+                visit(value);
+            }
+        }
+    }
+    function collectRuns(p) {
+        if (p === null || p === undefined)
+            return "";
+        if (typeof p !== "object")
+            return "";
+        const out = [];
+        function walkRuns(n) {
+            if (n === null || n === undefined)
+                return;
+            if (Array.isArray(n)) {
+                for (const item of n)
+                    walkRuns(item);
+                return;
+            }
+            if (typeof n !== "object")
+                return;
+            const obj = n;
+            for (const [k, v] of Object.entries(obj)) {
+                if (k === "a:t") {
+                    const items = Array.isArray(v) ? v : [v];
+                    for (const item of items) {
+                        if (typeof item === "string") {
+                            out.push(item);
+                        }
+                        else if (item !== null && typeof item === "object") {
+                            const txt = item["#text"];
+                            if (typeof txt === "string")
+                                out.push(txt);
+                            else if (typeof txt === "number")
+                                out.push(String(txt));
+                        }
+                        else if (typeof item === "number") {
+                            out.push(String(item));
+                        }
+                    }
+                }
+                else if (k !== "#text" && k !== ":@") {
+                    walkRuns(v);
+                }
+            }
+        }
+        walkRuns(p);
+        return out.join("");
+    }
+    visit(node);
+    return paragraphs;
+}
+/** Extract text from a PPTX file using JSZip + fast-xml-parser. */
+export async function extractPptx(filePath) {
+    const jszipMod = await importOptional("jszip");
+    const xmlMod = await importOptional("fast-xml-parser");
+    const loadAsync = jszipMod.loadAsync ?? jszipMod.default?.loadAsync;
+    if (typeof loadAsync !== "function") {
+        throw new Error("Missing optional dependency 'jszip'. Install with: npm install jszip");
+    }
+    const buf = fs.readFileSync(filePath);
+    const zip = await loadAsync(buf);
+    const slidePaths = Object.keys(zip.files)
+        .filter((p) => /^ppt\/slides\/slide\d+\.xml$/.test(p))
+        .sort((a, b) => slideIndex(a) - slideIndex(b));
+    const parser = new xmlMod.XMLParser({
+        ignoreAttributes: true,
+        textNodeName: "#text",
+        parseTagValue: false,
+        trimValues: false,
+    });
+    const sections = [];
+    for (let i = 0; i < slidePaths.length; i++) {
+        const slidePath = slidePaths[i];
+        if (slidePath === undefined)
+            continue;
+        const entry = zip.file(slidePath);
+        if (entry === null)
+            continue;
+        const xml = await entry.async("string");
+        const tree = parser.parse(xml);
+        const paragraphs = collectPptxParagraphs(tree)
+            .map((p) => p.trim())
+            .filter((p) => p.length > 0);
+        const slideText = paragraphs.join("\n\n");
+        sections.push(`## Slide ${i + 1}\n\n${slideText}`);
+    }
+    return normalizeExtracted(sections.join("\n\n"));
+}
+function slideIndex(p) {
+    const m = /slide(\d+)\.xml$/.exec(p);
+    return m && m[1] ? parseInt(m[1], 10) : 0;
+}
+// ── Dispatch ───────────────────────────────────────────────────────
+/**
+ * Extract text from a binary file to markdown, returning `ExtractResult`.
+ *
+ * Enforces a size cap (default 50 MiB) before reading the file; throws
+ * `ExtractCapExceeded` when exceeded. If `fmt` is not given, auto-detects
+ * from the file extension.
+ */
+export async function extract(inputPath, fmt = null, opts) {
+    let format = fmt === null || fmt === undefined ? undefined : fmt;
+    if (format === undefined) {
+        const ext = path.extname(inputPath).toLowerCase();
+        format = FORMAT_MAP[ext];
+        if (format === undefined) {
+            const supported = Object.keys(FORMAT_MAP).sort().join(", ");
+            throw new Error(`Unsupported format for '${path.basename(inputPath)}'. ` +
+                `Supported extensions: ${supported}`);
+        }
+    }
+    const stat = fs.statSync(inputPath);
+    const cap = opts?.maxBytes ?? DEFAULT_MAX_BYTES;
+    if (stat.size > cap) {
+        throw new ExtractCapExceeded(stat.size, cap);
+    }
+    let text;
+    if (format === "pdf")
+        text = await extractPdf(inputPath);
+    else if (format === "docx")
+        text = await extractDocx(inputPath);
+    else if (format === "pptx")
+        text = await extractPptx(inputPath);
+    else
+        throw new Error(`Unsupported format: ${format}`);
+    return { format, text, sizeBytes: stat.size };
+}
+// ── CLI ────────────────────────────────────────────────────────────
+const HELP_TEXT = `usage: extract_binary.js [-h] [--format {pdf,docx,pptx}] input_file
+
+Extract text from binary files (PDF, DOCX, PPTX). Emits JSON on stdout
+containing { format, text, sizeBytes }.
+
+positional arguments:
+  input_file            Path to the binary file
+
+options:
+  -h, --help            show this help message and exit
+  --format {pdf,docx,pptx}
+                        Force format (auto-detected if omitted)
+`;
+function parseCli(argv) {
+    const out = {};
+    let i = 0;
+    while (i < argv.length) {
+        const a = argv[i];
+        if (a === undefined) {
+            i++;
+            continue;
+        }
+        if (a === "-h" || a === "--help") {
+            out.help = true;
+            i++;
+            continue;
+        }
+        if (a === "--format") {
+            const v = argv[i + 1];
+            if (v === undefined) {
+                throw new Error("argument --format: expected one argument");
+            }
+            out.format = coerceFormat(v);
+            i += 2;
+            continue;
+        }
+        if (a.startsWith("--format=")) {
+            out.format = coerceFormat(a.slice("--format=".length));
+            i++;
+            continue;
+        }
+        if (a.startsWith("-")) {
+            throw new Error(`unrecognized arguments: ${a}`);
+        }
+        if (out.inputFile !== undefined) {
+            throw new Error(`unrecognized arguments: ${a}`);
+        }
+        out.inputFile = a;
+        i++;
+    }
+    return out;
+}
+function coerceFormat(v) {
+    if (v === "pdf" || v === "docx" || v === "pptx")
+        return v;
+    throw new Error(`argument --format: invalid choice: '${v}' (choose from 'pdf', 'docx', 'pptx')`);
+}
+export async function main(argv = process.argv.slice(2)) {
+    let args;
+    try {
+        args = parseCli(argv);
+    }
+    catch (e) {
+        process.stderr.write(`${e.message}\n`);
+        return 2;
+    }
+    if (args.help) {
+        process.stdout.write(HELP_TEXT);
+        return 0;
+    }
+    if (args.inputFile === undefined) {
+        process.stderr.write("the following arguments are required: input_file\n");
+        return 2;
+    }
+    const result = await extract(args.inputFile, args.format ?? null);
+    process.stdout.write(JSON.stringify(result) + "\n");
+    return 0;
+}
+const thisFile = fileURLToPath(import.meta.url);
+if (process.argv[1] && path.resolve(process.argv[1]) === thisFile) {
+    main().then((code) => process.exit(code), (e) => {
+        process.stderr.write(`${e.message}\n`);
+        process.exit(1);
+    });
+}
+//# sourceMappingURL=extract_binary.js.map

package/dist/toolkit/extract_binary.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"extract_binary.js","sourceRoot":"","sources":["../../src/toolkit/extract_binary.ts"],"names":[],"mappings":";AACA;;;;;;;;;;;;;;;;;;;GAmBG;AACH,OAAO,KAAK,EAAE,MAAM,SAAS,CAAC;AAC9B,OAAO,KAAK,IAAI,MAAM,WAAW,CAAC;AAClC,OAAO,EAAE,aAAa,EAAE,MAAM,UAAU,CAAC;AACzC,OAAO,EAAE,cAAc,EAAE,MAAM,gBAAgB,CAAC;AAEhD,sEAAsE;AAEtE,MAAM,CAAC,MAAM,UAAU,GAAsD;IAC3E,MAAM,EAAE,KAAK;IACb,OAAO,EAAE,MAAM;IACf,OAAO,EAAE,MAAM;CAChB,CAAC;AAIF,OAAO,EAAE,cAAc,EAAE,CAAC;AAE1B,sEAAsE;AAEtE,MAAM,iBAAiB,GAAG,EAAE,GAAG,IAAI,GAAG,IAAI,CAAC;AAY3C,MAAM,OAAO,kBAAmB,SAAQ,KAAK;IAEzB;IACA;IAFlB,YACkB,SAAiB,EACjB,QAAgB;QAEhC,KAAK,CAAC,aAAa,SAAS,gBAAgB,QAAQ,EAAE,CAAC,CAAC;QAHxC,cAAS,GAAT,SAAS,CAAQ;QACjB,aAAQ,GAAR,QAAQ,CAAQ;QAGhC,IAAI,CAAC,IAAI,GAAG,oBAAoB,CAAC;IACnC,CAAC;CACF;AAED,sEAAsE;AAEtE;;;;;;GAMG;AACH,MAAM,UAAU,kBAAkB,CAAC,IAAY;IAC7C,MAAM,QAAQ,GAAG,IAAI;SAClB,KAAK,CAAC,IAAI,CAAC;SACX,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,OAAO,CAAC,eAAe,EAAE,EAAE,CAAC,CAAC;SAChD,IAAI,CAAC,IAAI,CAAC,CAAC;IACd,MAAM,SAAS,GAAG,QAAQ,CAAC,OAAO,CAAC,SAAS,EAAE,MAAM,CAAC,CAAC;IACtD,OAAO,SAAS,CAAC,IAAI,EAAE,CAAC;AAC1B,CAAC;AA8BD;;;GAGG;AACH,MAAM,CAAC,KAAK,UAAU,UAAU,CAAC,QAAgB;IAC/C,MAAM,KAAK,GAAG,MAAM,cAAc,CAChC,iCAAiC,CAClC,CAAC;IACF,MAAM,IAAI,GAAG,EAAE,CAAC,YAAY,CAAC,QAAQ,CAAC,CAAC;IACvC,MAAM,KAAK,GAAG,IAAI,UAAU,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;IAC9C,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;IAChB,MAAM,GAAG,GAAG,MAAM,KAAK,CAAC,WAAW,CAAC,EAAE,IAAI,EAAE,KAAK,EAAE,SAAS,EAAE,CAAC,EAAE,CAAC,CAAC,OAAO,CAAC;IAC3E,IAAI,CAAC;QACH,MAAM,KAAK,GAAa,EAAE,CAAC;QAC3B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,GAAG,CAAC,QAAQ,EAAE,CAAC,EAAE,EAAE,CAAC;YACvC,MAAM,IAAI,GAAG,MAAM,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC;YAClC,MAAM,EAAE,GAAG,MAAM,IAAI,CAAC,cAAc,EAAE,CAAC;YACvC,MAAM,KAAK,GAAa,EAAE,CAAC;YAC3B,KAAK,MAAM,IAAI,IAAI,EAAE,CAAC,KAAK,EAAE,CAAC;gBAC5B,IAAI,KAAK,IAAI,IAAI,EAAE,CAAC;oBAClB,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;oBACrB,IAAI,IAAI,CAAC,MAAM;wBAAE,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;gBACpC,CAAC;YACH,CAAC;YACD,MAAM,MAAM,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;YACrC,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC;gBAAE,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QAC5C,CAAC;QACD,OAAO,kBAAkB,CAAC,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC;IAChD,CAAC;YAAS,CAAC;QACT,MAAM,GAAG,CAAC,OAAO,EAAE,CAAC;IACtB,CAAC;AACH,CAAC;AAeD,sFAAsF;AACtF,MAAM,CAAC,KAAK,UAAU,WAAW,CAAC,QAAgB;IAChD,MAAM,GAAG,GAAG,MAAM,cAAc,CAAgB,SAAS,CAAC,CAAC;IAC3D,MAAM,iBAAiB,GACrB,GAAG,CAAC,iBAAiB,IAAI,GAAG,CAAC,OAAO,EAAE,iBAAiB,CAAC;IAC1D,IAAI,OAAO,iBAAiB,KAAK,UAAU,EAAE,CAAC;QAC5C,MAAM,IAAI,KAAK,CACb,0EAA0E,CAC3E,CAAC;IACJ,CAAC;IACD,MAAM,MAAM,GAAG,EAAE,CAAC,YAAY,CAAC,QAAQ,CAAC,CAAC;IACzC,MAAM,MAAM,GAAG,MAAM,iBAAiB,CAAC,EAAE,MAAM,EAAE,CAAC,CAAC;IACnD,MAAM,SAAS,GAAG,MAAM,CAAC,KAAK;SAC3B,OAAO,CAAC,0BAA0B,EAAE,IAAI,CAAC;SACzC,OAAO,CAAC,OAAO,EAAE,IAAI,CAAC,CAAC;IAC1B,OAAO,kBAAkB,CAAC,SAAS,CAAC,CAAC;AACvC,CAAC;AAsBD,SAAS,qBAAqB,CAAC,IAAa;IAC1C,MAAM,UAAU,GAAa,EAAE,CAAC;IAEhC,SAAS,KAAK,CAAC,CAAU;QACvB,IAAI,CAAC,KAAK,IAAI,IAAI,CAAC,KAAK,SAAS;YAAE,OAAO;QAC1C,IAAI,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC;YACrB,KAAK,MAAM,IAAI,IAAI,CAAC;gBAAE,KAAK,CAAC,IAAI,CAAC,CAAC;YAClC,OAAO;QACT,CAAC;QACD,IAAI,OAAO,CAAC,KAAK,QAAQ;YAAE,OAAO;QAElC,MAAM,GAAG,GAAG,CAA4B,CAAC;QACzC,KAAK,MAAM,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC;YAC/C,IAAI,GAAG,KAAK,KAAK,EAAE,CAAC;gBAClB,MAAM,KAAK,GAAG,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC;gBACrD,KAAK,MAAM,CAAC,IAAI,KAAK,EAAE,CAAC;oBACtB,UAAU,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC,CAAC;gBAClC,CAAC;YACH,CAAC;iBAAM,CAAC;gBACN,KAAK,CAAC,KAAK,CAAC,CAAC;YACf,CAAC;QACH,CAAC;IACH,CAAC;IAED,SAAS,WAAW,CAAC,CAAU;QAC7B,IAAI,CAAC,KAAK,IAAI,IAAI,CAAC,KAAK,SAAS;YAAE,OAAO,EAAE,CAAC;QAC7C,IAAI,OAAO,CAAC,KAAK,QAAQ;YAAE,OAAO,EAAE,CAAC;QACrC,MAAM,GAAG,GAAa,EAAE,CAAC;QAEzB,SAAS,QAAQ,CAAC,CAAU;YAC1B,IAAI,CAAC,KAAK,IAAI,IAAI,CAAC,KAAK,SAAS;gBAAE,OAAO;YAC1C,IAAI,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC;gBACrB,KAAK,MAAM,IAAI,IAAI,CAAC;oBAAE,QAAQ,CAAC,IAAI,CAAC,CAAC;gBACrC,OAAO;YACT,CAAC;YACD,IAAI,OAAO,CAAC,KAAK,QAAQ;gBAAE,OAAO;YAClC,MAAM,GAAG,GAAG,CAA4B,CAAC;YACzC,KAAK,MAAM,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC;gBACzC,IAAI,CAAC,KAAK,KAAK,EAAE,CAAC;oBAChB,MAAM,KAAK,GAAG,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;oBACzC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;wBACzB,IAAI,OAAO,IAAI,KAAK,QAAQ,EAAE,CAAC;4BAC7B,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;wBACjB,CAAC;6BAAM,IAAI,IAAI,KAAK,IAAI,IAAI,OAAO,IAAI,KAAK,QAAQ,EAAE,CAAC;4BACrD,MAAM,GAAG,GAAI,IAAgC,CAAC,OAAO,CAAC,CAAC;4BACvD,IAAI,OAAO,GAAG,KAAK,QAAQ;gCAAE,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;iCACtC,IAAI,OAAO,GAAG,KAAK,QAAQ;gCAAE,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC;wBAC1D,CAAC;6BAAM,IAAI,OAAO,IAAI,KAAK,QAAQ,EAAE,CAAC;4BACpC,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC;wBACzB,CAAC;oBACH,CAAC;gBACH,CAAC;qBAAM,IAAI,CAAC,KAAK,OAAO,IAAI,CAAC,KAAK,IAAI,EAAE,CAAC;oBACvC,QAAQ,CAAC,CAAC,CAAC,CAAC;gBACd,CAAC;YACH,CAAC;QACH,CAAC;QAED,QAAQ,CAAC,CAAC,CAAC,CAAC;QACZ,OAAO,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IACtB,CAAC;IAED,KAAK,CAAC,IAAI,CAAC,CAAC;IACZ,OAAO,UAAU,CAAC;AACpB,CAAC;AAED,mEAAmE;AACnE,MAAM,CAAC,KAAK,UAAU,WAAW,CAAC,QAAgB;IAChD,MAAM,QAAQ,GAAG,MAAM,cAAc,CAAc,OAAO,CAAC,CAAC;IAC5D,MAAM,MAAM,GAAG,MAAM,cAAc,CAAkB,iBAAiB,CAAC,CAAC;IAExE,MAAM,SAAS,GAAG,QAAQ,CAAC,SAAS,IAAI,QAAQ,CAAC,OAAO,EAAE,SAAS,CAAC;IACpE,IAAI,OAAO,SAAS,KAAK,UAAU,EAAE,CAAC;QACpC,MAAM,IAAI,KAAK,CACb,sEAAsE,CACvE,CAAC;IACJ,CAAC;IAED,MAAM,GAAG,GAAG,EAAE,CAAC,YAAY,CAAC,QAAQ,CAAC,CAAC;IACtC,MAAM,GAAG,GAAG,MAAM,SAAS,CAAC,GAAG,CAAC,CAAC;IAEjC,MAAM,UAAU,GAAG,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC;SACtC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,8BAA8B,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;SACrD,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,UAAU,CAAC,CAAC,CAAC,GAAG,UAAU,CAAC,CAAC,CAAC,CAAC,CAAC;IAEjD,MAAM,MAAM,GAAG,IAAI,MAAM,CAAC,SAAS,CAAC;QAClC,gBAAgB,EAAE,IAAI;QACtB,YAAY,EAAE,OAAO;QACrB,aAAa,EAAE,KAAK;QACpB,UAAU,EAAE,KAAK;KAClB,CAAC,CAAC;IAEH,MAAM,QAAQ,GAAa,EAAE,CAAC;IAC9B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,UAAU,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAC3C,MAAM,SAAS,GAAG,UAAU,CAAC,CAAC,CAAC,CAAC;QAChC,IAAI,SAAS,KAAK,SAAS;YAAE,SAAS;QACtC,MAAM,KAAK,GAAG,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;QAClC,IAAI,KAAK,KAAK,IAAI;YAAE,SAAS;QAC7B,MAAM,GAAG,GAAG,MAAM,KAAK,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC;QACxC,MAAM,IAAI,GAAG,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;QAC/B,MAAM,UAAU,GAAG,qBAAqB,CAAC,IAAI,CAAC;aAC3C,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;aACpB,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;QAC/B,MAAM,SAAS,GAAG,UAAU,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QAC1C,QAAQ,CAAC,IAAI,CAAC,YAAY,CAAC,GAAG,CAAC,OAAO,SAAS,EAAE,CAAC,CAAC;IACrD,CAAC;IACD,OAAO,kBAAkB,CAAC,QAAQ,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC;AACnD,CAAC;AAED,SAAS,UAAU,CAAC,CAAS;IAC3B,MAAM,CAAC,GAAG,kBAAkB,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IACrC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;AAC5C,CAAC;AAED,sEAAsE;AAEtE;;;;;;GAMG;AACH,MAAM,CAAC,KAAK,UAAU,OAAO,CAC3B,SAAiB,EACjB,MAA2B,IAAI,EAC/B,IAAqB;IAErB,IAAI,MAAM,GACR,GAAG,KAAK,IAAI,IAAI,GAAG,KAAK,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,GAAG,CAAC;IACtD,IAAI,MAAM,KAAK,SAAS,EAAE,CAAC;QACzB,MAAM,GAAG,GAAG,IAAI,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,WAAW,EAAE,CAAC;QAClD,MAAM,GAAG,UAAU,CAAC,GAAG,CAAC,CAAC;QACzB,IAAI,MAAM,KAAK,SAAS,EAAE,CAAC;YACzB,MAAM,SAAS,GAAG,MAAM,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC,IAAI,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YAC5D,MAAM,IAAI,KAAK,CACb,2BAA2B,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC,KAAK;gBACtD,yBAAyB,SAAS,EAAE,CACvC,CAAC;QACJ,CAAC;IACH,CAAC;IAED,MAAM,IAAI,GAAG,EAAE,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC;IACpC,MAAM,GAAG,GAAG,IAAI,EAAE,QAAQ,IAAI,iBAAiB,CAAC;IAChD,IAAI,IAAI,CAAC,IAAI,GAAG,GAAG,EAAE,CAAC;QACpB,MAAM,IAAI,kBAAkB,CAAC,IAAI,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC;IAC/C,CAAC;IAED,IAAI,IAAY,CAAC;IACjB,IAAI,MAAM,KAAK,KAAK;QAAE,IAAI,GAAG,MAAM,UAAU,CAAC,SAAS,CAAC,CAAC;SACpD,IAAI,MAAM,KAAK,MAAM;QAAE,IAAI,GAAG,MAAM,WAAW,CAAC,SAAS,CAAC,CAAC;SAC3D,IAAI,MAAM,KAAK,MAAM;QAAE,IAAI,GAAG,MAAM,WAAW,CAAC,SAAS,CAAC,CAAC;;QAC3D,MAAM,IAAI,KAAK,CAAC,uBAAuB,MAAM,EAAE,CAAC,CAAC;IAEtD,OAAO,EAAE,MAAM,EAAE,IAAI,EAAE,SAAS,EAAE,IAAI,CAAC,IAAI,EAAE,CAAC;AAChD,CAAC;AAED,sEAAsE;AAEtE,MAAM,SAAS,GAAG;;;;;;;;;;;;CAYjB,CAAC;AAQF,SAAS,QAAQ,CAAC,IAAuB;IACvC,MAAM,GAAG,GAAY,EAAE,CAAC;IACxB,IAAI,CAAC,GAAG,CAAC,CAAC;IACV,OAAO,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC;QACvB,MAAM,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC;QAClB,IAAI,CAAC,KAAK,SAAS,EAAE,CAAC;YACpB,CAAC,EAAE,CAAC;YACJ,SAAS;QACX,CAAC;QACD,IAAI,CAAC,KAAK,IAAI,IAAI,CAAC,KAAK,QAAQ,EAAE,CAAC;YACjC,GAAG,CAAC,IAAI,GAAG,IAAI,CAAC;YAChB,CAAC,EAAE,CAAC;YACJ,SAAS;QACX,CAAC;QACD,IAAI,CAAC,KAAK,UAAU,EAAE,CAAC;YACrB,MAAM,CAAC,GAAG,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;YACtB,IAAI,CAAC,KAAK,SAAS,EAAE,CAAC;gBACpB,MAAM,IAAI,KAAK,CAAC,0CAA0C,CAAC,CAAC;YAC9D,CAAC;YACD,GAAG,CAAC,MAAM,GAAG,YAAY,CAAC,CAAC,CAAC,CAAC;YAC7B,CAAC,IAAI,CAAC,CAAC;YACP,SAAS;QACX,CAAC;QACD,IAAI,CAAC,CAAC,UAAU,CAAC,WAAW,CAAC,EAAE,CAAC;YAC9B,GAAG,CAAC,MAAM,GAAG,YAAY,CAAC,CAAC,CAAC,KAAK,CAAC,WAAW,CAAC,MAAM,CAAC,CAAC,CAAC;YACvD,CAAC,EAAE,CAAC;YACJ,SAAS;QACX,CAAC;QACD,IAAI,CAAC,CAAC,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;YACtB,MAAM,IAAI,KAAK,CAAC,2BAA2B,CAAC,EAAE,CAAC,CAAC;QAClD,CAAC;QACD,IAAI,GAAG,CAAC,SAAS,KAAK,SAAS,EAAE,CAAC;YAChC,MAAM,IAAI,KAAK,CAAC,2BAA2B,CAAC,EAAE,CAAC,CAAC;QAClD,CAAC;QACD,GAAG,CAAC,SAAS,GAAG,CAAC,CAAC;QAClB,CAAC,EAAE,CAAC;IACN,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAED,SAAS,YAAY,CAAC,CAAS;IAC7B,IAAI,CAAC,KAAK,KAAK,IAAI,CAAC,KAAK,MAAM,IAAI,CAAC,KAAK,MAAM;QAAE,OAAO,CAAC,CAAC;IAC1D,MAAM,IAAI,KAAK,CACb,uCAAuC,CAAC,uCAAuC,CAChF,CAAC;AACJ,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,IAAI,CACxB,OAA0B,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC;IAE/C,IAAI,IAAa,CAAC;IAClB,IAAI,CAAC;QACH,IAAI,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC;IACxB,CAAC;IAAC,OAAO,CAAC,EAAE,CAAC;QACX,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,GAAI,CAAW,CAAC,OAAO,IAAI,CAAC,CAAC;QAClD,OAAO,CAAC,CAAC;IACX,CAAC;IAED,IAAI,IAAI,CAAC,IAAI,EAAE,CAAC;QACd,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC;QAChC,OAAO,CAAC,CAAC;IACX,CAAC;IACD,IAAI,IAAI,CAAC,SAAS,KAAK,SAAS,EAAE,CAAC;QACjC,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,oDAAoD,CAAC,CAAC;QAC3E,OAAO,CAAC,CAAC;IACX,CAAC;IAED,MAAM,MAAM,GAAG,MAAM,OAAO,CAAC,IAAI,CAAC,SAAS,EAAE,IAAI,CAAC,MAAM,IAAI,IAAI,CAAC,CAAC;IAClE,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC,GAAG,IAAI,CAAC,CAAC;IACpD,OAAO,CAAC,CAAC;AACX,CAAC;AAED,MAAM,QAAQ,GAAG,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;AAChD,IAAI,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,IAAI,CAAC,OAAO,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,KAAK,QAAQ,EAAE,CAAC;IAClE,IAAI,EAAE,CAAC,IAAI,CACT,CAAC,IAAI,EAAE,EAAE,CAAC,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,EAC5B,CAAC,CAAC,EAAE,EAAE;QACJ,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,GAAI,CAAW,CAAC,OAAO,IAAI,CAAC,CAAC;QAClD,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAClB,CAAC,CACF,CAAC;AACJ,CAAC"}

package/dist/toolkit/extract_multimodal.d.ts

@@ -0,0 +1,22 @@
+#!/usr/bin/env node
+export type MultimodalMode = "on" | "optional" | "off";
+export type MultimodalFormat = "vision" | "audio_video" | "youtube" | "skipped";
+export interface MultimodalResult {
+    format: MultimodalFormat;
+    warning?: string;
+    handoff?: string;
+    input?: string;
+}
+export declare const VISION_EXTENSIONS: ReadonlySet<string>;
+export declare const AUDIO_VIDEO_EXTENSIONS: ReadonlySet<string>;
+export interface MultimodalConfig {
+    enabled?: unknown;
+}
+/**
+ * Route an input (file path or URL) to the right multimodal handler.
+ * Never throws. `format === "skipped"` + populated `warning` is the
+ * universal "nothing happened" signal.
+ */
+export declare function dispatchMultimodal(input: string, cfg?: MultimodalConfig): MultimodalResult;
+export declare function main(argv?: readonly string[]): number;
+//# sourceMappingURL=extract_multimodal.d.ts.map

package/dist/toolkit/extract_multimodal.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"extract_multimodal.d.ts","sourceRoot":"","sources":["../../src/toolkit/extract_multimodal.ts"],"names":[],"mappings":";AA+BA,MAAM,MAAM,cAAc,GAAG,IAAI,GAAG,UAAU,GAAG,KAAK,CAAC;AAUvD,MAAM,MAAM,gBAAgB,GAAG,QAAQ,GAAG,aAAa,GAAG,SAAS,GAAG,SAAS,CAAC;AAEhF,MAAM,WAAW,gBAAgB;IAC/B,MAAM,EAAE,gBAAgB,CAAC;IACzB,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,KAAK,CAAC,EAAE,MAAM,CAAC;CAChB;AAID,eAAO,MAAM,iBAAiB,EAAE,WAAW,CAAC,MAAM,CAOhD,CAAC;AAEH,eAAO,MAAM,sBAAsB,EAAE,WAAW,CAAC,MAAM,CAWrD,CAAC;AAuCH,MAAM,WAAW,gBAAgB;IAC/B,OAAO,CAAC,EAAE,OAAO,CAAC;CACnB;AAED;;;;GAIG;AACH,wBAAgB,kBAAkB,CAChC,KAAK,EAAE,MAAM,EACb,GAAG,GAAE,gBAAqB,GACzB,gBAAgB,CA+ClB;AA4DD,wBAAgB,IAAI,CAClB,IAAI,GAAE,SAAS,MAAM,EAA0B,GAC9C,MAAM,CA8BR"}

package/dist/toolkit/extract_multimodal.js

@@ -0,0 +1,209 @@
+#!/usr/bin/env node
+/**
+ * extract_multimodal.ts — optional extraction path for images, audio,
+ * video, and YouTube URLs.
+ *
+ * Unlike `extract_binary.ts` which uses npm optional deps (`pdfjs-dist`,
+ * `mammoth`, `jszip`), this module treats its external tools as
+ * optional-on-PATH:
+ *
+ *   - **Vision** (`.png .jpg .jpeg .webp .gif .svg`): no tool. Claude
+ *     Code's orchestrator reads the image directly via `Read` (native
+ *     multimodal). This module just flags the handoff.
+ *   - **Audio/video** (`.mp4 .mov .mkv .webm .avi .m4v .mp3 .wav .m4a
+ *     .ogg`): needs `faster-whisper` on `PATH`. Missing → skip with
+ *     install-hint warning.
+ *   - **YouTube / remote AV** (`youtu.be/*`, `youtube.com/watch?v=*`):
+ *     needs `yt-dlp` + `faster-whisper`. Either missing → skip.
+ *
+ * Gated by `multimodal.enabled` config: `"off"` (always skip),
+ * `"optional"` (probe PATH), `"on"` (same as optional but the caller
+ * asserts tools should be present; still skips rather than raises).
+ *
+ * CLI: emits JSON on stdout; the warning, if any, also goes to stderr.
+ * Exits 0 on success/skip, 2 on CLI misuse.
+ */
+import * as path from "node:path";
+import { fileURLToPath } from "node:url";
+import { isBinaryOnPath } from "./_optional.js";
+function normalizeMode(raw) {
+    if (raw === "on" || raw === true)
+        return "on";
+    if (raw === "off" || raw === false)
+        return "off";
+    return "optional";
+}
+// ── Extension classification ────────────────────────────────────────
+export const VISION_EXTENSIONS = new Set([
+    ".png",
+    ".jpg",
+    ".jpeg",
+    ".webp",
+    ".gif",
+    ".svg",
+]);
+export const AUDIO_VIDEO_EXTENSIONS = new Set([
+    ".mp4",
+    ".mov",
+    ".mkv",
+    ".webm",
+    ".avi",
+    ".m4v",
+    ".mp3",
+    ".wav",
+    ".m4a",
+    ".ogg",
+]);
+const YOUTUBE_URL_RE = /^https?:\/\/(?:www\.)?(?:youtu\.be\/|youtube\.com\/watch\?v=)/i;
+function isYoutubeUrl(s) {
+    return YOUTUBE_URL_RE.test(s);
+}
+function extOf(s) {
+    return path.extname(s).toLowerCase();
+}
+// ── Warning strings (standardized — callers may grep for these) ─────
+const WARN_PREFIX = "multimodal:";
+function warnDisabled() {
+    return `${WARN_PREFIX} disabled in config (multimodal.enabled: off). Set to "optional" or "on" to re-enable.`;
+}
+function warnMissingWhisper() {
+    return `${WARN_PREFIX} faster-whisper not on PATH. Install with 'pipx install faster-whisper' (or 'pip install faster-whisper').`;
+}
+function warnMissingYtDlp() {
+    return `${WARN_PREFIX} yt-dlp not on PATH. Install with 'brew install yt-dlp' or 'pipx install yt-dlp', then re-run.`;
+}
+function warnMissingBoth() {
+    return `${WARN_PREFIX} both yt-dlp and faster-whisper missing on PATH. YouTube ingest needs both. Install: 'brew install yt-dlp && pipx install faster-whisper'.`;
+}
+function warnUnknownFormat(input) {
+    return `${WARN_PREFIX} '${input}' does not match any multimodal extension or URL pattern. Supported: ${[...VISION_EXTENSIONS].join(", ")} (vision); ${[...AUDIO_VIDEO_EXTENSIONS].join(", ")} (audio/video); YouTube URLs.`;
+}
+/**
+ * Route an input (file path or URL) to the right multimodal handler.
+ * Never throws. `format === "skipped"` + populated `warning` is the
+ * universal "nothing happened" signal.
+ */
+export function dispatchMultimodal(input, cfg = {}) {
+    const mode = normalizeMode(cfg.enabled);
+    if (mode === "off") {
+        return { format: "skipped", warning: warnDisabled() };
+    }
+    if (isYoutubeUrl(input)) {
+        const hasYtDlp = isBinaryOnPath("yt-dlp");
+        const hasWhisper = isBinaryOnPath("faster-whisper");
+        if (!hasYtDlp && !hasWhisper) {
+            return { format: "skipped", warning: warnMissingBoth() };
+        }
+        if (!hasYtDlp) {
+            return { format: "skipped", warning: warnMissingYtDlp() };
+        }
+        if (!hasWhisper) {
+            return { format: "skipped", warning: warnMissingWhisper() };
+        }
+        return {
+            format: "youtube",
+            handoff: "yt-dlp -x --audio-format wav | faster-whisper",
+            input,
+        };
+    }
+    const ext = extOf(input);
+    if (VISION_EXTENSIONS.has(ext)) {
+        return {
+            format: "vision",
+            handoff: "orchestrator-reads-image",
+            input,
+        };
+    }
+    if (AUDIO_VIDEO_EXTENSIONS.has(ext)) {
+        if (!isBinaryOnPath("faster-whisper")) {
+            return { format: "skipped", warning: warnMissingWhisper() };
+        }
+        return {
+            format: "audio_video",
+            handoff: "faster-whisper",
+            input,
+        };
+    }
+    return { format: "skipped", warning: warnUnknownFormat(input) };
+}
+function parseArgs(argv) {
+    const out = {};
+    let i = 0;
+    while (i < argv.length) {
+        const a = argv[i];
+        if (a === undefined) {
+            i++;
+            continue;
+        }
+        if (a === "-h" || a === "--help") {
+            out.help = true;
+            i++;
+            continue;
+        }
+        if (a === "--enabled") {
+            out.enabled = argv[i + 1] ?? "";
+            i += 2;
+            continue;
+        }
+        if (a.startsWith("--enabled=")) {
+            out.enabled = a.slice("--enabled=".length);
+            i++;
+            continue;
+        }
+        if (a.startsWith("--")) {
+            throw new Error(`unrecognized argument: ${a}`);
+        }
+        if (out.input === undefined) {
+            out.input = a;
+            i++;
+            continue;
+        }
+        throw new Error(`unexpected extra positional: ${a}`);
+    }
+    return out;
+}
+const HELP_TEXT = `usage: extract_multimodal.js <input> [--enabled on|off|optional]
+
+Classify an image / audio / video / YouTube URL and probe for the
+required external tool. Returns JSON describing the dispatch result.
+
+positional arguments:
+  <input>               File path or URL
+
+options:
+  --enabled MODE        Override multimodal.enabled (on|off|optional)
+  -h, --help            Show this help and exit
+`;
+export function main(argv = process.argv.slice(2)) {
+    let args;
+    try {
+        args = parseArgs(argv);
+    }
+    catch (e) {
+        process.stderr.write(`${e.message}\n`);
+        return 2;
+    }
+    if (args.help) {
+        process.stdout.write(HELP_TEXT);
+        return 0;
+    }
+    if (args.input === undefined || args.input === "") {
+        process.stderr.write("required: <input> (file path or URL)\n");
+        return 2;
+    }
+    const cfg = {};
+    if (args.enabled !== undefined && args.enabled !== "") {
+        cfg.enabled = args.enabled;
+    }
+    const result = dispatchMultimodal(args.input, cfg);
+    process.stdout.write(JSON.stringify(result, null, 2) + "\n");
+    if (result.warning !== undefined) {
+        process.stderr.write(`${result.warning}\n`);
+    }
+    return 0;
+}
+const thisFile = fileURLToPath(import.meta.url);
+if (process.argv[1] && path.resolve(process.argv[1]) === thisFile) {
+    process.exit(main());
+}
+//# sourceMappingURL=extract_multimodal.js.map

package/dist/toolkit/extract_multimodal.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"extract_multimodal.js","sourceRoot":"","sources":["../../src/toolkit/extract_multimodal.ts"],"names":[],"mappings":";AACA;;;;;;;;;;;;;;;;;;;;;;;GAuBG;AACH,OAAO,KAAK,IAAI,MAAM,WAAW,CAAC;AAClC,OAAO,EAAE,aAAa,EAAE,MAAM,UAAU,CAAC;AACzC,OAAO,EAAE,cAAc,EAAE,MAAM,gBAAgB,CAAC;AAMhD,SAAS,aAAa,CAAC,GAAY;IACjC,IAAI,GAAG,KAAK,IAAI,IAAI,GAAG,KAAK,IAAI;QAAE,OAAO,IAAI,CAAC;IAC9C,IAAI,GAAG,KAAK,KAAK,IAAI,GAAG,KAAK,KAAK;QAAE,OAAO,KAAK,CAAC;IACjD,OAAO,UAAU,CAAC;AACpB,CAAC;AAaD,uEAAuE;AAEvE,MAAM,CAAC,MAAM,iBAAiB,GAAwB,IAAI,GAAG,CAAC;IAC5D,MAAM;IACN,MAAM;IACN,OAAO;IACP,OAAO;IACP,MAAM;IACN,MAAM;CACP,CAAC,CAAC;AAEH,MAAM,CAAC,MAAM,sBAAsB,GAAwB,IAAI,GAAG,CAAC;IACjE,MAAM;IACN,MAAM;IACN,MAAM;IACN,OAAO;IACP,MAAM;IACN,MAAM;IACN,MAAM;IACN,MAAM;IACN,MAAM;IACN,MAAM;CACP,CAAC,CAAC;AAEH,MAAM,cAAc,GAClB,gEAAgE,CAAC;AAEnE,SAAS,YAAY,CAAC,CAAS;IAC7B,OAAO,cAAc,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;AAChC,CAAC;AAED,SAAS,KAAK,CAAC,CAAS;IACtB,OAAO,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,WAAW,EAAE,CAAC;AACvC,CAAC;AAED,uEAAuE;AAEvE,MAAM,WAAW,GAAG,aAAa,CAAC;AAElC,SAAS,YAAY;IACnB,OAAO,GAAG,WAAW,wFAAwF,CAAC;AAChH,CAAC;AAED,SAAS,kBAAkB;IACzB,OAAO,GAAG,WAAW,4GAA4G,CAAC;AACpI,CAAC;AAED,SAAS,gBAAgB;IACvB,OAAO,GAAG,WAAW,gGAAgG,CAAC;AACxH,CAAC;AAED,SAAS,eAAe;IACtB,OAAO,GAAG,WAAW,4IAA4I,CAAC;AACpK,CAAC;AAED,SAAS,iBAAiB,CAAC,KAAa;IACtC,OAAO,GAAG,WAAW,KAAK,KAAK,wEAAwE,CAAC,GAAG,iBAAiB,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,cAAc,CAAC,GAAG,sBAAsB,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,+BAA+B,CAAC;AAC9N,CAAC;AAQD;;;;GAIG;AACH,MAAM,UAAU,kBAAkB,CAChC,KAAa,EACb,MAAwB,EAAE;IAE1B,MAAM,IAAI,GAAG,aAAa,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;IACxC,IAAI,IAAI,KAAK,KAAK,EAAE,CAAC;QACnB,OAAO,EAAE,MAAM,EAAE,SAAS,EAAE,OAAO,EAAE,YAAY,EAAE,EAAE,CAAC;IACxD,CAAC;IAED,IAAI,YAAY,CAAC,KAAK,CAAC,EAAE,CAAC;QACxB,MAAM,QAAQ,GAAG,cAAc,CAAC,QAAQ,CAAC,CAAC;QAC1C,MAAM,UAAU,GAAG,cAAc,CAAC,gBAAgB,CAAC,CAAC;QACpD,IAAI,CAAC,QAAQ,IAAI,CAAC,UAAU,EAAE,CAAC;YAC7B,OAAO,EAAE,MAAM,EAAE,SAAS,EAAE,OAAO,EAAE,eAAe,EAAE,EAAE,CAAC;QAC3D,CAAC;QACD,IAAI,CAAC,QAAQ,EAAE,CAAC;YACd,OAAO,EAAE,MAAM,EAAE,SAAS,EAAE,OAAO,EAAE,gBAAgB,EAAE,EAAE,CAAC;QAC5D,CAAC;QACD,IAAI,CAAC,UAAU,EAAE,CAAC;YAChB,OAAO,EAAE,MAAM,EAAE,SAAS,EAAE,OAAO,EAAE,kBAAkB,EAAE,EAAE,CAAC;QAC9D,CAAC;QACD,OAAO;YACL,MAAM,EAAE,SAAS;YACjB,OAAO,EAAE,+CAA+C;YACxD,KAAK;SACN,CAAC;IACJ,CAAC;IAED,MAAM,GAAG,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC;IAEzB,IAAI,iBAAiB,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC;QAC/B,OAAO;YACL,MAAM,EAAE,QAAQ;YAChB,OAAO,EAAE,0BAA0B;YACnC,KAAK;SACN,CAAC;IACJ,CAAC;IAED,IAAI,sBAAsB,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC;QACpC,IAAI,CAAC,cAAc,CAAC,gBAAgB,CAAC,EAAE,CAAC;YACtC,OAAO,EAAE,MAAM,EAAE,SAAS,EAAE,OAAO,EAAE,kBAAkB,EAAE,EAAE,CAAC;QAC9D,CAAC;QACD,OAAO;YACL,MAAM,EAAE,aAAa;YACrB,OAAO,EAAE,gBAAgB;YACzB,KAAK;SACN,CAAC;IACJ,CAAC;IAED,OAAO,EAAE,MAAM,EAAE,SAAS,EAAE,OAAO,EAAE,iBAAiB,CAAC,KAAK,CAAC,EAAE,CAAC;AAClE,CAAC;AAUD,SAAS,SAAS,CAAC,IAAuB;IACxC,MAAM,GAAG,GAAe,EAAE,CAAC;IAC3B,IAAI,CAAC,GAAG,CAAC,CAAC;IACV,OAAO,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC;QACvB,MAAM,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC;QAClB,IAAI,CAAC,KAAK,SAAS,EAAE,CAAC;YACpB,CAAC,EAAE,CAAC;YACJ,SAAS;QACX,CAAC;QACD,IAAI,CAAC,KAAK,IAAI,IAAI,CAAC,KAAK,QAAQ,EAAE,CAAC;YACjC,GAAG,CAAC,IAAI,GAAG,IAAI,CAAC;YAChB,CAAC,EAAE,CAAC;YACJ,SAAS;QACX,CAAC;QACD,IAAI,CAAC,KAAK,WAAW,EAAE,CAAC;YACtB,GAAG,CAAC,OAAO,GAAG,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;YAChC,CAAC,IAAI,CAAC,CAAC;YACP,SAAS;QACX,CAAC;QACD,IAAI,CAAC,CAAC,UAAU,CAAC,YAAY,CAAC,EAAE,CAAC;YAC/B,GAAG,CAAC,OAAO,GAAG,CAAC,CAAC,KAAK,CAAC,YAAY,CAAC,MAAM,CAAC,CAAC;YAC3C,CAAC,EAAE,CAAC;YACJ,SAAS;QACX,CAAC;QACD,IAAI,CAAC,CAAC,UAAU,CAAC,IAAI,CAAC,EAAE,CAAC;YACvB,MAAM,IAAI,KAAK,CAAC,0BAA0B,CAAC,EAAE,CAAC,CAAC;QACjD,CAAC;QACD,IAAI,GAAG,CAAC,KAAK,KAAK,SAAS,EAAE,CAAC;YAC5B,GAAG,CAAC,KAAK,GAAG,CAAC,CAAC;YACd,CAAC,EAAE,CAAC;YACJ,SAAS;QACX,CAAC;QACD,MAAM,IAAI,KAAK,CAAC,gCAAgC,CAAC,EAAE,CAAC,CAAC;IACvD,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAED,MAAM,SAAS,GAAG;;;;;;;;;;;CAWjB,CAAC;AAEF,MAAM,UAAU,IAAI,CAClB,OAA0B,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC;IAE/C,IAAI,IAAgB,CAAC;IACrB,IAAI,CAAC;QACH,IAAI,GAAG,SAAS,CAAC,IAAI,CAAC,CAAC;IACzB,CAAC;IAAC,OAAO,CAAC,EAAE,CAAC;QACX,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,GAAI,CAAW,CAAC,OAAO,IAAI,CAAC,CAAC;QAClD,OAAO,CAAC,CAAC;IACX,CAAC;IAED,IAAI,IAAI,CAAC,IAAI,EAAE,CAAC;QACd,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC;QAChC,OAAO,CAAC,CAAC;IACX,CAAC;IAED,IAAI,IAAI,CAAC,KAAK,KAAK,SAAS,IAAI,IAAI,CAAC,KAAK,KAAK,EAAE,EAAE,CAAC;QAClD,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,wCAAwC,CAAC,CAAC;QAC/D,OAAO,CAAC,CAAC;IACX,CAAC;IAED,MAAM,GAAG,GAAqB,EAAE,CAAC;IACjC,IAAI,IAAI,CAAC,OAAO,KAAK,SAAS,IAAI,IAAI,CAAC,OAAO,KAAK,EAAE,EAAE,CAAC;QACtD,GAAG,CAAC,OAAO,GAAG,IAAI,CAAC,OAAO,CAAC;IAC7B,CAAC;IACD,MAAM,MAAM,GAAG,kBAAkB,CAAC,IAAI,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC;IAEnD,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC;IAC7D,IAAI,MAAM,CAAC,OAAO,KAAK,SAAS,EAAE,CAAC;QACjC,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,GAAG,MAAM,CAAC,OAAO,IAAI,CAAC,CAAC;IAC9C,CAAC;IACD,OAAO,CAAC,CAAC;AACX,CAAC;AAED,MAAM,QAAQ,GAAG,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;AAChD,IAAI,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,IAAI,CAAC,OAAO,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,KAAK,QAAQ,EAAE,CAAC;IAClE,OAAO,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC;AACvB,CAAC"}

package/dist/toolkit/fetch_attachment.d.ts

@@ -0,0 +1,21 @@
+export interface FetchAttachmentOptions {
+    maxBytes?: number;
+    timeoutMs?: number;
+    headers?: Record<string, string>;
+    fetchImpl?: typeof fetch;
+}
+export interface FetchAttachmentResult {
+    rawBytes: Uint8Array;
+    contentType: string;
+    filename: string;
+    checksum: string;
+    extracted: {
+        format: "pdf" | "docx" | "pptx" | "text" | "skipped";
+        text: string | null;
+        warning?: string;
+    };
+    sizeBytes: number;
+    sourceUrl: string;
+}
+export declare function fetchAttachment(url: string, opts?: FetchAttachmentOptions): Promise<FetchAttachmentResult>;
+//# sourceMappingURL=fetch_attachment.d.ts.map

package/dist/toolkit/fetch_attachment.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"fetch_attachment.d.ts","sourceRoot":"","sources":["../../src/toolkit/fetch_attachment.ts"],"names":[],"mappings":"AA4BA,MAAM,WAAW,sBAAsB;IACrC,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,OAAO,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IACjC,SAAS,CAAC,EAAE,OAAO,KAAK,CAAC;CAC1B;AAED,MAAM,WAAW,qBAAqB;IACpC,QAAQ,EAAE,UAAU,CAAC;IACrB,WAAW,EAAE,MAAM,CAAC;IACpB,QAAQ,EAAE,MAAM,CAAC;IACjB,QAAQ,EAAE,MAAM,CAAC;IACjB,SAAS,EAAE;QACT,MAAM,EAAE,KAAK,GAAG,MAAM,GAAG,MAAM,GAAG,MAAM,GAAG,SAAS,CAAC;QACrD,IAAI,EAAE,MAAM,GAAG,IAAI,CAAC;QACpB,OAAO,CAAC,EAAE,MAAM,CAAC;KAClB,CAAC;IACF,SAAS,EAAE,MAAM,CAAC;IAClB,SAAS,EAAE,MAAM,CAAC;CACnB;AAiCD,wBAAsB,eAAe,CACnC,GAAG,EAAE,MAAM,EACX,IAAI,GAAE,sBAA2B,GAChC,OAAO,CAAC,qBAAqB,CAAC,CA0HhC"}