@unisim/doc 0.1.0

package/dist/read/doc.js

@@ -0,0 +1,360 @@
+// ---------------------------------------------------------------------------
+// DOC (Word 97–2003) → RichDoc, as TEXT ONLY. That limit is the whole story.
+//
+// A .doc is not an older .docx. It is a Microsoft Compound File — a FAT
+// filesystem in a file, with directories and a sector allocation table — and
+// inside it the `WordDocument` stream is a C struct dump from 1997. The text is
+// not in one place and not in one encoding: a PIECE TABLE maps ranges of the
+// document onto byte ranges of the stream, each piece independently either
+// CP1252 or UTF-16, and the pieces are in storage order rather than reading
+// order. Recovering the words means walking that table. Recovering the
+// FORMATTING means also walking the CHPX/PAPX binary property exception runs
+// through a two-level page tree of 512-byte bins — a second machine of similar
+// size, for bold and headings.
+//
+// So: this reads the words, in the right order, in the right encoding, and
+// stops. Paragraph breaks are kept; bold, headings, tables, lists and pictures
+// are not, and the UI says so BEFORE the button is pressed rather than leaving
+// somebody to notice their headings vanished on page three.
+//
+// The honest alternative is telling people to open it in Word and save as
+// .docx, which the notice does — but a .doc you are handed and cannot open
+// still converts to something readable here, and that is worth having.
+// ---------------------------------------------------------------------------
+import { addNotice, tidy } from '../model.js';
+/** The compound-file signature: D0 CF 11 E0 A1 B1 1A E1. */
+const OLE_SIGNATURE = [0xd0, 0xcf, 0x11, 0xe0, 0xa1, 0xb1, 0x1a, 0xe1];
+export async function readDoc(file) {
+    const bytes = new Uint8Array(await file.arrayBuffer());
+    if (!OLE_SIGNATURE.every((b, i) => bytes[i] === b)) {
+        // RTF and DOCX both get saved with a .doc extension often enough to be
+        // worth naming rather than failing as "damaged".
+        if (bytes[0] === 0x7b && bytes[1] === 0x5c) {
+            throw new Error('This is really an RTF file with a .doc name — rename it to .rtf and it will convert.');
+        }
+        if (bytes[0] === 0x50 && bytes[1] === 0x4b) {
+            throw new Error('This is really a .docx with a .doc name — rename it to .docx and it will convert in full.');
+        }
+        throw new Error('This doesn’t look like a Word document inside.');
+    }
+    const cfb = new CompoundFile(bytes);
+    const stream = cfb.read('WordDocument');
+    if (!stream)
+        throw new Error('This compound file has no Word document in it.');
+    const doc = { blocks: [], notices: [] };
+    const text = extractText(stream, cfb);
+    for (const piece of splitParagraphs(text)) {
+        doc.blocks.push(piece.kind === 'break'
+            ? { kind: 'pagebreak' }
+            : { kind: 'paragraph', runs: [{ text: piece.text }] });
+    }
+    if (!doc.blocks.length)
+        throw new Error('No readable text was found in this .doc.');
+    addNotice(doc, 'This is the old Word format, so only its TEXT could be read — bold, headings, lists, tables, page breaks and pictures are not in the result. Opening it in Word and saving as .docx keeps all of that.');
+    // ⚠️ Said separately, and said even though it will not apply to most files,
+    // because it is the one loss here that can EMBARRASS somebody rather than
+    // merely disappoint them. A .doc's deleted text is stored as ordinary text in
+    // the piece table with a revision flag hung off it in the CHPX runs, and this
+    // reader does not walk those — so a document with tracked changes turned on
+    // converts with the deletions still in it, reading as though they were never
+    // made. Detecting it properly needs the whole property-exception machine that
+    // the header explains is out of scope, so it is disclosed instead of guessed
+    // at. It sits beside the file on the row, before anyone sends it on.
+    addNotice(doc, 'If this document had tracked changes in it, text that was deleted may still appear — the old format stores deletions as ordinary text, and only Word can tell them apart. Check the result before sending it on.');
+    doc.title = file.name.replace(/\.docx?$/i, '');
+    return tidy(doc);
+}
+// ── The piece table ──────────────────────────────────────────────────────────
+/**
+ * Walk the FIB and the piece table to reassemble the document text.
+ *
+ * The route, which is not guessable from the bytes:
+ *   1. The FIB at offset 0 says which stream the text lives in (`fWhichTblStm`
+ *      picks `1Table` or `0Table`) and where the CLX is inside it.
+ *   2. The CLX is a list of properties followed by ONE `Pcdt` (type 0x02),
+ *      which holds the piece table.
+ *   3. The piece table is a PLCF: n+1 character positions, then n 8-byte
+ *      descriptors. Descriptor bytes 2–5 are an `fc` whose BIT 30 means
+ *      "this piece is CP1252, and the real offset is fc/2".
+ */
+function extractText(wordStream, cfb) {
+    const view = new DataView(wordStream.buffer, wordStream.byteOffset, wordStream.byteLength);
+    const flags = view.getUint16(0x000a, true);
+    const useTable1 = (flags & 0x0200) !== 0;
+    const table = cfb.read(useTable1 ? '1Table' : '0Table') ?? cfb.read('0Table') ?? cfb.read('1Table');
+    if (!table) {
+        // Word 6/95 had no table stream at all and stored text unpiece-tabled.
+        return decodeCp1252(wordStream.subarray(view.getUint32(0x0018, true), view.getUint32(0x001c, true)));
+    }
+    const clxOffset = view.getUint32(0x01a2, true);
+    const clxLength = view.getUint32(0x01a6, true);
+    if (clxOffset + clxLength > table.length || clxLength < 5) {
+        throw new Error('This .doc’s index is damaged — the text can’t be located.');
+    }
+    const clx = table.subarray(clxOffset, clxOffset + clxLength);
+    // Skip the leading Prc blocks (type 0x01, each with a 2-byte length) to find
+    // the one Pcdt (type 0x02) that follows them.
+    let at = 0;
+    while (at < clx.length && clx[at] === 0x01) {
+        const size = new DataView(clx.buffer, clx.byteOffset + at + 1, 2).getUint16(0, true);
+        at += 3 + size;
+    }
+    if (clx[at] !== 0x02)
+        throw new Error('This .doc’s piece table is missing.');
+    const clxView = new DataView(clx.buffer, clx.byteOffset, clx.byteLength);
+    const pieceTableSize = clxView.getUint32(at + 1, true);
+    const pieces = clx.subarray(at + 5, at + 5 + pieceTableSize);
+    const pieceView = new DataView(pieces.buffer, pieces.byteOffset, pieces.byteLength);
+    // n pieces means n+1 positions (4 bytes each) and n descriptors (8 each).
+    const count = Math.floor((pieces.length - 4) / 12);
+    if (count <= 0)
+        throw new Error('This .doc’s piece table is empty.');
+    let out = '';
+    for (let i = 0; i < count; i += 1) {
+        const start = pieceView.getUint32(i * 4, true);
+        const end = pieceView.getUint32((i + 1) * 4, true);
+        const descriptor = 4 * (count + 1) + i * 8;
+        const fc = pieceView.getUint32(descriptor + 2, true);
+        // Bit 30 is the compression flag, and it is the crux of the whole format.
+        const compressed = (fc & 0x40000000) !== 0;
+        const offset = compressed ? (fc & 0x3fffffff) / 2 : fc & 0x3fffffff;
+        const characters = end - start;
+        if (characters <= 0)
+            continue;
+        const byteLength = compressed ? characters : characters * 2;
+        const slice = wordStream.subarray(offset, offset + byteLength);
+        out += compressed ? decodeCp1252(slice) : decodeUtf16(slice);
+    }
+    return out;
+}
+/**
+ * Windows-1252, which is what "compressed" means here — one byte per character,
+ * NOT ASCII. `TextDecoder('windows-1252')` is in every browser and gets the
+ * 0x80–0x9F block (smart quotes, dashes) right, which is exactly the range a
+ * Word document is full of and the range a naive `String.fromCharCode` mangles.
+ */
+function decodeCp1252(bytes) {
+    return new TextDecoder('windows-1252').decode(bytes);
+}
+function decodeUtf16(bytes) {
+    return new TextDecoder('utf-16le').decode(bytes);
+}
+/**
+ * Word's control characters, turned back into paragraphs.
+ *
+ * \r is the paragraph mark — NOT \n, which does not appear. \x07 ends a table
+ * cell and a table row; \x0c is a page break; \x0b a line break within a
+ * paragraph; \x13–\x15 wrap field instructions, whose contents are code
+ * (`HYPERLINK "…"`, `PAGE \\* MERGEFORMAT`) rather than text and would
+ * otherwise appear inline as gibberish.
+ */
+function splitParagraphs(text) {
+    const pieces = [];
+    let cleaned = '';
+    let inFieldInstruction = false;
+    /** Turn whatever text has accumulated into paragraph pieces. */
+    const flush = () => {
+        const paragraphs = cleaned
+            .split('\r')
+            // A literal NUL pads the tail of the last piece in a great many files.
+            // Left in, it draws as a hollow box in the PDF rather than as nothing.
+            .map((p) => p.replace(/\t/g, '  ').replace(/\0/g, '').trim())
+            .filter((p, i, all) => p.length > 0 || (i > 0 && all[i - 1].length > 0));
+        for (const text of paragraphs)
+            pieces.push({ kind: 'text', text });
+        cleaned = '';
+    };
+    for (const ch of text) {
+        const code = ch.charCodeAt(0);
+        if (code === 0x13) {
+            inFieldInstruction = true;
+            continue;
+        } // field begin
+        if (code === 0x14) {
+            inFieldInstruction = false;
+            continue;
+        } // separator: result follows
+        if (code === 0x15) {
+            inFieldInstruction = false;
+            continue;
+        } // field end
+        if (inFieldInstruction)
+            continue;
+        if (code === 0x07) {
+            cleaned += '\r';
+            continue;
+        } // cell/row end
+        if (code === 0x0b) {
+            cleaned += '\n';
+            continue;
+        } // line break
+        // An explicit page-break character — what Word writes for Ctrl+Enter. It
+        // becomes a real break rather than a paragraph mark, because a .doc whose
+        // author broke each section by hand otherwise arrives as one unbroken run.
+        //
+        // ⚠️ THIS CATCHES ONLY ONE OF THE TWO WAYS A .doc HOLDS A PAGE BREAK. The
+        // other is a paragraph PROPERTY in the PAPX runs — which is what
+        // LibreOffice writes, and what the file header explains this reader does
+        // not walk. So some page breaks survive and some do not, and the notice on
+        // the row lists page breaks among what the old format loses.
+        if (code === 0x0c) {
+            flush();
+            pieces.push({ kind: 'break' });
+            continue;
+        }
+        if (code === 0x1e) {
+            cleaned += '-';
+            continue;
+        } // non-breaking hyphen
+        if (code === 0x1f)
+            continue; // optional hyphen
+        if (code === 0x02)
+            continue; // footnote mark
+        if (code < 0x20 && code !== 0x09 && code !== 0x0d)
+            continue;
+        cleaned += ch;
+    }
+    flush();
+    return pieces;
+}
+// ── The compound file ────────────────────────────────────────────────────────
+/**
+ * Enough of the Microsoft Compound File format to pull two named streams out.
+ *
+ * It is a filesystem: a header, a FAT of sector chains, a directory of entries,
+ * and — for anything under 4096 bytes — a second miniature allocation inside a
+ * ministream, because the designers did not want a 64-byte stream costing a
+ * 512-byte sector. Both paths are needed: `1Table` is routinely large and
+ * `WordDocument` always is, but a small document puts one of them in the mini
+ * FAT and reading it as a normal chain returns garbage.
+ */
+class CompoundFile {
+    bytes;
+    view;
+    sectorSize;
+    miniSectorSize;
+    fat = [];
+    miniFat = [];
+    directory = [];
+    miniStream = new Uint8Array(0);
+    constructor(bytes) {
+        this.bytes = bytes;
+        this.view = new DataView(bytes.buffer, bytes.byteOffset, bytes.byteLength);
+        this.sectorSize = 1 << this.view.getUint16(0x1e, true);
+        this.miniSectorSize = 1 << this.view.getUint16(0x20, true);
+        this.readFat();
+        this.readDirectory();
+    }
+    read(name) {
+        const entry = this.directory.find((e) => e.name === name);
+        if (!entry || entry.size === 0)
+            return null;
+        return entry.size < 4096 && this.miniStream.length
+            ? this.readChain(entry.start, entry.size, this.miniFat, this.miniSectorSize, this.miniStream)
+            : this.readChain(entry.start, entry.size, this.fat, this.sectorSize, null);
+    }
+    sectorAt(index) {
+        return (index + 1) * this.sectorSize;
+    }
+    readFat() {
+        const difatCount = this.view.getUint32(0x48, true);
+        const sectors = [];
+        // The first 109 FAT sector numbers are in the header itself; anything
+        // beyond that continues in DIFAT sectors, which only very large files need.
+        for (let i = 0; i < 109; i += 1) {
+            const sector = this.view.getUint32(0x4c + i * 4, true);
+            if (sector === 0xffffffff)
+                break;
+            sectors.push(sector);
+        }
+        let difat = this.view.getUint32(0x44, true);
+        for (let n = 0; n < difatCount && difat !== 0xffffffff && difat !== 0xfffffffe; n += 1) {
+            const base = this.sectorAt(difat);
+            const perSector = this.sectorSize / 4 - 1;
+            for (let i = 0; i < perSector; i += 1) {
+                const sector = this.view.getUint32(base + i * 4, true);
+                if (sector === 0xffffffff)
+                    break;
+                sectors.push(sector);
+            }
+            difat = this.view.getUint32(base + perSector * 4, true);
+        }
+        for (const sector of sectors) {
+            const base = this.sectorAt(sector);
+            if (base + this.sectorSize > this.bytes.length)
+                break;
+            for (let i = 0; i < this.sectorSize / 4; i += 1) {
+                this.fat.push(this.view.getUint32(base + i * 4, true));
+            }
+        }
+        let miniFatSector = this.view.getUint32(0x3c, true);
+        const miniFatCount = this.view.getUint32(0x40, true);
+        for (let n = 0; n < miniFatCount && miniFatSector < 0xfffffffa; n += 1) {
+            const base = this.sectorAt(miniFatSector);
+            if (base + this.sectorSize > this.bytes.length)
+                break;
+            for (let i = 0; i < this.sectorSize / 4; i += 1) {
+                this.miniFat.push(this.view.getUint32(base + i * 4, true));
+            }
+            miniFatSector = this.fat[miniFatSector] ?? 0xfffffffe;
+        }
+    }
+    readDirectory() {
+        const first = this.view.getUint32(0x30, true);
+        const raw = this.readChain(first, Number.MAX_SAFE_INTEGER, this.fat, this.sectorSize, null);
+        const view = new DataView(raw.buffer, raw.byteOffset, raw.byteLength);
+        for (let at = 0; at + 128 <= raw.length; at += 128) {
+            const nameLength = view.getUint16(at + 64, true);
+            if (nameLength < 2 || nameLength > 64)
+                continue;
+            // The name is UTF-16 and the length INCLUDES its terminator, which is why
+            // this is `nameLength - 2` and not `nameLength`.
+            const name = new TextDecoder('utf-16le').decode(raw.subarray(at, at + nameLength - 2));
+            const type = raw[at + 66];
+            const start = view.getUint32(at + 116, true);
+            const size = view.getUint32(at + 120, true);
+            this.directory.push({ name, type, start, size });
+        }
+        // Entry 0 is the root, and its "stream" is the ministream that every small
+        // stream is carved out of.
+        const root = this.directory[0];
+        if (root && root.size > 0) {
+            this.miniStream = this.readChain(root.start, root.size, this.fat, this.sectorSize, null);
+        }
+    }
+    /**
+     * Follow a sector chain.
+     *
+     * The guard on `seen` is not paranoia about hostile files: a truncated
+     * download produces a FAT whose chain loops back on itself, and without it
+     * this allocates until the tab dies.
+     */
+    readChain(start, size, fat, sectorSize, container) {
+        const parts = [];
+        const seen = new Set();
+        let sector = start;
+        let remaining = size;
+        while (sector < 0xfffffffa && remaining > 0) {
+            if (seen.has(sector))
+                break;
+            seen.add(sector);
+            const base = container ? sector * sectorSize : this.sectorAt(sector);
+            const source = container ?? this.bytes;
+            if (base >= source.length)
+                break;
+            const take = Math.min(sectorSize, remaining, source.length - base);
+            parts.push(source.subarray(base, base + take));
+            remaining -= take;
+            sector = fat[sector] ?? 0xfffffffe;
+        }
+        const total = parts.reduce((n, p) => n + p.length, 0);
+        const out = new Uint8Array(total);
+        let at = 0;
+        for (const part of parts) {
+            out.set(part, at);
+            at += part.length;
+        }
+        return out;
+    }
+}
+//# sourceMappingURL=doc.js.map

package/dist/read/doc.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"doc.js","sourceRoot":"","sources":["../../src/read/doc.ts"],"names":[],"mappings":"AAAA,8EAA8E;AAC9E,6EAA6E;AAC7E,EAAE;AACF,wEAAwE;AACxE,6EAA6E;AAC7E,gFAAgF;AAChF,6EAA6E;AAC7E,2EAA2E;AAC3E,4EAA4E;AAC5E,uEAAuE;AACvE,6EAA6E;AAC7E,+EAA+E;AAC/E,+BAA+B;AAC/B,EAAE;AACF,2EAA2E;AAC3E,+EAA+E;AAC/E,+EAA+E;AAC/E,4DAA4D;AAC5D,EAAE;AACF,0EAA0E;AAC1E,2EAA2E;AAC3E,uEAAuE;AACvE,8EAA8E;AAE9E,OAAO,EAAE,SAAS,EAAE,IAAI,EAA4B,MAAM,aAAa,CAAA;AAEvE,4DAA4D;AAC5D,MAAM,aAAa,GAAG,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,CAAA;AActE,MAAM,CAAC,KAAK,UAAU,OAAO,CAAC,IAAU;IACtC,MAAM,KAAK,GAAG,IAAI,UAAU,CAAC,MAAM,IAAI,CAAC,WAAW,EAAE,CAAC,CAAA;IAEtD,IAAI,CAAC,aAAa,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC;QACnD,uEAAuE;QACvE,iDAAiD;QACjD,IAAI,KAAK,CAAC,CAAC,CAAC,KAAK,IAAI,IAAI,KAAK,CAAC,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;YAC3C,MAAM,IAAI,KAAK,CAAC,sFAAsF,CAAC,CAAA;QACzG,CAAC;QACD,IAAI,KAAK,CAAC,CAAC,CAAC,KAAK,IAAI,IAAI,KAAK,CAAC,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;YAC3C,MAAM,IAAI,KAAK,CAAC,2FAA2F,CAAC,CAAA;QAC9G,CAAC;QACD,MAAM,IAAI,KAAK,CAAC,gDAAgD,CAAC,CAAA;IACnE,CAAC;IAED,MAAM,GAAG,GAAG,IAAI,YAAY,CAAC,KAAK,CAAC,CAAA;IACnC,MAAM,MAAM,GAAG,GAAG,CAAC,IAAI,CAAC,cAAc,CAAC,CAAA;IACvC,IAAI,CAAC,MAAM;QAAE,MAAM,IAAI,KAAK,CAAC,gDAAgD,CAAC,CAAA;IAE9E,MAAM,GAAG,GAAY,EAAE,MAAM,EAAE,EAAE,EAAE,OAAO,EAAE,EAAE,EAAE,CAAA;IAChD,MAAM,IAAI,GAAG,WAAW,CAAC,MAAM,EAAE,GAAG,CAAC,CAAA;IAErC,KAAK,MAAM,KAAK,IAAI,eAAe,CAAC,IAAI,CAAC,EAAE,CAAC;QAC1C,GAAG,CAAC,MAAM,CAAC,IAAI,CACb,KAAK,CAAC,IAAI,KAAK,OAAO;YACpB,CAAC,CAAC,EAAE,IAAI,EAAE,WAAW,EAAE;YACvB,CAAC,CAAE,EAAE,IAAI,EAAE,WAAW,EAAE,IAAI,EAAE,CAAC,EAAE,IAAI,EAAE,KAAK,CAAC,IAAI,EAAE,CAAC,EAAmB,CAC1E,CAAA;IACH,CAAC;IAED,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,MAAM;QAAE,MAAM,IAAI,KAAK,CAAC,0CAA0C,CAAC,CAAA;IAEnF,SAAS,CACP,GAAG,EACH,wMAAwM,CACzM,CAAA;IACD,4EAA4E;IAC5E,0EAA0E;IAC1E,8EAA8E;IAC9E,8EAA8E;IAC9E,4EAA4E;IAC5E,6EAA6E;IAC7E,8EAA8E;IAC9E,6EAA6E;IAC7E,qEAAqE;IACrE,SAAS,CACP,GAAG,EACH,kNAAkN,CACnN,CAAA;IACD,GAAG,CAAC,KAAK,GAAG,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,WAAW,EAAE,EAAE,CAAC,CAAA;IAC9C,OAAO,IAAI,CAAC,GAAG,CAAC,CAAA;AAClB,CAAC;AAED,gFAAgF;AAEhF;;;;;;;;;;;GAWG;AACH,SAAS,WAAW,CAAC,UAAsB,EAAE,GAAiB;IAC5D,MAAM,IAAI,GAAG,IAAI,QAAQ,CAAC,UAAU,CAAC,MAAM,EAAE,UAAU,CAAC,UAAU,EAAE,UAAU,CAAC,UAAU,CAAC,CAAA;IAE1F,MAAM,KAAK,GAAG,IAAI,CAAC,SAAS,CAAC,MAAM,EAAE,IAAI,CAAC,CAAA;IAC1C,MAAM,SAAS,GAAG,CAAC,KAAK,GAAG,MAAM,CAAC,KAAK,CAAC,CAAA;IACxC,MAAM,KAAK,GAAG,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,QAAQ,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAA;IACnG,IAAI,CAAC,KAAK,EAAE,CAAC;QACX,uEAAuE;QACvE,OAAO,YAAY,CAAC,UAAU,CAAC,QAAQ,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,EAAE,IAAI,CAAC,EAAE,IAAI,CAAC,SAAS,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC,CAAC,CAAA;IACtG,CAAC;IAED,MAAM,SAAS,GAAG,IAAI,CAAC,SAAS,CAAC,MAAM,EAAE,IAAI,CAAC,CAAA;IAC9C,MAAM,SAAS,GAAG,IAAI,CAAC,SAAS,CAAC,MAAM,EAAE,IAAI,CAAC,CAAA;IAC9C,IAAI,SAAS,GAAG,SAAS,GAAG,KAAK,CAAC,MAAM,IAAI,SAAS,GAAG,CAAC,EAAE,CAAC;QAC1D,MAAM,IAAI,KAAK,CAAC,2DAA2D,CAAC,CAAA;IAC9E,CAAC;IACD,MAAM,GAAG,GAAG,KAAK,CAAC,QAAQ,CAAC,SAAS,EAAE,SAAS,GAAG,SAAS,CAAC,CAAA;IAE5D,6EAA6E;IAC7E,8CAA8C;IAC9C,IAAI,EAAE,GAAG,CAAC,CAAA;IACV,OAAO,EAAE,GAAG,GAAG,CAAC,MAAM,IAAI,GAAG,CAAC,EAAE,CAAC,KAAK,IAAI,EAAE,CAAC;QAC3C,MAAM,IAAI,GAAG,IAAI,QAAQ,CAAC,GAAG,CAAC,MAAM,EAAE,GAAG,CAAC,UAAU,GAAG,EAAE,GAAG,CAAC,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,EAAE,IAAI,CAAC,CAAA;QACpF,EAAE,IAAI,CAAC,GAAG,IAAI,CAAA;IAChB,CAAC;IACD,IAAI,GAAG,CAAC,EAAE,CAAC,KAAK,IAAI;QAAE,MAAM,IAAI,KAAK,CAAC,qCAAqC,CAAC,CAAA;IAE5E,MAAM,OAAO,GAAG,IAAI,QAAQ,CAAC,GAAG,CAAC,MAAM,EAAE,GAAG,CAAC,UAAU,EAAE,GAAG,CAAC,UAAU,CAAC,CAAA;IACxE,MAAM,cAAc,GAAG,OAAO,CAAC,SAAS,CAAC,EAAE,GAAG,CAAC,EAAE,IAAI,CAAC,CAAA;IACtD,MAAM,MAAM,GAAG,GAAG,CAAC,QAAQ,CAAC,EAAE,GAAG,CAAC,EAAE,EAAE,GAAG,CAAC,GAAG,cAAc,CAAC,CAAA;IAC5D,MAAM,SAAS,GAAG,IAAI,QAAQ,CAAC,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,UAAU,EAAE,MAAM,CAAC,UAAU,CAAC,CAAA;IAEnF,0EAA0E;IAC1E,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC,GAAG,EAAE,CAAC,CAAA;IAClD,IAAI,KAAK,IAAI,CAAC;QAAE,MAAM,IAAI,KAAK,CAAC,mCAAmC,CAAC,CAAA;IAEpE,IAAI,GAAG,GAAG,EAAE,CAAA;IACZ,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;QAClC,MAAM,KAAK,GAAG,SAAS,CAAC,SAAS,CAAC,CAAC,GAAG,CAAC,EAAE,IAAI,CAAC,CAAA;QAC9C,MAAM,GAAG,GAAG,SAAS,CAAC,SAAS,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,EAAE,IAAI,CAAC,CAAA;QAClD,MAAM,UAAU,GAAG,CAAC,GAAG,CAAC,KAAK,GAAG,CAAC,CAAC,GAAG,CAAC,GAAG,CAAC,CAAA;QAC1C,MAAM,EAAE,GAAG,SAAS,CAAC,SAAS,CAAC,UAAU,GAAG,CAAC,EAAE,IAAI,CAAC,CAAA;QAEpD,0EAA0E;QAC1E,MAAM,UAAU,GAAG,CAAC,EAAE,GAAG,UAAU,CAAC,KAAK,CAAC,CAAA;QAC1C,MAAM,MAAM,GAAG,UAAU,CAAC,CAAC,CAAC,CAAC,EAAE,GAAG,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,GAAG,UAAU,CAAA;QACnE,MAAM,UAAU,GAAG,GAAG,GAAG,KAAK,CAAA;QAC9B,IAAI,UAAU,IAAI,CAAC;YAAE,SAAQ;QAE7B,MAAM,UAAU,GAAG,UAAU,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,UAAU,GAAG,CAAC,CAAA;QAC3D,MAAM,KAAK,GAAG,UAAU,CAAC,QAAQ,CAAC,MAAM,EAAE,MAAM,GAAG,UAAU,CAAC,CAAA;QAC9D,GAAG,IAAI,UAAU,CAAC,CAAC,CAAC,YAAY,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,WAAW,CAAC,KAAK,CAAC,CAAA;IAC9D,CAAC;IAED,OAAO,GAAG,CAAA;AACZ,CAAC;AAED;;;;;GAKG;AACH,SAAS,YAAY,CAAC,KAAiB;IACrC,OAAO,IAAI,WAAW,CAAC,cAAc,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAA;AACtD,CAAC;AAED,SAAS,WAAW,CAAC,KAAiB;IACpC,OAAO,IAAI,WAAW,CAAC,UAAU,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAA;AAClD,CAAC;AAED;;;;;;;;GAQG;AACH,SAAS,eAAe,CAAC,IAAY;IACnC,MAAM,MAAM,GAAY,EAAE,CAAA;IAC1B,IAAI,OAAO,GAAG,EAAE,CAAA;IAChB,IAAI,kBAAkB,GAAG,KAAK,CAAA;IAE9B,gEAAgE;IAChE,MAAM,KAAK,GAAG,GAAG,EAAE;QACjB,MAAM,UAAU,GAAG,OAAO;aACvB,KAAK,CAAC,IAAI,CAAC;YACZ,uEAAuE;YACvE,uEAAuE;aACtE,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;aAC5D,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,GAAG,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,IAAI,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAA;QAC1E,KAAK,MAAM,IAAI,IAAI,UAAU;YAAE,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC,CAAA;QAClE,OAAO,GAAG,EAAE,CAAA;IACd,CAAC,CAAA;IAED,KAAK,MAAM,EAAE,IAAI,IAAI,EAAE,CAAC;QACtB,MAAM,IAAI,GAAG,EAAE,CAAC,UAAU,CAAC,CAAC,CAAC,CAAA;QAC7B,IAAI,IAAI,KAAK,IAAI,EAAE,CAAC;YAAC,kBAAkB,GAAG,IAAI,CAAC;YAAC,SAAQ;QAAC,CAAC,CAAE,cAAc;QAC1E,IAAI,IAAI,KAAK,IAAI,EAAE,CAAC;YAAC,kBAAkB,GAAG,KAAK,CAAC;YAAC,SAAQ;QAAC,CAAC,CAAC,4BAA4B;QACxF,IAAI,IAAI,KAAK,IAAI,EAAE,CAAC;YAAC,kBAAkB,GAAG,KAAK,CAAC;YAAC,SAAQ;QAAC,CAAC,CAAC,YAAY;QACxE,IAAI,kBAAkB;YAAE,SAAQ;QAEhC,IAAI,IAAI,KAAK,IAAI,EAAE,CAAC;YAAC,OAAO,IAAI,IAAI,CAAC;YAAC,SAAQ;QAAC,CAAC,CAAE,eAAe;QACjE,IAAI,IAAI,KAAK,IAAI,EAAE,CAAC;YAAC,OAAO,IAAI,IAAI,CAAC;YAAC,SAAQ;QAAC,CAAC,CAAE,aAAa;QAC/D,yEAAyE;QACzE,0EAA0E;QAC1E,2EAA2E;QAC3E,EAAE;QACF,0EAA0E;QAC1E,iEAAiE;QACjE,yEAAyE;QACzE,2EAA2E;QAC3E,6DAA6D;QAC7D,IAAI,IAAI,KAAK,IAAI,EAAE,CAAC;YAAC,KAAK,EAAE,CAAC;YAAC,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,CAAC,CAAC;YAAC,SAAQ;QAAC,CAAC;QACxE,IAAI,IAAI,KAAK,IAAI,EAAE,CAAC;YAAC,OAAO,IAAI,GAAG,CAAC;YAAC,SAAQ;QAAC,CAAC,CAAG,sBAAsB;QACxE,IAAI,IAAI,KAAK,IAAI;YAAE,SAAQ,CAAuB,kBAAkB;QACpE,IAAI,IAAI,KAAK,IAAI;YAAE,SAAQ,CAAuB,gBAAgB;QAClE,IAAI,IAAI,GAAG,IAAI,IAAI,IAAI,KAAK,IAAI,IAAI,IAAI,KAAK,IAAI;YAAE,SAAQ;QAC3D,OAAO,IAAI,EAAE,CAAA;IACf,CAAC;IAED,KAAK,EAAE,CAAA;IACP,OAAO,MAAM,CAAA;AACf,CAAC;AAED,gFAAgF;AAEhF;;;;;;;;;GASG;AACH,MAAM,YAAY;IACR,KAAK,CAAY;IACjB,IAAI,CAAU;IACd,UAAU,CAAQ;IAClB,cAAc,CAAQ;IACtB,GAAG,GAAa,EAAE,CAAA;IAClB,OAAO,GAAa,EAAE,CAAA;IACtB,SAAS,GAAe,EAAE,CAAA;IAC1B,UAAU,GAAe,IAAI,UAAU,CAAC,CAAC,CAAC,CAAA;IAElD,YAAY,KAAiB;QAC3B,IAAI,CAAC,KAAK,GAAG,KAAK,CAAA;QAClB,IAAI,CAAC,IAAI,GAAG,IAAI,QAAQ,CAAC,KAAK,CAAC,MAAM,EAAE,KAAK,CAAC,UAAU,EAAE,KAAK,CAAC,UAAU,CAAC,CAAA;QAC1E,IAAI,CAAC,UAAU,GAAG,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,EAAE,IAAI,CAAC,CAAA;QACtD,IAAI,CAAC,cAAc,GAAG,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,EAAE,IAAI,CAAC,CAAA;QAC1D,IAAI,CAAC,OAAO,EAAE,CAAA;QACd,IAAI,CAAC,aAAa,EAAE,CAAA;IACtB,CAAC;IAED,IAAI,CAAC,IAAY;QACf,MAAM,KAAK,GAAG,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,IAAI,CAAC,CAAA;QACzD,IAAI,CAAC,KAAK,IAAI,KAAK,CAAC,IAAI,KAAK,CAAC;YAAE,OAAO,IAAI,CAAA;QAC3C,OAAO,KAAK,CAAC,IAAI,GAAG,IAAI,IAAI,IAAI,CAAC,UAAU,CAAC,MAAM;YAChD,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,KAAK,EAAE,KAAK,CAAC,IAAI,EAAE,IAAI,CAAC,OAAO,EAAE,IAAI,CAAC,cAAc,EAAE,IAAI,CAAC,UAAU,CAAC;YAC7F,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,KAAK,EAAE,KAAK,CAAC,IAAI,EAAE,IAAI,CAAC,GAAG,EAAE,IAAI,CAAC,UAAU,EAAE,IAAI,CAAC,CAAA;IAC9E,CAAC;IAEO,QAAQ,CAAC,KAAa;QAC5B,OAAO,CAAC,KAAK,GAAG,CAAC,CAAC,GAAG,IAAI,CAAC,UAAU,CAAA;IACtC,CAAC;IAEO,OAAO;QACb,MAAM,UAAU,GAAG,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,EAAE,IAAI,CAAC,CAAA;QAClD,MAAM,OAAO,GAAa,EAAE,CAAA;QAE5B,sEAAsE;QACtE,4EAA4E;QAC5E,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,GAAG,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;YAChC,MAAM,MAAM,GAAG,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,GAAG,CAAC,GAAG,CAAC,EAAE,IAAI,CAAC,CAAA;YACtD,IAAI,MAAM,KAAK,UAAU;gBAAE,MAAK;YAChC,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;QACtB,CAAC;QAED,IAAI,KAAK,GAAG,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,EAAE,IAAI,CAAC,CAAA;QAC3C,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,UAAU,IAAI,KAAK,KAAK,UAAU,IAAI,KAAK,KAAK,UAAU,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;YACvF,MAAM,IAAI,GAAG,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAA;YACjC,MAAM,SAAS,GAAG,IAAI,CAAC,UAAU,GAAG,CAAC,GAAG,CAAC,CAAA;YACzC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,SAAS,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;gBACtC,MAAM,MAAM,GAAG,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,GAAG,CAAC,GAAG,CAAC,EAAE,IAAI,CAAC,CAAA;gBACtD,IAAI,MAAM,KAAK,UAAU;oBAAE,MAAK;gBAChC,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;YACtB,CAAC;YACD,KAAK,GAAG,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,GAAG,SAAS,GAAG,CAAC,EAAE,IAAI,CAAC,CAAA;QACzD,CAAC;QAED,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;YAC7B,MAAM,IAAI,GAAG,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAA;YAClC,IAAI,IAAI,GAAG,IAAI,CAAC,UAAU,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM;gBAAE,MAAK;YACrD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,UAAU,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;gBAChD,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,GAAG,CAAC,GAAG,CAAC,EAAE,IAAI,CAAC,CAAC,CAAA;YACxD,CAAC;QACH,CAAC;QAED,IAAI,aAAa,GAAG,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,EAAE,IAAI,CAAC,CAAA;QACnD,MAAM,YAAY,GAAG,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,EAAE,IAAI,CAAC,CAAA;QACpD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,YAAY,IAAI,aAAa,GAAG,UAAU,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;YACvE,MAAM,IAAI,GAAG,IAAI,CAAC,QAAQ,CAAC,aAAa,CAAC,CAAA;YACzC,IAAI,IAAI,GAAG,IAAI,CAAC,UAAU,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM;gBAAE,MAAK;YACrD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,UAAU,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;gBAChD,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,GAAG,CAAC,GAAG,CAAC,EAAE,IAAI,CAAC,CAAC,CAAA;YAC5D,CAAC;YACD,aAAa,GAAG,IAAI,CAAC,GAAG,CAAC,aAAa,CAAC,IAAI,UAAU,CAAA;QACvD,CAAC;IACH,CAAC;IAEO,aAAa;QACnB,MAAM,KAAK,GAAG,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,EAAE,IAAI,CAAC,CAAA;QAC7C,MAAM,GAAG,GAAG,IAAI,CAAC,SAAS,CAAC,KAAK,EAAE,MAAM,CAAC,gBAAgB,EAAE,IAAI,CAAC,GAAG,EAAE,IAAI,CAAC,UAAU,EAAE,IAAI,CAAC,CAAA;QAC3F,MAAM,IAAI,GAAG,IAAI,QAAQ,CAAC,GAAG,CAAC,MAAM,EAAE,GAAG,CAAC,UAAU,EAAE,GAAG,CAAC,UAAU,CAAC,CAAA;QAErE,KAAK,IAAI,EAAE,GAAG,CAAC,EAAE,EAAE,GAAG,GAAG,IAAI,GAAG,CAAC,MAAM,EAAE,EAAE,IAAI,GAAG,EAAE,CAAC;YACnD,MAAM,UAAU,GAAG,IAAI,CAAC,SAAS,CAAC,EAAE,GAAG,EAAE,EAAE,IAAI,CAAC,CAAA;YAChD,IAAI,UAAU,GAAG,CAAC,IAAI,UAAU,GAAG,EAAE;gBAAE,SAAQ;YAC/C,0EAA0E;YAC1E,iDAAiD;YACjD,MAAM,IAAI,GAAG,IAAI,WAAW,CAAC,UAAU,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,QAAQ,CAAC,EAAE,EAAE,EAAE,GAAG,UAAU,GAAG,CAAC,CAAC,CAAC,CAAA;YACtF,MAAM,IAAI,GAAG,GAAG,CAAC,EAAE,GAAG,EAAE,CAAC,CAAA;YACzB,MAAM,KAAK,GAAG,IAAI,CAAC,SAAS,CAAC,EAAE,GAAG,GAAG,EAAE,IAAI,CAAC,CAAA;YAC5C,MAAM,IAAI,GAAG,IAAI,CAAC,SAAS,CAAC,EAAE,GAAG,GAAG,EAAE,IAAI,CAAC,CAAA;YAC3C,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,KAAK,EAAE,IAAI,EAAE,CAAC,CAAA;QAClD,CAAC;QAED,2EAA2E;QAC3E,2BAA2B;QAC3B,MAAM,IAAI,GAAG,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,CAAA;QAC9B,IAAI,IAAI,IAAI,IAAI,CAAC,IAAI,GAAG,CAAC,EAAE,CAAC;YAC1B,IAAI,CAAC,UAAU,GAAG,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,KAAK,EAAE,IAAI,CAAC,IAAI,EAAE,IAAI,CAAC,GAAG,EAAE,IAAI,CAAC,UAAU,EAAE,IAAI,CAAC,CAAA;QAC1F,CAAC;IACH,CAAC;IAED;;;;;;OAMG;IACK,SAAS,CACf,KAAa,EACb,IAAY,EACZ,GAAa,EACb,UAAkB,EAClB,SAA4B;QAE5B,MAAM,KAAK,GAAiB,EAAE,CAAA;QAC9B,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAA;QAC9B,IAAI,MAAM,GAAG,KAAK,CAAA;QAClB,IAAI,SAAS,GAAG,IAAI,CAAA;QAEpB,OAAO,MAAM,GAAG,UAAU,IAAI,SAAS,GAAG,CAAC,EAAE,CAAC;YAC5C,IAAI,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC;gBAAE,MAAK;YAC3B,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,CAAA;YAChB,MAAM,IAAI,GAAG,SAAS,CAAC,CAAC,CAAC,MAAM,GAAG,UAAU,CAAC,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAA;YACpE,MAAM,MAAM,GAAG,SAAS,IAAI,IAAI,CAAC,KAAK,CAAA;YACtC,IAAI,IAAI,IAAI,MAAM,CAAC,MAAM;gBAAE,MAAK;YAChC,MAAM,IAAI,GAAG,IAAI,CAAC,GAAG,CAAC,UAAU,EAAE,SAAS,EAAE,MAAM,CAAC,MAAM,GAAG,IAAI,CAAC,CAAA;YAClE,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,IAAI,EAAE,IAAI,GAAG,IAAI,CAAC,CAAC,CAAA;YAC9C,SAAS,IAAI,IAAI,CAAA;YACjB,MAAM,GAAG,GAAG,CAAC,MAAM,CAAC,IAAI,UAAU,CAAA;QACpC,CAAC;QAED,MAAM,KAAK,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC,CAAA;QACrD,MAAM,GAAG,GAAG,IAAI,UAAU,CAAC,KAAK,CAAC,CAAA;QACjC,IAAI,EAAE,GAAG,CAAC,CAAA;QACV,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YACzB,GAAG,CAAC,GAAG,CAAC,IAAI,EAAE,EAAE,CAAC,CAAA;YACjB,EAAE,IAAI,IAAI,CAAC,MAAM,CAAA;QACnB,CAAC;QACD,OAAO,GAAG,CAAA;IACZ,CAAC;CACF"}

package/dist/read/docx.d.ts

@@ -0,0 +1,3 @@
+import { type RichDoc } from '../model.js';
+export declare function readDocx(file: File): Promise<RichDoc>;
+//# sourceMappingURL=docx.d.ts.map

package/dist/read/docx.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"docx.d.ts","sourceRoot":"","sources":["../../src/read/docx.ts"],"names":[],"mappings":"AAqBA,OAAO,EAEsB,KAAK,OAAO,EACxC,MAAM,aAAa,CAAA;AAkBpB,wBAAsB,QAAQ,CAAC,IAAI,EAAE,IAAI,GAAG,OAAO,CAAC,OAAO,CAAC,CAyB3D"}