@unisim/doc 0.1.0

package/dist/read/rtf.js

@@ -0,0 +1,341 @@
+// ---------------------------------------------------------------------------
+// RTF → RichDoc.
+//
+// RTF is a text format with braces, which makes it look far more tractable than
+// the binary .doc — and it mostly is. A tokeniser over `\control`, `{group}`
+// and literal text gets bold, italic, underline, paragraphs and headings, which
+// is most of what a converted document needs.
+//
+// THREE THINGS THAT ARE NOT OBVIOUS AND BREAK IT IF MISSED
+// --------------------------------------------------------
+//   * FORMATTING IS SCOPED TO THE GROUP. `{\b bold}` ends when the brace does,
+//     so state has to be pushed and popped with the braces rather than tracked
+//     flat. Flat tracking makes one stray `\b` bold the rest of the document.
+//   * `\'xx` IS A BYTE, NOT A CHARACTER, and which character it is depends on
+//     the document's codepage (`\ansicpg1252`). Bytes are collected and decoded
+//     as a group so a multi-byte sequence survives.
+//   * DESTINATIONS MUST BE SKIPPED. `{\fonttbl…}`, `{\colortbl…}`, `{\*\…}` and
+//     `{\stylesheet…}` are all text that is NOT document text; without skipping
+//     them a converted file opens with a list of every font it uses.
+//
+// Pictures are not read. `\pict` holds hex-encoded WMF or PNG, and the WMF half
+// browsers cannot draw at all, so the group is skipped and a notice is raised.
+// ---------------------------------------------------------------------------
+import { addNotice, mergeRuns, tidy } from '../model.js';
+/** Groups whose contents are metadata rather than document text. */
+const SKIPPED_DESTINATIONS = new Set([
+    'fonttbl', 'colortbl', 'stylesheet', 'listtable', 'listoverridetable',
+    'info', 'pict', 'header', 'footer', 'headerl', 'headerr', 'headerf',
+    'footerl', 'footerr', 'footerf', 'footnote', 'comment', 'generator',
+    'themedata', 'colorschememapping', 'latentstyles', 'datastore', 'xmlnstbl',
+    'nonshppict', 'shppict', 'field', 'filetbl', 'revtbl', 'annotation',
+    'bkmkstart', 'bkmkend', 'template', 'atnid', 'atnauthor',
+]);
+export async function readRtf(file) {
+    const source = await file.text();
+    if (!source.startsWith('{\\rt')) {
+        throw new Error('This doesn’t look like an RTF file inside.');
+    }
+    const doc = { blocks: [], notices: [] };
+    const codepage = /\\ansicpg(\d+)/.exec(source)?.[1] ?? '1252';
+    const decoder = makeDecoder(codepage);
+    const stack = [];
+    let state = {
+        bold: false, italic: false, underline: false, strike: false,
+        heading: 0, list: false, skip: false, unicodeSkip: 1,
+    };
+    let runs = [];
+    let pendingBytes = [];
+    let sawPicture = false;
+    let listItems = [];
+    let listOrdered = false;
+    const flushBytes = () => {
+        if (!pendingBytes.length)
+            return;
+        const text = decoder(new Uint8Array(pendingBytes));
+        pendingBytes = [];
+        if (!state.skip)
+            runs.push({ text, ...styleOf(state) });
+    };
+    const flushList = () => {
+        if (listItems.length) {
+            doc.blocks.push({ kind: 'list', ordered: listOrdered, items: listItems });
+            listItems = [];
+        }
+    };
+    const endParagraph = () => {
+        flushBytes();
+        const merged = mergeRuns(runs);
+        runs = [];
+        if (!merged.length) {
+            flushList();
+            doc.blocks.push({ kind: 'paragraph', runs: [{ text: ' ' }] });
+            return;
+        }
+        if (state.list) {
+            listItems.push({ runs: merged, level: 0 });
+            return;
+        }
+        flushList();
+        if (state.heading > 0) {
+            const level = Math.min(4, Math.max(1, state.heading));
+            doc.blocks.push({ kind: 'heading', level, runs: merged });
+        }
+        else {
+            doc.blocks.push({ kind: 'paragraph', runs: merged });
+        }
+    };
+    let i = 0;
+    while (i < source.length) {
+        const ch = source[i];
+        if (ch === '{') {
+            flushBytes();
+            stack.push({ ...state });
+            i += 1;
+            continue;
+        }
+        if (ch === '}') {
+            flushBytes();
+            const restored = stack.pop();
+            if (restored)
+                state = restored;
+            i += 1;
+            continue;
+        }
+        if (ch === '\\') {
+            // An escaped literal: \\ \{ \} are the characters themselves.
+            const next = source[i + 1];
+            if (next === '\\' || next === '{' || next === '}') {
+                pendingBytes.push(next.charCodeAt(0));
+                i += 2;
+                continue;
+            }
+            // A hex byte.
+            if (next === "'") {
+                const hex = source.slice(i + 2, i + 4);
+                pendingBytes.push(parseInt(hex, 16) || 0);
+                i += 4;
+                continue;
+            }
+            // ⚠️ `\*` MUST BE HANDLED HERE, before the control-word regex, and the
+            // reason is worth keeping: `*` is not a letter, so the regex below does
+            // not match it, and the fallthrough consumed the backslash and then
+            // treated the `*` as ORDINARY TEXT. Every `{\*\generator …}` and
+            // `{\*\listtable …}` LibreOffice writes therefore contributed a literal
+            // asterisk, and a converted RTF opened with "*****" before its title.
+            // It means "skip this group unless you know the destination", and we
+            // know none of them.
+            if (next === '*') {
+                flushBytes();
+                state.skip = true;
+                i += 2;
+                continue;
+            }
+            const match = /^\\([a-zA-Z]+)(-?\d+)? ?/.exec(source.slice(i));
+            if (!match) {
+                i += 1;
+                continue;
+            }
+            flushBytes();
+            const word = match[1];
+            const value = match[2] === undefined ? null : Number(match[2]);
+            i += match[0].length;
+            // A destination marker: `{\*\foo …}` means "skip this whole group if you
+            // don't know \foo", which for our purposes is always.
+            if (word === '*') {
+                state.skip = true;
+                continue;
+            }
+            if (SKIPPED_DESTINATIONS.has(word)) {
+                if (word === 'pict')
+                    sawPicture = true;
+                state.skip = true;
+                continue;
+            }
+            switch (word) {
+                case 'par':
+                case 'sect':
+                    endParagraph();
+                    break;
+                case 'line':
+                    if (!state.skip)
+                        runs.push({ text: '\n', ...styleOf(state) });
+                    break;
+                case 'page':
+                    endParagraph();
+                    doc.blocks.push({ kind: 'pagebreak' });
+                    break;
+                case 'tab':
+                    if (!state.skip)
+                        runs.push({ text: '  ', ...styleOf(state) });
+                    break;
+                case 'b':
+                    state.bold = value !== 0;
+                    break;
+                case 'i':
+                    state.italic = value !== 0;
+                    break;
+                case 'ul':
+                    state.underline = value !== 0;
+                    break;
+                case 'ulnone':
+                    state.underline = false;
+                    break;
+                case 'strike':
+                    state.strike = value !== 0;
+                    break;
+                case 'plain':
+                    state.bold = state.italic = state.underline = state.strike = false;
+                    break;
+                case 'outlinelevel':
+                    state.heading = value === null ? 0 : value + 1;
+                    break;
+                case 's':
+                    // `\s1`–`\s9` are the built-in heading styles in almost every
+                    // producer's stylesheet. Not guaranteed, but right far more often
+                    // than treating a styled heading as body text.
+                    state.heading = value !== null && value >= 1 && value <= 9 ? value : 0;
+                    break;
+                case 'pard':
+                    // Resets paragraph formatting — including list membership and any
+                    // heading level, which is what ends a list.
+                    state.heading = 0;
+                    state.list = false;
+                    break;
+                case 'ls':
+                case 'listtext':
+                    state.list = true;
+                    break;
+                case 'pntext':
+                    // The literal bullet or number a non-list-aware reader should show.
+                    // Skipped, because the writers draw their own markers and this would
+                    // double them up.
+                    state.skip = true;
+                    break;
+                case 'pnlvlblt':
+                    listOrdered = false;
+                    state.list = true;
+                    break;
+                case 'pnlvlbody':
+                    listOrdered = true;
+                    state.list = true;
+                    break;
+                case 'u': {
+                    // A Unicode character, given as a SIGNED 16-bit value — so 8212
+                    // (em dash) arrives fine but anything above U+7FFF arrives negative.
+                    if (value !== null && !state.skip) {
+                        const code = value < 0 ? value + 65536 : value;
+                        runs.push({ text: String.fromCharCode(code), ...styleOf(state) });
+                    }
+                    // `\uc` says how many CHARACTERS of ANSI fallback follow, spelling
+                    // the same glyph for a reader too old to understand `\u`. They must
+                    // be swallowed or the document gets the character twice.
+                    i = skipFallback(source, i, state.unicodeSkip);
+                    break;
+                }
+                case 'uc':
+                    state.unicodeSkip = value ?? 1;
+                    break;
+                case 'cell':
+                    if (!state.skip)
+                        runs.push({ text: '  ', ...styleOf(state) });
+                    break;
+                case 'row':
+                    endParagraph();
+                    break;
+                default:
+                    break;
+            }
+            continue;
+        }
+        if (ch === '\n' || ch === '\r') {
+            i += 1;
+            continue;
+        }
+        pendingBytes.push(ch.charCodeAt(0));
+        i += 1;
+    }
+    endParagraph();
+    flushList();
+    if (sawPicture) {
+        addNotice(doc, 'Pictures were left out — RTF stores them in a Windows format browsers can’t draw.');
+    }
+    // RTF's tables are rows of cells with no structure around them, so they come
+    // through as tab-separated lines rather than as a laid-out grid.
+    if (/\\trowd/.test(source)) {
+        addNotice(doc, 'Tables came through as plain lines — RTF doesn’t describe them as a grid.');
+    }
+    if (!doc.blocks.length)
+        throw new Error('No readable text was found in this RTF.');
+    doc.title = /\{\\title ([^}]*)\}/.exec(source)?.[1]?.trim() || undefined;
+    return tidy(doc);
+}
+/**
+ * Step over the ANSI fallback that follows a `\uN`.
+ *
+ * ⚠️ `\uc1` means ONE CHARACTER, NOT ONE STRING POSITION, and conflating the
+ * two is the bug this function exists to prevent. The fallback for an em dash
+ * is written `\'97` — four characters spelling one byte — so advancing the
+ * cursor by `unicodeSkip` left `'97` behind as literal text, and every em dash
+ * in a converted RTF came out as `—'97`. A fallback character can be:
+ *
+ *   \'xx   a hex byte, four characters
+ *   \\ \{  an escaped literal, two characters
+ *   x      an ordinary character, one
+ *
+ * A control word among them ends the run early — it is the next instruction,
+ * not a fallback — so the loop stops rather than eating it.
+ */
+function skipFallback(source, at, count) {
+    let i = at;
+    for (let n = 0; n < count && i < source.length; n += 1) {
+        if (source[i] === '\\') {
+            const next = source[i + 1];
+            if (next === "'") {
+                i += 4;
+                continue;
+            }
+            if (next === '\\' || next === '{' || next === '}') {
+                i += 2;
+                continue;
+            }
+            break;
+        }
+        if (source[i] === '{' || source[i] === '}')
+            break;
+        i += 1;
+    }
+    return i;
+}
+function styleOf(state) {
+    const style = {};
+    if (state.bold)
+        style.bold = true;
+    if (state.italic)
+        style.italic = true;
+    if (state.underline)
+        style.underline = true;
+    if (state.strike)
+        style.strike = true;
+    return style;
+}
+/**
+ * A decoder for the document's declared codepage.
+ *
+ * `TextDecoder` knows these by their web names, and an unknown one throws
+ * rather than falling back — so an exotic `\ansicpg` would take the whole
+ * conversion down. Windows-1252 is the overwhelmingly common case and a decent
+ * last resort.
+ */
+function makeDecoder(codepage) {
+    const label = codepage === '65001' ? 'utf-8' : `windows-${codepage}`;
+    try {
+        const decoder = new TextDecoder(label);
+        return (bytes) => decoder.decode(bytes);
+    }
+    catch {
+        const fallback = new TextDecoder('windows-1252');
+        return (bytes) => fallback.decode(bytes);
+    }
+}
+//# sourceMappingURL=rtf.js.map

package/dist/read/rtf.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"rtf.js","sourceRoot":"","sources":["../../src/read/rtf.ts"],"names":[],"mappings":"AAAA,8EAA8E;AAC9E,iBAAiB;AACjB,EAAE;AACF,gFAAgF;AAChF,6EAA6E;AAC7E,gFAAgF;AAChF,8CAA8C;AAC9C,EAAE;AACF,2DAA2D;AAC3D,2DAA2D;AAC3D,+EAA+E;AAC/E,+EAA+E;AAC/E,8EAA8E;AAC9E,8EAA8E;AAC9E,gFAAgF;AAChF,oDAAoD;AACpD,gFAAgF;AAChF,gFAAgF;AAChF,qEAAqE;AACrE,EAAE;AACF,gFAAgF;AAChF,+EAA+E;AAC/E,8EAA8E;AAE9E,OAAO,EAAE,SAAS,EAAE,SAAS,EAAE,IAAI,EAAqD,MAAM,aAAa,CAAA;AAE3G,oEAAoE;AACpE,MAAM,oBAAoB,GAAG,IAAI,GAAG,CAAC;IACnC,SAAS,EAAE,UAAU,EAAE,YAAY,EAAE,WAAW,EAAE,mBAAmB;IACrE,MAAM,EAAE,MAAM,EAAE,QAAQ,EAAE,QAAQ,EAAE,SAAS,EAAE,SAAS,EAAE,SAAS;IACnE,SAAS,EAAE,SAAS,EAAE,SAAS,EAAE,UAAU,EAAE,SAAS,EAAE,WAAW;IACnE,WAAW,EAAE,oBAAoB,EAAE,cAAc,EAAE,WAAW,EAAE,UAAU;IAC1E,YAAY,EAAE,SAAS,EAAE,OAAO,EAAE,SAAS,EAAE,QAAQ,EAAE,YAAY;IACnE,WAAW,EAAE,SAAS,EAAE,UAAU,EAAE,OAAO,EAAE,WAAW;CACzD,CAAC,CAAA;AAgBF,MAAM,CAAC,KAAK,UAAU,OAAO,CAAC,IAAU;IACtC,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,IAAI,EAAE,CAAA;IAChC,IAAI,CAAC,MAAM,CAAC,UAAU,CAAC,OAAO,CAAC,EAAE,CAAC;QAChC,MAAM,IAAI,KAAK,CAAC,4CAA4C,CAAC,CAAA;IAC/D,CAAC;IAED,MAAM,GAAG,GAAY,EAAE,MAAM,EAAE,EAAE,EAAE,OAAO,EAAE,EAAE,EAAE,CAAA;IAChD,MAAM,QAAQ,GAAG,gBAAgB,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,MAAM,CAAA;IAC7D,MAAM,OAAO,GAAG,WAAW,CAAC,QAAQ,CAAC,CAAA;IAErC,MAAM,KAAK,GAAY,EAAE,CAAA;IACzB,IAAI,KAAK,GAAU;QACjB,IAAI,EAAE,KAAK,EAAE,MAAM,EAAE,KAAK,EAAE,SAAS,EAAE,KAAK,EAAE,MAAM,EAAE,KAAK;QAC3D,OAAO,EAAE,CAAC,EAAE,IAAI,EAAE,KAAK,EAAE,IAAI,EAAE,KAAK,EAAE,WAAW,EAAE,CAAC;KACrD,CAAA;IAED,IAAI,IAAI,GAAU,EAAE,CAAA;IACpB,IAAI,YAAY,GAAa,EAAE,CAAA;IAC/B,IAAI,UAAU,GAAG,KAAK,CAAA;IACtB,IAAI,SAAS,GAAe,EAAE,CAAA;IAC9B,IAAI,WAAW,GAAG,KAAK,CAAA;IAEvB,MAAM,UAAU,GAAG,GAAG,EAAE;QACtB,IAAI,CAAC,YAAY,CAAC,MAAM;YAAE,OAAM;QAChC,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,UAAU,CAAC,YAAY,CAAC,CAAC,CAAA;QAClD,YAAY,GAAG,EAAE,CAAA;QACjB,IAAI,CAAC,KAAK,CAAC,IAAI;YAAE,IAAI,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC,CAAA;IACzD,CAAC,CAAA;IAED,MAAM,SAAS,GAAG,GAAG,EAAE;QACrB,IAAI,SAAS,CAAC,MAAM,EAAE,CAAC;YACrB,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,OAAO,EAAE,WAAW,EAAE,KAAK,EAAE,SAAS,EAAE,CAAC,CAAA;YACzE,SAAS,GAAG,EAAE,CAAA;QAChB,CAAC;IACH,CAAC,CAAA;IAED,MAAM,YAAY,GAAG,GAAG,EAAE;QACxB,UAAU,EAAE,CAAA;QACZ,MAAM,MAAM,GAAG,SAAS,CAAC,IAAI,CAAC,CAAA;QAC9B,IAAI,GAAG,EAAE,CAAA;QACT,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC;YACnB,SAAS,EAAE,CAAA;YACX,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,WAAW,EAAE,IAAI,EAAE,CAAC,EAAE,IAAI,EAAE,GAAG,EAAE,CAAC,EAAE,CAAC,CAAA;YAC7D,OAAM;QACR,CAAC;QACD,IAAI,KAAK,CAAC,IAAI,EAAE,CAAC;YACf,SAAS,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,EAAE,CAAC,CAAA;YAC1C,OAAM;QACR,CAAC;QACD,SAAS,EAAE,CAAA;QACX,IAAI,KAAK,CAAC,OAAO,GAAG,CAAC,EAAE,CAAC;YACtB,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,CAAC,OAAO,CAAC,CAAkB,CAAA;YACtE,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,SAAS,EAAE,KAAK,EAAE,IAAI,EAAE,MAAM,EAAE,CAAC,CAAA;QAC3D,CAAC;aAAM,CAAC;YACN,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,WAAW,EAAE,IAAI,EAAE,MAAM,EAAkB,CAAC,CAAA;QACtE,CAAC;IACH,CAAC,CAAA;IAED,IAAI,CAAC,GAAG,CAAC,CAAA;IACT,OAAO,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC;QACzB,MAAM,EAAE,GAAG,MAAM,CAAC,CAAC,CAAC,CAAA;QAEpB,IAAI,EAAE,KAAK,GAAG,EAAE,CAAC;YACf,UAAU,EAAE,CAAA;YACZ,KAAK,CAAC,IAAI,CAAC,EAAE,GAAG,KAAK,EAAE,CAAC,CAAA;YACxB,CAAC,IAAI,CAAC,CAAA;YACN,SAAQ;QACV,CAAC;QAED,IAAI,EAAE,KAAK,GAAG,EAAE,CAAC;YACf,UAAU,EAAE,CAAA;YACZ,MAAM,QAAQ,GAAG,KAAK,CAAC,GAAG,EAAE,CAAA;YAC5B,IAAI,QAAQ;gBAAE,KAAK,GAAG,QAAQ,CAAA;YAC9B,CAAC,IAAI,CAAC,CAAA;YACN,SAAQ;QACV,CAAC;QAED,IAAI,EAAE,KAAK,IAAI,EAAE,CAAC;YAChB,8DAA8D;YAC9D,MAAM,IAAI,GAAG,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,CAAA;YAC1B,IAAI,IAAI,KAAK,IAAI,IAAI,IAAI,KAAK,GAAG,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;gBAClD,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,CAAA;gBACrC,CAAC,IAAI,CAAC,CAAA;gBACN,SAAQ;YACV,CAAC;YAED,cAAc;YACd,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;gBACjB,MAAM,GAAG,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,CAAA;gBACtC,YAAY,CAAC,IAAI,CAAC,QAAQ,CAAC,GAAG,EAAE,EAAE,CAAC,IAAI,CAAC,CAAC,CAAA;gBACzC,CAAC,IAAI,CAAC,CAAA;gBACN,SAAQ;YACV,CAAC;YAED,uEAAuE;YACvE,wEAAwE;YACxE,oEAAoE;YACpE,iEAAiE;YACjE,wEAAwE;YACxE,sEAAsE;YACtE,qEAAqE;YACrE,qBAAqB;YACrB,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;gBACjB,UAAU,EAAE,CAAA;gBACZ,KAAK,CAAC,IAAI,GAAG,IAAI,CAAA;gBACjB,CAAC,IAAI,CAAC,CAAA;gBACN,SAAQ;YACV,CAAC;YAED,MAAM,KAAK,GAAG,0BAA0B,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAA;YAC9D,IAAI,CAAC,KAAK,EAAE,CAAC;gBACX,CAAC,IAAI,CAAC,CAAA;gBACN,SAAQ;YACV,CAAC;YACD,UAAU,EAAE,CAAA;YACZ,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,CAAA;YACrB,MAAM,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,KAAK,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAA;YAC9D,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,CAAA;YAEpB,yEAAyE;YACzE,sDAAsD;YACtD,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;gBACjB,KAAK,CAAC,IAAI,GAAG,IAAI,CAAA;gBACjB,SAAQ;YACV,CAAC;YAED,IAAI,oBAAoB,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC;gBACnC,IAAI,IAAI,KAAK,MAAM;oBAAE,UAAU,GAAG,IAAI,CAAA;gBACtC,KAAK,CAAC,IAAI,GAAG,IAAI,CAAA;gBACjB,SAAQ;YACV,CAAC;YAED,QAAQ,IAAI,EAAE,CAAC;gBACb,KAAK,KAAK,CAAC;gBACX,KAAK,MAAM;oBACT,YAAY,EAAE,CAAA;oBACd,MAAK;gBACP,KAAK,MAAM;oBACT,IAAI,CAAC,KAAK,CAAC,IAAI;wBAAE,IAAI,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC,CAAA;oBAC7D,MAAK;gBACP,KAAK,MAAM;oBACT,YAAY,EAAE,CAAA;oBACd,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,WAAW,EAAE,CAAC,CAAA;oBACtC,MAAK;gBACP,KAAK,KAAK;oBACR,IAAI,CAAC,KAAK,CAAC,IAAI;wBAAE,IAAI,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC,CAAA;oBAC7D,MAAK;gBACP,KAAK,GAAG;oBAAE,KAAK,CAAC,IAAI,GAAG,KAAK,KAAK,CAAC,CAAC;oBAAC,MAAK;gBACzC,KAAK,GAAG;oBAAE,KAAK,CAAC,MAAM,GAAG,KAAK,KAAK,CAAC,CAAC;oBAAC,MAAK;gBAC3C,KAAK,IAAI;oBAAE,KAAK,CAAC,SAAS,GAAG,KAAK,KAAK,CAAC,CAAC;oBAAC,MAAK;gBAC/C,KAAK,QAAQ;oBAAE,KAAK,CAAC,SAAS,GAAG,KAAK,CAAC;oBAAC,MAAK;gBAC7C,KAAK,QAAQ;oBAAE,KAAK,CAAC,MAAM,GAAG,KAAK,KAAK,CAAC,CAAC;oBAAC,MAAK;gBAChD,KAAK,OAAO;oBACV,KAAK,CAAC,IAAI,GAAG,KAAK,CAAC,MAAM,GAAG,KAAK,CAAC,SAAS,GAAG,KAAK,CAAC,MAAM,GAAG,KAAK,CAAA;oBAClE,MAAK;gBACP,KAAK,cAAc;oBACjB,KAAK,CAAC,OAAO,GAAG,KAAK,KAAK,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAA;oBAC9C,MAAK;gBACP,KAAK,GAAG;oBACN,8DAA8D;oBAC9D,kEAAkE;oBAClE,+CAA+C;oBAC/C,KAAK,CAAC,OAAO,GAAG,KAAK,KAAK,IAAI,IAAI,KAAK,IAAI,CAAC,IAAI,KAAK,IAAI,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAA;oBACtE,MAAK;gBACP,KAAK,MAAM;oBACT,kEAAkE;oBAClE,4CAA4C;oBAC5C,KAAK,CAAC,OAAO,GAAG,CAAC,CAAA;oBACjB,KAAK,CAAC,IAAI,GAAG,KAAK,CAAA;oBAClB,MAAK;gBACP,KAAK,IAAI,CAAC;gBACV,KAAK,UAAU;oBACb,KAAK,CAAC,IAAI,GAAG,IAAI,CAAA;oBACjB,MAAK;gBACP,KAAK,QAAQ;oBACX,oEAAoE;oBACpE,qEAAqE;oBACrE,kBAAkB;oBAClB,KAAK,CAAC,IAAI,GAAG,IAAI,CAAA;oBACjB,MAAK;gBACP,KAAK,UAAU;oBACb,WAAW,GAAG,KAAK,CAAA;oBACnB,KAAK,CAAC,IAAI,GAAG,IAAI,CAAA;oBACjB,MAAK;gBACP,KAAK,WAAW;oBACd,WAAW,GAAG,IAAI,CAAA;oBAClB,KAAK,CAAC,IAAI,GAAG,IAAI,CAAA;oBACjB,MAAK;gBACP,KAAK,GAAG,CAAC,CAAC,CAAC;oBACT,gEAAgE;oBAChE,qEAAqE;oBACrE,IAAI,KAAK,KAAK,IAAI,IAAI,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC;wBAClC,MAAM,IAAI,GAAG,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,KAAK,CAAA;wBAC9C,IAAI,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,MAAM,CAAC,YAAY,CAAC,IAAI,CAAC,EAAE,GAAG,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC,CAAA;oBACnE,CAAC;oBACD,mEAAmE;oBACnE,oEAAoE;oBACpE,yDAAyD;oBACzD,CAAC,GAAG,YAAY,CAAC,MAAM,EAAE,CAAC,EAAE,KAAK,CAAC,WAAW,CAAC,CAAA;oBAC9C,MAAK;gBACP,CAAC;gBACD,KAAK,IAAI;oBACP,KAAK,CAAC,WAAW,GAAG,KAAK,IAAI,CAAC,CAAA;oBAC9B,MAAK;gBACP,KAAK,MAAM;oBACT,IAAI,CAAC,KAAK,CAAC,IAAI;wBAAE,IAAI,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC,CAAA;oBAC7D,MAAK;gBACP,KAAK,KAAK;oBACR,YAAY,EAAE,CAAA;oBACd,MAAK;gBACP;oBACE,MAAK;YACT,CAAC;YACD,SAAQ;QACV,CAAC;QAED,IAAI,EAAE,KAAK,IAAI,IAAI,EAAE,KAAK,IAAI,EAAE,CAAC;YAC/B,CAAC,IAAI,CAAC,CAAA;YACN,SAAQ;QACV,CAAC;QAED,YAAY,CAAC,IAAI,CAAC,EAAE,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,CAAA;QACnC,CAAC,IAAI,CAAC,CAAA;IACR,CAAC;IAED,YAAY,EAAE,CAAA;IACd,SAAS,EAAE,CAAA;IAEX,IAAI,UAAU,EAAE,CAAC;QACf,SAAS,CAAC,GAAG,EAAE,mFAAmF,CAAC,CAAA;IACrG,CAAC;IACD,6EAA6E;IAC7E,iEAAiE;IACjE,IAAI,SAAS,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC;QAC3B,SAAS,CAAC,GAAG,EAAE,2EAA2E,CAAC,CAAA;IAC7F,CAAC;IAED,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,MAAM;QAAE,MAAM,IAAI,KAAK,CAAC,yCAAyC,CAAC,CAAA;IAClF,GAAG,CAAC,KAAK,GAAG,qBAAqB,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,IAAI,SAAS,CAAA;IACxE,OAAO,IAAI,CAAC,GAAG,CAAC,CAAA;AAClB,CAAC;AAED;;;;;;;;;;;;;;;GAeG;AACH,SAAS,YAAY,CAAC,MAAc,EAAE,EAAU,EAAE,KAAa;IAC7D,IAAI,CAAC,GAAG,EAAE,CAAA;IACV,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,IAAI,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;QACvD,IAAI,MAAM,CAAC,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;YACvB,MAAM,IAAI,GAAG,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,CAAA;YAC1B,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;gBAAC,CAAC,IAAI,CAAC,CAAC;gBAAC,SAAQ;YAAC,CAAC;YACtC,IAAI,IAAI,KAAK,IAAI,IAAI,IAAI,KAAK,GAAG,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;gBAAC,CAAC,IAAI,CAAC,CAAC;gBAAC,SAAQ;YAAC,CAAC;YACvE,MAAK;QACP,CAAC;QACD,IAAI,MAAM,CAAC,CAAC,CAAC,KAAK,GAAG,IAAI,MAAM,CAAC,CAAC,CAAC,KAAK,GAAG;YAAE,MAAK;QACjD,CAAC,IAAI,CAAC,CAAA;IACR,CAAC;IACD,OAAO,CAAC,CAAA;AACV,CAAC;AAED,SAAS,OAAO,CAAC,KAAY;IAC3B,MAAM,KAAK,GAAiB,EAAE,CAAA;IAC9B,IAAI,KAAK,CAAC,IAAI;QAAE,KAAK,CAAC,IAAI,GAAG,IAAI,CAAA;IACjC,IAAI,KAAK,CAAC,MAAM;QAAE,KAAK,CAAC,MAAM,GAAG,IAAI,CAAA;IACrC,IAAI,KAAK,CAAC,SAAS;QAAE,KAAK,CAAC,SAAS,GAAG,IAAI,CAAA;IAC3C,IAAI,KAAK,CAAC,MAAM;QAAE,KAAK,CAAC,MAAM,GAAG,IAAI,CAAA;IACrC,OAAO,KAAK,CAAA;AACd,CAAC;AAED;;;;;;;GAOG;AACH,SAAS,WAAW,CAAC,QAAgB;IACnC,MAAM,KAAK,GAAG,QAAQ,KAAK,OAAO,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,WAAW,QAAQ,EAAE,CAAA;IACpE,IAAI,CAAC;QACH,MAAM,OAAO,GAAG,IAAI,WAAW,CAAC,KAAK,CAAC,CAAA;QACtC,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,CAAA;IACzC,CAAC;IAAC,MAAM,CAAC;QACP,MAAM,QAAQ,GAAG,IAAI,WAAW,CAAC,cAAc,CAAC,CAAA;QAChD,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,QAAQ,CAAC,MAAM,CAAC,KAAK,CAAC,CAAA;IAC1C,CAAC;AACH,CAAC"}

package/dist/read/text.d.ts

@@ -0,0 +1,4 @@
+import { type RichDoc } from '../model.js';
+export declare function readText(file: File): Promise<RichDoc>;
+export declare function parseTextDoc(source: string, filename?: string): RichDoc;
+//# sourceMappingURL=text.d.ts.map

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

@@ -0,0 +1 @@
+{"version":3,"file":"text.d.ts","sourceRoot":"","sources":["../../src/read/text.ts"],"names":[],"mappings":"AAeA,OAAO,EAAoB,KAAK,OAAO,EAAE,MAAM,aAAa,CAAA;AAE5D,wBAAsB,QAAQ,CAAC,IAAI,EAAE,IAAI,GAAG,OAAO,CAAC,OAAO,CAAC,CAE3D;AAED,wBAAgB,YAAY,CAAC,MAAM,EAAE,MAAM,EAAE,QAAQ,CAAC,EAAE,MAAM,GAAG,OAAO,CAwBvE"}

package/dist/read/text.js

@@ -0,0 +1,62 @@
+// ---------------------------------------------------------------------------
+// Plain text → RichDoc.
+//
+// The smallest reader, and the one that has to guess most. A .txt has no
+// structure at all, so the only question is what a BLANK LINE means — and the
+// answer differs between a wrapped-at-72-columns file from an email client and
+// a file with one paragraph per line from a note app. Both are common.
+//
+// The rule: a blank line always separates paragraphs. Within a block of lines,
+// they are JOINED if the file looks hard-wrapped (most lines near the same
+// length, none very long) and kept as separate lines otherwise. Joining a file
+// that was not hard-wrapped ruins a list; not joining one that was gives a PDF
+// with a ragged right edge every 72 characters.
+// ---------------------------------------------------------------------------
+import { tidy } from '../model.js';
+export async function readText(file) {
+    return parseTextDoc(await file.text(), file.name);
+}
+export function parseTextDoc(source, filename) {
+    const lines = source.replace(/\r\n?/g, '\n').split('\n');
+    const blocks = [];
+    const rewrap = looksHardWrapped(lines);
+    let buffer = [];
+    const flush = () => {
+        if (!buffer.length)
+            return;
+        const text = rewrap ? buffer.join(' ') : buffer.join('\n');
+        blocks.push({ kind: 'paragraph', runs: [{ text }] });
+        buffer = [];
+    };
+    for (const line of lines) {
+        if (!line.trim()) {
+            flush();
+            continue;
+        }
+        buffer.push(line.replace(/\t/g, '    ').trimEnd());
+    }
+    flush();
+    if (!blocks.length)
+        throw new Error('That text file is empty.');
+    return tidy({ blocks, notices: [], title: filename?.replace(/\.(txt|text|log)$/i, '') });
+}
+/**
+ * Does this file look hard-wrapped at a fixed column?
+ *
+ * The test is the shape of the line lengths, not any one line: a hard-wrapped
+ * file has most of its lines bunched just under a limit and none far over it.
+ * A file of one-paragraph-per-line has a wide spread and long lines.
+ */
+function looksHardWrapped(lines) {
+    const filled = lines.map((l) => l.trimEnd().length).filter((n) => n > 0);
+    if (filled.length < 4)
+        return false;
+    const longest = Math.max(...filled);
+    if (longest > 120)
+        return false; // something is a full paragraph on one line
+    // How many lines sit in the top quarter of the observed width. Hard wrapping
+    // pushes almost every line there; free-form text does not.
+    const nearLimit = filled.filter((n) => n >= longest * 0.75).length;
+    return nearLimit / filled.length > 0.6;
+}
+//# sourceMappingURL=text.js.map

package/dist/read/text.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"text.js","sourceRoot":"","sources":["../../src/read/text.ts"],"names":[],"mappings":"AAAA,8EAA8E;AAC9E,wBAAwB;AACxB,EAAE;AACF,yEAAyE;AACzE,8EAA8E;AAC9E,+EAA+E;AAC/E,uEAAuE;AACvE,EAAE;AACF,+EAA+E;AAC/E,2EAA2E;AAC3E,+EAA+E;AAC/E,+EAA+E;AAC/E,gDAAgD;AAChD,8EAA8E;AAE9E,OAAO,EAAE,IAAI,EAA4B,MAAM,aAAa,CAAA;AAE5D,MAAM,CAAC,KAAK,UAAU,QAAQ,CAAC,IAAU;IACvC,OAAO,YAAY,CAAC,MAAM,IAAI,CAAC,IAAI,EAAE,EAAE,IAAI,CAAC,IAAI,CAAC,CAAA;AACnD,CAAC;AAED,MAAM,UAAU,YAAY,CAAC,MAAc,EAAE,QAAiB;IAC5D,MAAM,KAAK,GAAG,MAAM,CAAC,OAAO,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAA;IACxD,MAAM,MAAM,GAAY,EAAE,CAAA;IAC1B,MAAM,MAAM,GAAG,gBAAgB,CAAC,KAAK,CAAC,CAAA;IAEtC,IAAI,MAAM,GAAa,EAAE,CAAA;IACzB,MAAM,KAAK,GAAG,GAAG,EAAE;QACjB,IAAI,CAAC,MAAM,CAAC,MAAM;YAAE,OAAM;QAC1B,MAAM,IAAI,GAAG,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;QAC1D,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,WAAW,EAAE,IAAI,EAAE,CAAC,EAAE,IAAI,EAAE,CAAC,EAAE,CAAC,CAAA;QACpD,MAAM,GAAG,EAAE,CAAA;IACb,CAAC,CAAA;IAED,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,EAAE,CAAC;YACjB,KAAK,EAAE,CAAA;YACP,SAAQ;QACV,CAAC;QACD,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC,OAAO,EAAE,CAAC,CAAA;IACpD,CAAC;IACD,KAAK,EAAE,CAAA;IAEP,IAAI,CAAC,MAAM,CAAC,MAAM;QAAE,MAAM,IAAI,KAAK,CAAC,0BAA0B,CAAC,CAAA;IAC/D,OAAO,IAAI,CAAC,EAAE,MAAM,EAAE,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,QAAQ,EAAE,OAAO,CAAC,oBAAoB,EAAE,EAAE,CAAC,EAAE,CAAC,CAAA;AAC1F,CAAC;AAED;;;;;;GAMG;AACH,SAAS,gBAAgB,CAAC,KAAwB;IAChD,MAAM,MAAM,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC,MAAM,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,CAAA;IACxE,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC;QAAE,OAAO,KAAK,CAAA;IAEnC,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,MAAM,CAAC,CAAA;IACnC,IAAI,OAAO,GAAG,GAAG;QAAE,OAAO,KAAK,CAAA,CAAC,4CAA4C;IAE5E,6EAA6E;IAC7E,2DAA2D;IAC3D,MAAM,SAAS,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,IAAI,OAAO,GAAG,IAAI,CAAC,CAAC,MAAM,CAAA;IAClE,OAAO,SAAS,GAAG,MAAM,CAAC,MAAM,GAAG,GAAG,CAAA;AACxC,CAAC"}

package/dist/unzip.d.ts

@@ -0,0 +1,23 @@
+export declare class ZipArchive {
+    private bytes;
+    private view;
+    private entries;
+    private constructor();
+    static open(source: Blob | ArrayBuffer): Promise<ZipArchive>;
+    /** Every path in the archive, in directory order. */
+    get names(): string[];
+    has(name: string): boolean;
+    bytesOf(name: string): Promise<Uint8Array | null>;
+    textOf(name: string): Promise<string | null>;
+    blobOf(name: string, type: string): Promise<Blob | null>;
+    /**
+     * Find the end-of-central-directory record and read every entry from it.
+     *
+     * The record is at the very end unless the archive has a comment, so it is
+     * searched for backwards over the last 64 KB — the most a comment can be.
+     */
+    private readDirectory;
+}
+/** Whether the browser can read ZIP-based documents at all. */
+export declare function unzipSupported(): boolean;
+//# sourceMappingURL=unzip.d.ts.map

package/dist/unzip.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"unzip.d.ts","sourceRoot":"","sources":["../src/unzip.ts"],"names":[],"mappings":"AAkCA,qBAAa,UAAU;IACrB,OAAO,CAAC,KAAK,CAAY;IACzB,OAAO,CAAC,IAAI,CAAU;IACtB,OAAO,CAAC,OAAO,CAA2B;IAE1C,OAAO;WAKM,IAAI,CAAC,MAAM,EAAE,IAAI,GAAG,WAAW,GAAG,OAAO,CAAC,UAAU,CAAC;IAOlE,qDAAqD;IACrD,IAAI,KAAK,IAAI,MAAM,EAAE,CAEpB;IAED,GAAG,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO;IAIpB,OAAO,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAAC,UAAU,GAAG,IAAI,CAAC;IAwBjD,MAAM,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAAC,MAAM,GAAG,IAAI,CAAC;IAO5C,MAAM,CAAC,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,GAAG,OAAO,CAAC,IAAI,GAAG,IAAI,CAAC;IAK9D;;;;;OAKG;IACH,OAAO,CAAC,aAAa;CAoDtB;AAgBD,+DAA+D;AAC/D,wBAAgB,cAAc,IAAI,OAAO,CAExC"}

package/dist/unzip.js

@@ -0,0 +1,152 @@
+// ---------------------------------------------------------------------------
+// A ZIP reader, in one page and with no dependency.
+//
+// A .docx and a .odt are both ZIPs of XML, so reading either starts here. The
+// app already WRITES ZIPs (`zip.ts`) for "Download all"; this is the other
+// direction, and it needs one thing that writing did not: DEFLATE.
+//
+// `DecompressionStream('deflate-raw')` is that, built into the browser — the
+// same primitive PalsPayIn leans on for its relay payloads. It is why this file
+// is a hundred lines instead of a fflate dependency: the compression is the
+// hard part and the platform already has it, leaving only the container, which
+// is a central directory of fixed-width records.
+//
+// WHY THE CENTRAL DIRECTORY AND NOT THE LOCAL HEADERS
+// ---------------------------------------------------
+// Walking local headers front-to-back looks simpler and is a trap: an entry
+// written by a streaming producer sets its sizes to zero and puts the real ones
+// in a data descriptor AFTER the data, so you cannot know where the entry ends
+// without decompressing it first. The central directory at the tail always has
+// the true sizes. Word and LibreOffice both produce files that need this.
+// ---------------------------------------------------------------------------
+const END_OF_CENTRAL = 0x06054b50;
+const CENTRAL_HEADER = 0x02014b50;
+const ZIP64_END_LOCATOR = 0x07064b50;
+export class ZipArchive {
+    bytes;
+    view;
+    entries = new Map();
+    constructor(bytes) {
+        this.bytes = bytes;
+        this.view = new DataView(bytes.buffer, bytes.byteOffset, bytes.byteLength);
+    }
+    static async open(source) {
+        const buffer = source instanceof Blob ? await source.arrayBuffer() : source;
+        const archive = new ZipArchive(new Uint8Array(buffer));
+        archive.readDirectory();
+        return archive;
+    }
+    /** Every path in the archive, in directory order. */
+    get names() {
+        return [...this.entries.keys()];
+    }
+    has(name) {
+        return this.entries.has(name);
+    }
+    async bytesOf(name) {
+        const entry = this.entries.get(name);
+        if (!entry)
+            return null;
+        // The local header repeats the name and extra field, and its lengths are
+        // the ones that count — a producer may write a different extra field here
+        // than in the central directory, so the data offset must be computed from
+        // this copy rather than from the directory's.
+        const local = entry.offset;
+        if (this.view.getUint32(local, true) !== 0x04034b50) {
+            throw new Error(`${name} is not where this archive says it is.`);
+        }
+        const nameLength = this.view.getUint16(local + 26, true);
+        const extraLength = this.view.getUint16(local + 28, true);
+        const start = local + 30 + nameLength + extraLength;
+        const raw = this.bytes.subarray(start, start + entry.compressedSize);
+        if (entry.method === 0)
+            return raw;
+        if (entry.method !== 8) {
+            throw new Error(`This file uses a compression method (${entry.method}) browsers can’t undo.`);
+        }
+        return inflateRaw(raw);
+    }
+    async textOf(name) {
+        const bytes = await this.bytesOf(name);
+        // Every XML part in an Office or OpenDocument package is UTF-8; the
+        // declaration says so and nothing in the wild disagrees.
+        return bytes ? new TextDecoder('utf-8').decode(bytes) : null;
+    }
+    async blobOf(name, type) {
+        const bytes = await this.bytesOf(name);
+        return bytes ? new Blob([bytes], { type }) : null;
+    }
+    /**
+     * Find the end-of-central-directory record and read every entry from it.
+     *
+     * The record is at the very end unless the archive has a comment, so it is
+     * searched for backwards over the last 64 KB — the most a comment can be.
+     */
+    readDirectory() {
+        const size = this.bytes.length;
+        if (size < 22)
+            throw new Error('That file is too small to be a document.');
+        let eocd = -1;
+        const floor = Math.max(0, size - 22 - 0xffff);
+        for (let i = size - 22; i >= floor; i -= 1) {
+            if (this.view.getUint32(i, true) === END_OF_CENTRAL) {
+                eocd = i;
+                break;
+            }
+        }
+        if (eocd < 0)
+            throw new Error('That doesn’t look like a Word or OpenDocument file.');
+        let count = this.view.getUint16(eocd + 10, true);
+        let directoryAt = this.view.getUint32(eocd + 16, true);
+        // ZIP64. A .docx never needs it, but a .odt with a few hundred megabytes of
+        // embedded images can, and the 32-bit fields read as 0xffffffff when it
+        // does — which without this points the directory scan at nothing.
+        if (directoryAt === 0xffffffff || count === 0xffff) {
+            for (let i = eocd - 20; i >= 0; i -= 1) {
+                if (this.view.getUint32(i, true) === ZIP64_END_LOCATOR) {
+                    const zip64At = Number(this.view.getBigUint64(i + 8, true));
+                    count = Number(this.view.getBigUint64(zip64At + 32, true));
+                    directoryAt = Number(this.view.getBigUint64(zip64At + 48, true));
+                    break;
+                }
+            }
+        }
+        let cursor = directoryAt;
+        for (let i = 0; i < count; i += 1) {
+            if (cursor + 46 > size || this.view.getUint32(cursor, true) !== CENTRAL_HEADER)
+                break;
+            const method = this.view.getUint16(cursor + 10, true);
+            const compressedSize = this.view.getUint32(cursor + 20, true);
+            const nameLength = this.view.getUint16(cursor + 28, true);
+            const extraLength = this.view.getUint16(cursor + 30, true);
+            const commentLength = this.view.getUint16(cursor + 32, true);
+            const offset = this.view.getUint32(cursor + 42, true);
+            // Names are UTF-8 when bit 11 of the flags is set and CP437 otherwise.
+            // Every part path we look up is ASCII, where the two agree, so decoding
+            // as UTF-8 throughout is safe and saves carrying a CP437 table.
+            const name = new TextDecoder('utf-8').decode(this.bytes.subarray(cursor + 46, cursor + 46 + nameLength));
+            this.entries.set(name, { name, method, offset, compressedSize });
+            cursor += 46 + nameLength + extraLength + commentLength;
+        }
+        if (!this.entries.size)
+            throw new Error('That archive has nothing in it.');
+    }
+}
+/**
+ * Raw DEFLATE, via the browser.
+ *
+ * 'deflate-raw' rather than 'deflate': ZIP stores the bare compressed stream
+ * with no zlib header or Adler checksum around it, and asking for 'deflate'
+ * makes the browser reject the first two bytes as a bad header.
+ */
+async function inflateRaw(data) {
+    const stream = new Blob([data])
+        .stream()
+        .pipeThrough(new DecompressionStream('deflate-raw'));
+    return new Uint8Array(await new Response(stream).arrayBuffer());
+}
+/** Whether the browser can read ZIP-based documents at all. */
+export function unzipSupported() {
+    return typeof DecompressionStream !== 'undefined';
+}
+//# sourceMappingURL=unzip.js.map

package/dist/unzip.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"unzip.js","sourceRoot":"","sources":["../src/unzip.ts"],"names":[],"mappings":"AAAA,8EAA8E;AAC9E,oDAAoD;AACpD,EAAE;AACF,8EAA8E;AAC9E,2EAA2E;AAC3E,mEAAmE;AACnE,EAAE;AACF,6EAA6E;AAC7E,gFAAgF;AAChF,4EAA4E;AAC5E,+EAA+E;AAC/E,iDAAiD;AACjD,EAAE;AACF,sDAAsD;AACtD,sDAAsD;AACtD,4EAA4E;AAC5E,gFAAgF;AAChF,+EAA+E;AAC/E,+EAA+E;AAC/E,0EAA0E;AAC1E,8EAA8E;AAE9E,MAAM,cAAc,GAAG,UAAU,CAAA;AACjC,MAAM,cAAc,GAAG,UAAU,CAAA;AACjC,MAAM,iBAAiB,GAAG,UAAU,CAAA;AAUpC,MAAM,OAAO,UAAU;IACb,KAAK,CAAY;IACjB,IAAI,CAAU;IACd,OAAO,GAAG,IAAI,GAAG,EAAiB,CAAA;IAE1C,YAAoB,KAAiB;QACnC,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;IAC5E,CAAC;IAED,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,MAA0B;QAC1C,MAAM,MAAM,GAAG,MAAM,YAAY,IAAI,CAAC,CAAC,CAAC,MAAM,MAAM,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC,MAAM,CAAA;QAC3E,MAAM,OAAO,GAAG,IAAI,UAAU,CAAC,IAAI,UAAU,CAAC,MAAM,CAAC,CAAC,CAAA;QACtD,OAAO,CAAC,aAAa,EAAE,CAAA;QACvB,OAAO,OAAO,CAAA;IAChB,CAAC;IAED,qDAAqD;IACrD,IAAI,KAAK;QACP,OAAO,CAAC,GAAG,IAAI,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC,CAAA;IACjC,CAAC;IAED,GAAG,CAAC,IAAY;QACd,OAAO,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,CAAA;IAC/B,CAAC;IAED,KAAK,CAAC,OAAO,CAAC,IAAY;QACxB,MAAM,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,CAAA;QACpC,IAAI,CAAC,KAAK;YAAE,OAAO,IAAI,CAAA;QAEvB,yEAAyE;QACzE,0EAA0E;QAC1E,0EAA0E;QAC1E,8CAA8C;QAC9C,MAAM,KAAK,GAAG,KAAK,CAAC,MAAM,CAAA;QAC1B,IAAI,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,KAAK,EAAE,IAAI,CAAC,KAAK,UAAU,EAAE,CAAC;YACpD,MAAM,IAAI,KAAK,CAAC,GAAG,IAAI,wCAAwC,CAAC,CAAA;QAClE,CAAC;QACD,MAAM,UAAU,GAAG,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,KAAK,GAAG,EAAE,EAAE,IAAI,CAAC,CAAA;QACxD,MAAM,WAAW,GAAG,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,KAAK,GAAG,EAAE,EAAE,IAAI,CAAC,CAAA;QACzD,MAAM,KAAK,GAAG,KAAK,GAAG,EAAE,GAAG,UAAU,GAAG,WAAW,CAAA;QACnD,MAAM,GAAG,GAAG,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,KAAK,EAAE,KAAK,GAAG,KAAK,CAAC,cAAc,CAAC,CAAA;QAEpE,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,GAAG,CAAA;QAClC,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACvB,MAAM,IAAI,KAAK,CAAC,wCAAwC,KAAK,CAAC,MAAM,wBAAwB,CAAC,CAAA;QAC/F,CAAC;QACD,OAAO,UAAU,CAAC,GAAG,CAAC,CAAA;IACxB,CAAC;IAED,KAAK,CAAC,MAAM,CAAC,IAAY;QACvB,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,CAAA;QACtC,oEAAoE;QACpE,yDAAyD;QACzD,OAAO,KAAK,CAAC,CAAC,CAAC,IAAI,WAAW,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,CAAA;IAC9D,CAAC;IAED,KAAK,CAAC,MAAM,CAAC,IAAY,EAAE,IAAY;QACrC,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,CAAA;QACtC,OAAO,KAAK,CAAC,CAAC,CAAC,IAAI,IAAI,CAAC,CAAC,KAAiB,CAAC,EAAE,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,CAAA;IAC/D,CAAC;IAED;;;;;OAKG;IACK,aAAa;QACnB,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM,CAAA;QAC9B,IAAI,IAAI,GAAG,EAAE;YAAE,MAAM,IAAI,KAAK,CAAC,0CAA0C,CAAC,CAAA;QAE1E,IAAI,IAAI,GAAG,CAAC,CAAC,CAAA;QACb,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,GAAG,EAAE,GAAG,MAAM,CAAC,CAAA;QAC7C,KAAK,IAAI,CAAC,GAAG,IAAI,GAAG,EAAE,EAAE,CAAC,IAAI,KAAK,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;YAC3C,IAAI,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,EAAE,IAAI,CAAC,KAAK,cAAc,EAAE,CAAC;gBACpD,IAAI,GAAG,CAAC,CAAA;gBACR,MAAK;YACP,CAAC;QACH,CAAC;QACD,IAAI,IAAI,GAAG,CAAC;YAAE,MAAM,IAAI,KAAK,CAAC,qDAAqD,CAAC,CAAA;QAEpF,IAAI,KAAK,GAAG,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,GAAG,EAAE,EAAE,IAAI,CAAC,CAAA;QAChD,IAAI,WAAW,GAAG,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,GAAG,EAAE,EAAE,IAAI,CAAC,CAAA;QAEtD,4EAA4E;QAC5E,wEAAwE;QACxE,kEAAkE;QAClE,IAAI,WAAW,KAAK,UAAU,IAAI,KAAK,KAAK,MAAM,EAAE,CAAC;YACnD,KAAK,IAAI,CAAC,GAAG,IAAI,GAAG,EAAE,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;gBACvC,IAAI,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,EAAE,IAAI,CAAC,KAAK,iBAAiB,EAAE,CAAC;oBACvD,MAAM,OAAO,GAAG,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC,GAAG,CAAC,EAAE,IAAI,CAAC,CAAC,CAAA;oBAC3D,KAAK,GAAG,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,YAAY,CAAC,OAAO,GAAG,EAAE,EAAE,IAAI,CAAC,CAAC,CAAA;oBAC1D,WAAW,GAAG,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,YAAY,CAAC,OAAO,GAAG,EAAE,EAAE,IAAI,CAAC,CAAC,CAAA;oBAChE,MAAK;gBACP,CAAC;YACH,CAAC;QACH,CAAC;QAED,IAAI,MAAM,GAAG,WAAW,CAAA;QACxB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;YAClC,IAAI,MAAM,GAAG,EAAE,GAAG,IAAI,IAAI,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,EAAE,IAAI,CAAC,KAAK,cAAc;gBAAE,MAAK;YACrF,MAAM,MAAM,GAAG,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,GAAG,EAAE,EAAE,IAAI,CAAC,CAAA;YACrD,MAAM,cAAc,GAAG,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,GAAG,EAAE,EAAE,IAAI,CAAC,CAAA;YAC7D,MAAM,UAAU,GAAG,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,GAAG,EAAE,EAAE,IAAI,CAAC,CAAA;YACzD,MAAM,WAAW,GAAG,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,GAAG,EAAE,EAAE,IAAI,CAAC,CAAA;YAC1D,MAAM,aAAa,GAAG,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,GAAG,EAAE,EAAE,IAAI,CAAC,CAAA;YAC5D,MAAM,MAAM,GAAG,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,GAAG,EAAE,EAAE,IAAI,CAAC,CAAA;YACrD,uEAAuE;YACvE,wEAAwE;YACxE,gEAAgE;YAChE,MAAM,IAAI,GAAG,IAAI,WAAW,CAAC,OAAO,CAAC,CAAC,MAAM,CAC1C,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,MAAM,GAAG,EAAE,EAAE,MAAM,GAAG,EAAE,GAAG,UAAU,CAAC,CAC3D,CAAA;YACD,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,IAAI,EAAE,EAAE,IAAI,EAAE,MAAM,EAAE,MAAM,EAAE,cAAc,EAAE,CAAC,CAAA;YAChE,MAAM,IAAI,EAAE,GAAG,UAAU,GAAG,WAAW,GAAG,aAAa,CAAA;QACzD,CAAC;QAED,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI;YAAE,MAAM,IAAI,KAAK,CAAC,iCAAiC,CAAC,CAAA;IAC5E,CAAC;CACF;AAED;;;;;;GAMG;AACH,KAAK,UAAU,UAAU,CAAC,IAAgB;IACxC,MAAM,MAAM,GAAG,IAAI,IAAI,CAAC,CAAC,IAAgB,CAAC,CAAC;SACxC,MAAM,EAAE;SACR,WAAW,CAAC,IAAI,mBAAmB,CAAC,aAAa,CAAC,CAAC,CAAA;IACtD,OAAO,IAAI,UAAU,CAAC,MAAM,IAAI,QAAQ,CAAC,MAAM,CAAC,CAAC,WAAW,EAAE,CAAC,CAAA;AACjE,CAAC;AAED,+DAA+D;AAC/D,MAAM,UAAU,cAAc;IAC5B,OAAO,OAAO,mBAAmB,KAAK,WAAW,CAAA;AACnD,CAAC"}

package/dist/write/data.d.ts

@@ -0,0 +1,16 @@
+import type { Grid } from '../model.js';
+export declare function gridToCsv(grid: Grid): string;
+export interface JsonOptions {
+    /** Turn "42" into 42 and "true" into true. */
+    inferTypes: boolean;
+}
+/**
+ * A grid as an array of objects.
+ *
+ * Columns become keys. Where the source had no header row, the keys are
+ * `column1`, `column2`… rather than the first row's values — using data as keys
+ * silently deletes a row, and a file with no header is exactly the file where
+ * every row matters.
+ */
+export declare function gridToJson(grid: Grid, options?: JsonOptions): string;
+//# sourceMappingURL=data.d.ts.map

package/dist/write/data.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"data.d.ts","sourceRoot":"","sources":["../../src/write/data.ts"],"names":[],"mappings":"AASA,OAAO,KAAK,EAAE,IAAI,EAAE,MAAM,aAAa,CAAA;AAEvC,wBAAgB,SAAS,CAAC,IAAI,EAAE,IAAI,GAAG,MAAM,CAK5C;AAcD,MAAM,WAAW,WAAW;IAC1B,8CAA8C;IAC9C,UAAU,EAAE,OAAO,CAAA;CACpB;AAED;;;;;;;GAOG;AACH,wBAAgB,UAAU,CAAC,IAAI,EAAE,IAAI,EAAE,OAAO,GAAE,WAAkC,GAAG,MAAM,CAc1F"}