@unisim/doc 0.1.0

package/dist/pdfcore.js

@@ -0,0 +1,398 @@
+// ---------------------------------------------------------------------------
+// The PDF writer — objects, pages, fonts, text, rules and images.
+//
+// WHY THIS IS STILL HAND-WRITTEN
+// ------------------------------
+// `pdf.ts` used to end with a note saying that if text was ever wanted, the
+// right move was to take a real PDF library rather than grow the writer. The
+// Files tab is exactly that moment, and the note was re-read before it was
+// overruled. What changed the answer:
+//
+//   * The expensive half of "documents to PDF" is not the file format, it is
+//     the LAYOUT — line breaking, pagination, lists, tables. Universal PDF
+//     already has that engine (`src/lib/markdownToPdf.ts`), and only its bottom
+//     inch touches pdf-lib: `widthOfTextAtSize`, `drawText`, `drawRectangle`,
+//     `drawLine`. Porting it needed those four primitives, not a library.
+//   * The base-14 fonts are IN every PDF reader. Using them means no font
+//     embedding, no subsetting, no CMap — the genuinely hard parts of a PDF
+//     library — in exchange for one table of glyph widths per font (`doc/
+//     metrics.ts`).
+//   * pdf-lib is ~380 KB gzipped. This app's pitch is that it is small and runs
+//     offline from the first visit, and a lazily-fetched library breaks the
+//     second half of that: your first document conversion would need the
+//     network. LAME and libFLAC are fetched on demand because there is no other
+//     way to have MP3 and FLAC at all. There is another way to have text.
+//
+// WHAT THIS DELIBERATELY DOES NOT DO
+// ----------------------------------
+//   * No font EMBEDDING, so only what WinAnsi can spell — Latin-1 plus the
+//     Windows-1252 extras (smart quotes, dashes, bullet, ellipsis, euro).
+//     Cyrillic, Greek, Hebrew, Arabic and CJK cannot be written at all. They
+//     are COUNTED and reported (`EncodeReport`) rather than silently replaced,
+//     because a document that comes back as ??? with no explanation is worse
+//     than one that refuses.
+//   * No reading or editing an existing PDF. That needs a parser, which is a
+//     different program. Universal PDF is the app for that.
+// ---------------------------------------------------------------------------
+import { FONT_METRICS } from './metrics.js';
+export function rgb(r, g, b) {
+    return { r, g, b };
+}
+// ── WinAnsi encoding ─────────────────────────────────────────────────────────
+// Codes 32–126 and 160–255 are Unicode's own values, so they need no table.
+// 128–159 is where Windows-1252 differs from Latin-1, and it is the useful
+// part: every smart quote, dash and bullet a word processor emits lives here.
+// Universal PDF's markdown writer flattens these to ASCII ('’' → "'") because
+// pdf-lib's standard fonts throw on them; we can spell them properly, so a
+// converted document keeps its typography instead of losing it on the way out.
+const WIN_ANSI_HIGH = {
+    '€': 0x80, // euro
+    '‚': 0x82, // single low quote
+    'ƒ': 0x83, // florin
+    '„': 0x84, // double low quote
+    '…': 0x85, // ellipsis
+    '†': 0x86, // dagger
+    '‡': 0x87, // double dagger
+    'ˆ': 0x88, // circumflex
+    '‰': 0x89, // per mille
+    'Š': 0x8a, // S caron
+    '‹': 0x8b, // single left guillemet
+    'Œ': 0x8c, // OE
+    'Ž': 0x8e, // Z caron
+    '‘': 0x91, // left single quote
+    '’': 0x92, // right single quote / apostrophe
+    '“': 0x93, // left double quote
+    '”': 0x94, // right double quote
+    '•': 0x95, // bullet
+    '–': 0x96, // en dash
+    '—': 0x97, // em dash
+    '˜': 0x98, // small tilde
+    '™': 0x99, // trademark
+    'š': 0x9a, // s caron
+    '›': 0x9b, // single right guillemet
+    'œ': 0x9c, // oe
+    'ž': 0x9e, // z caron
+    'Ÿ': 0x9f, // Y dieresis
+};
+/**
+ * Characters that have no WinAnsi code but a fair ASCII stand-in. Losing the
+ * shape of an arrow is a much smaller loss than losing the line it was on, so
+ * these substitute silently and are NOT counted as dropped.
+ */
+const SUBSTITUTES = {
+    '→': '->', '←': '<-', '↔': '<->',
+    '⇒': '=>', '⇐': '<=',
+    '≤': '<=', '≥': '>=', '≠': '!=',
+    '−': '-', ' ': ' ', '​': '', ' ': ' ', ' ': ' ',
+    '‑': '-', '‒': '-', '―': '-',
+    '′': "'", '″': '"',
+    '✓': 'v', '✔': 'v', '✗': 'x', '✘': 'x',
+    '●': '*', '○': 'o', '◦': 'o', '▪': '-', '▫': '-', '‣': '-', '⁃': '-',
+    '▶': '>', '◀': '<', '▸': '>',
+    '✅': '[ok]', '❌': '[x]',
+    '
': ' ', '
': ' ',
+    'fi': 'fi', 'fl': 'fl',
+};
+/**
+ * A JS string as WinAnsi bytes.
+ *
+ * Anything unrepresentable becomes '?' and is recorded on `report`, so the UI
+ * can say *which* characters went missing rather than leaving somebody to spot
+ * it themselves halfway down page four.
+ */
+export function toWinAnsi(text, report) {
+    const out = [];
+    for (const ch of text) {
+        const sub = SUBSTITUTES[ch];
+        if (sub !== undefined) {
+            for (const c of sub)
+                out.push(c.charCodeAt(0));
+            continue;
+        }
+        const code = ch.codePointAt(0);
+        if (code === 9) {
+            out.push(32);
+            continue;
+        } // tab — the layout owns indentation
+        if ((code >= 32 && code <= 126) || (code >= 160 && code <= 255)) {
+            out.push(code);
+            continue;
+        }
+        const high = WIN_ANSI_HIGH[ch];
+        if (high !== undefined) {
+            out.push(high);
+            continue;
+        }
+        report?.dropped.add(ch);
+        out.push(0x3f); // '?'
+    }
+    return out;
+}
+/** The width of a string at a given size, in points. */
+export function widthOfText(font, text, size) {
+    const metrics = FONT_METRICS[font];
+    let units = 0;
+    for (const code of toWinAnsi(text))
+        units += metrics.widths[code] ?? metrics.fallback;
+    return (units * size) / 1000;
+}
+// ── The document ─────────────────────────────────────────────────────────────
+/** The base-14 names, in the order font objects are written. */
+const FONT_NAMES = {
+    helv: 'Helvetica',
+    helvB: 'Helvetica-Bold',
+    helvI: 'Helvetica-Oblique',
+    helvBI: 'Helvetica-BoldOblique',
+    times: 'Times-Roman',
+    timesB: 'Times-Bold',
+    timesI: 'Times-Italic',
+    timesBI: 'Times-BoldItalic',
+    cour: 'Courier',
+    courB: 'Courier-Bold',
+};
+const FONT_IDS = Object.keys(FONT_NAMES);
+/**
+ * One page, accumulating content-stream operators.
+ *
+ * The op list is built as strings and joined at save time. A page of dense text
+ * is a few thousand short strings, which is nothing next to the image bytes
+ * beside it, and the alternative — encoding to bytes per op — makes every
+ * drawing call allocate.
+ */
+export class PdfPage {
+    // ⚠️ NOT `readonly` any more, and only `setSize` may change them. A page that
+    // has been drawn on cannot be resized — the content stream's coordinates are
+    // already in terms of the old box — which is why the caller in `write/pdf.ts`
+    // only ever calls it on an untouched page. Kept mutable for exactly that one
+    // case: a document whose FIRST section is landscape has to change the page
+    // the constructor already opened, or every landscape document gains a blank
+    // portrait sheet in front of it.
+    width;
+    height;
+    ops = [];
+    images = [];
+    links = [];
+    /** Fonts actually used, so a page's /Resources only names what it draws. */
+    usedFonts = new Set();
+    report;
+    constructor(width, height, report) {
+        this.width = width;
+        this.height = height;
+        this.report = report;
+    }
+    /**
+     * Change this page's paper size. See the note on `width`/`height` above:
+     * valid only before anything is drawn.
+     */
+    setSize(width, height) {
+        if (this.ops.length) {
+            throw new Error('A page that has been drawn on cannot be resized.');
+        }
+        this.width = width;
+        this.height = height;
+    }
+    /**
+     * `y` is the BASELINE, and the origin is bottom-left — PDF's own convention,
+     * kept rather than flipped so the ported layout engine reads the same in both
+     * apps and a bug in one is recognisable in the other.
+     */
+    drawText(text, opts) {
+        if (!text)
+            return;
+        this.usedFonts.add(opts.font);
+        const bytes = toWinAnsi(text, this.report);
+        this.ops.push('BT', `${num(opts.color.r)} ${num(opts.color.g)} ${num(opts.color.b)} rg`, `/${opts.font} ${num(opts.size)} Tf`, `${num(opts.x)} ${num(opts.y)} Td`, `${pdfStringFromBytes(bytes)} Tj`, 'ET');
+    }
+    drawRect(opts) {
+        if (opts.width <= 0 || opts.height <= 0)
+            return;
+        const rectOp = `${num(opts.x)} ${num(opts.y)} ${num(opts.width)} ${num(opts.height)} re`;
+        if (opts.color)
+            this.ops.push(`${num(opts.color.r)} ${num(opts.color.g)} ${num(opts.color.b)} rg`);
+        if (opts.borderColor) {
+            this.ops.push(`${num(opts.borderColor.r)} ${num(opts.borderColor.g)} ${num(opts.borderColor.b)} RG`, `${num(opts.borderWidth ?? 1)} w`);
+        }
+        this.ops.push(rectOp, opts.color && opts.borderColor ? 'B' : opts.color ? 'f' : 'S');
+    }
+    drawImage(image, opts) {
+        const index = this.images.length;
+        this.images.push(image);
+        this.ops.push('q', `${num(opts.width)} 0 0 ${num(opts.height)} ${num(opts.x)} ${num(opts.y)} cm`, `/Im${index} Do`, 'Q');
+    }
+    /**
+     * A clickable region. Kept separate from the drawing that makes it visible:
+     * a link's underline is content and its hotspot is an annotation, and the two
+     * are only in the same place because we put them there.
+     */
+    addLink(x1, y1, x2, y2, url) {
+        this.links.push({ rect: [x1, y1, x2, y2], url });
+    }
+    /** @internal */
+    build() {
+        return {
+            content: this.ops.join('\n') + '\n',
+            images: this.images,
+            links: this.links,
+            fonts: [...this.usedFonts],
+        };
+    }
+}
+export class PdfDocument {
+    pages = [];
+    /** Collected across every page, so the UI can report it once at the end. */
+    report = { dropped: new Set() };
+    addPage(width, height) {
+        const page = new PdfPage(width, height, this.report);
+        this.pages.push(page);
+        return page;
+    }
+    get pageCount() {
+        return this.pages.length;
+    }
+    /**
+     * Serialise. Object numbers are allocated as we go rather than computed from
+     * a formula — the old `3 + i * 3` arithmetic in `pdf.ts` was correct only
+     * while every page had exactly one image and no annotations, and a formula
+     * that has to be re-derived every time the object graph changes is a bug
+     * waiting for the next feature.
+     */
+    save(meta = {}) {
+        if (!this.pages.length)
+            throw new Error('A PDF needs at least one page.');
+        const chunks = [];
+        const offsets = [];
+        let length = 0;
+        let nextObject = 1;
+        const push = (data) => {
+            const bytes = typeof data === 'string' ? latin1(data) : data;
+            chunks.push(bytes);
+            length += bytes.length;
+        };
+        const alloc = () => nextObject++;
+        const begin = (n) => {
+            offsets[n] = length;
+            push(`${n} 0 obj\n`);
+        };
+        const end = () => push('endobj\n');
+        push('%PDF-1.4\n');
+        // A comment of high bytes — what tells a transfer that treats the file as
+        // text that it is binary. Every writer emits it; leaving it out breaks the
+        // file on exactly the systems that are hardest to debug.
+        push(new Uint8Array([0x25, 0xe2, 0xe3, 0xcf, 0xd3, 0x0a]));
+        const catalogId = alloc();
+        const pagesId = alloc();
+        // One font object per base-14 face, shared by every page. Ten small
+        // dictionaries is less machinery than tracking which page wants which.
+        const fontIds = new Map();
+        for (const id of FONT_IDS) {
+            const objectId = alloc();
+            fontIds.set(id, objectId);
+            begin(objectId);
+            push(`<< /Type /Font /Subtype /Type1 /BaseFont /${FONT_NAMES[id]} ` +
+                `/Encoding /WinAnsiEncoding >>\n`);
+            end();
+        }
+        const pageIds = [];
+        for (const page of this.pages) {
+            const built = page.build();
+            const pageId = alloc();
+            pageIds.push(pageId);
+            const contentId = alloc();
+            const imageIds = built.images.map(() => alloc());
+            const linkIds = built.links.map(() => alloc());
+            const fontResources = built.fonts
+                .map((f) => `/${f} ${fontIds.get(f)} 0 R`)
+                .join(' ');
+            const xobjectResources = imageIds
+                .map((id, i) => `/Im${i} ${id} 0 R`)
+                .join(' ');
+            const resources = `<< ${fontResources ? `/Font << ${fontResources} >> ` : ''}` +
+                `${xobjectResources ? `/XObject << ${xobjectResources} >> ` : ''}>>`;
+            begin(pageId);
+            push(`<< /Type /Page /Parent ${pagesId} 0 R ` +
+                `/MediaBox [0 0 ${num(page.width)} ${num(page.height)}] ` +
+                `/Resources ${resources} /Contents ${contentId} 0 R` +
+                (linkIds.length ? ` /Annots [${linkIds.map((id) => `${id} 0 R`).join(' ')}]` : '') +
+                ` >>\n`);
+            end();
+            const contentBytes = latin1(built.content);
+            begin(contentId);
+            push(`<< /Length ${contentBytes.length} >>\nstream\n`);
+            push(contentBytes);
+            push('\nendstream\n');
+            end();
+            built.images.forEach((image, i) => {
+                begin(imageIds[i]);
+                push(`<< /Type /XObject /Subtype /Image /Width ${image.width} /Height ${image.height} ` +
+                    `/ColorSpace /DeviceRGB /BitsPerComponent 8 /Filter /DCTDecode ` +
+                    `/Length ${image.jpeg.length} >>\nstream\n`);
+                push(image.jpeg);
+                push('\nendstream\n');
+                end();
+            });
+            built.links.forEach((link, i) => {
+                begin(linkIds[i]);
+                push(`<< /Type /Annot /Subtype /Link /Rect [${link.rect.map(num).join(' ')}] ` +
+                    `/Border [0 0 0] /A << /Type /Action /S /URI /URI ${pdfString(link.url)} >> >>\n`);
+                end();
+            });
+        }
+        begin(pagesId);
+        push(`<< /Type /Pages /Count ${pageIds.length} ` +
+            `/Kids [${pageIds.map((id) => `${id} 0 R`).join(' ')}] >>\n`);
+        end();
+        begin(catalogId);
+        push(`<< /Type /Catalog /Pages ${pagesId} 0 R >>\n`);
+        end();
+        const infoId = alloc();
+        begin(infoId);
+        push(`<< ${meta.title ? `/Title ${pdfString(meta.title)} ` : ''}` +
+            `${meta.author ? `/Author ${pdfString(meta.author)} ` : ''}` +
+            `/Producer ${pdfString(meta.producer ?? 'Universal Converter')} ` +
+            `/Creator ${pdfString(meta.producer ?? 'Universal Converter')} >>\n`);
+        end();
+        const xrefAt = length;
+        const count = nextObject;
+        push(`xref\n0 ${count}\n`);
+        push('0000000000 65535 f \n');
+        for (let n = 1; n < count; n += 1) {
+            push(`${String(offsets[n] ?? 0).padStart(10, '0')} 00000 n \n`);
+        }
+        push(`trailer\n<< /Size ${count} /Root ${catalogId} 0 R /Info ${infoId} 0 R >>\n` +
+            `startxref\n${xrefAt}\n%%EOF\n`);
+        return new Blob(chunks, { type: 'application/pdf' });
+    }
+}
+// ── Serialisation helpers ────────────────────────────────────────────────────
+/**
+ * Numbers in a content stream. Fixed to 2dp and stripped of a trailing '.00':
+ * a raw JS float can serialise as `1e-7`, which is not a PDF number and makes
+ * the whole stream unparseable from that operator on.
+ */
+function num(value) {
+    if (!Number.isFinite(value))
+        return '0';
+    const fixed = value.toFixed(2);
+    return fixed.replace(/\.?0+$/, '') || '0';
+}
+/** A PDF literal string from already-encoded WinAnsi bytes. */
+function pdfStringFromBytes(bytes) {
+    let out = '(';
+    for (const b of bytes) {
+        if (b === 0x28 || b === 0x29 || b === 0x5c)
+            out += '\\';
+        out += String.fromCharCode(b);
+    }
+    return out + ')';
+}
+function pdfString(text) {
+    return pdfStringFromBytes(toWinAnsi(text));
+}
+/** Latin-1 bytes. Every string that reaches this is already WinAnsi or ASCII. */
+function latin1(s) {
+    const out = new Uint8Array(s.length);
+    for (let i = 0; i < s.length; i += 1)
+        out[i] = s.charCodeAt(i) & 0xff;
+    return out;
+}
+//# sourceMappingURL=pdfcore.js.map

package/dist/pdfcore.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"pdfcore.js","sourceRoot":"","sources":["../src/pdfcore.ts"],"names":[],"mappings":"AAAA,8EAA8E;AAC9E,kEAAkE;AAClE,EAAE;AACF,iCAAiC;AACjC,iCAAiC;AACjC,4EAA4E;AAC5E,6EAA6E;AAC7E,2EAA2E;AAC3E,sCAAsC;AACtC,EAAE;AACF,6EAA6E;AAC7E,2EAA2E;AAC3E,gFAAgF;AAChF,8EAA8E;AAC9E,0EAA0E;AAC1E,0EAA0E;AAC1E,4EAA4E;AAC5E,0EAA0E;AAC1E,oBAAoB;AACpB,gFAAgF;AAChF,4EAA4E;AAC5E,yEAAyE;AACzE,gFAAgF;AAChF,0EAA0E;AAC1E,EAAE;AACF,qCAAqC;AACrC,qCAAqC;AACrC,2EAA2E;AAC3E,0EAA0E;AAC1E,6EAA6E;AAC7E,+EAA+E;AAC/E,6EAA6E;AAC7E,6BAA6B;AAC7B,6EAA6E;AAC7E,4DAA4D;AAC5D,8EAA8E;AAE9E,OAAO,EAAE,YAAY,EAAe,MAAM,cAAc,CAAA;AAUxD,MAAM,UAAU,GAAG,CAAC,CAAS,EAAE,CAAS,EAAE,CAAS;IACjD,OAAO,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,CAAA;AACpB,CAAC;AASD,gFAAgF;AAChF,4EAA4E;AAC5E,2EAA2E;AAC3E,8EAA8E;AAC9E,8EAA8E;AAC9E,2EAA2E;AAC3E,+EAA+E;AAC/E,MAAM,aAAa,GAA2B;IAC5C,GAAG,EAAE,IAAI,EAAE,OAAO;IAClB,GAAG,EAAE,IAAI,EAAE,mBAAmB;IAC9B,GAAG,EAAE,IAAI,EAAE,SAAS;IACpB,GAAG,EAAE,IAAI,EAAE,mBAAmB;IAC9B,GAAG,EAAE,IAAI,EAAE,WAAW;IACtB,GAAG,EAAE,IAAI,EAAE,SAAS;IACpB,GAAG,EAAE,IAAI,EAAE,gBAAgB;IAC3B,GAAG,EAAE,IAAI,EAAE,aAAa;IACxB,GAAG,EAAE,IAAI,EAAE,YAAY;IACvB,GAAG,EAAE,IAAI,EAAE,UAAU;IACrB,GAAG,EAAE,IAAI,EAAE,wBAAwB;IACnC,GAAG,EAAE,IAAI,EAAE,KAAK;IAChB,GAAG,EAAE,IAAI,EAAE,UAAU;IACrB,GAAG,EAAE,IAAI,EAAE,oBAAoB;IAC/B,GAAG,EAAE,IAAI,EAAE,kCAAkC;IAC7C,GAAG,EAAE,IAAI,EAAE,oBAAoB;IAC/B,GAAG,EAAE,IAAI,EAAE,qBAAqB;IAChC,GAAG,EAAE,IAAI,EAAE,SAAS;IACpB,GAAG,EAAE,IAAI,EAAE,UAAU;IACrB,GAAG,EAAE,IAAI,EAAE,UAAU;IACrB,GAAG,EAAE,IAAI,EAAE,cAAc;IACzB,GAAG,EAAE,IAAI,EAAE,YAAY;IACvB,GAAG,EAAE,IAAI,EAAE,UAAU;IACrB,GAAG,EAAE,IAAI,EAAE,yBAAyB;IACpC,GAAG,EAAE,IAAI,EAAE,KAAK;IAChB,GAAG,EAAE,IAAI,EAAE,UAAU;IACrB,GAAG,EAAE,IAAI,EAAE,aAAa;CACzB,CAAA;AAED;;;;GAIG;AACH,MAAM,WAAW,GAA2B;IAC1C,GAAG,EAAE,IAAI,EAAE,GAAG,EAAE,IAAI,EAAE,GAAG,EAAE,KAAK;IAChC,GAAG,EAAE,IAAI,EAAE,GAAG,EAAE,IAAI;IACpB,GAAG,EAAE,IAAI,EAAE,GAAG,EAAE,IAAI,EAAE,GAAG,EAAE,IAAI;IAC/B,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,EAAE,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG;IAC/C,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG;IAC5B,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG;IAClB,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG;IACtC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG;IACpE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG;IAC5B,GAAG,EAAE,MAAM,EAAE,GAAG,EAAE,KAAK;IACvB;CACD,EAAE,GAAG,EAAE;CACP,EAAE,GAAG;IACJ,GAAG,EAAE,IAAI,EAAE,GAAG,EAAE,IAAI;CACrB,CAAA;AAOD;;;;;;GAMG;AACH,MAAM,UAAU,SAAS,CAAC,IAAY,EAAE,MAAqB;IAC3D,MAAM,GAAG,GAAa,EAAE,CAAA;IACxB,KAAK,MAAM,EAAE,IAAI,IAAI,EAAE,CAAC;QACtB,MAAM,GAAG,GAAG,WAAW,CAAC,EAAE,CAAC,CAAA;QAC3B,IAAI,GAAG,KAAK,SAAS,EAAE,CAAC;YACtB,KAAK,MAAM,CAAC,IAAI,GAAG;gBAAE,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,CAAA;YAC9C,SAAQ;QACV,CAAC;QACD,MAAM,IAAI,GAAG,EAAE,CAAC,WAAW,CAAC,CAAC,CAAE,CAAA;QAC/B,IAAI,IAAI,KAAK,CAAC,EAAE,CAAC;YAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;YAAC,SAAQ;QAAC,CAAC,CAAC,oCAAoC;QAC/E,IAAI,CAAC,IAAI,IAAI,EAAE,IAAI,IAAI,IAAI,GAAG,CAAC,IAAI,CAAC,IAAI,IAAI,GAAG,IAAI,IAAI,IAAI,GAAG,CAAC,EAAE,CAAC;YAChE,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;YACd,SAAQ;QACV,CAAC;QACD,MAAM,IAAI,GAAG,aAAa,CAAC,EAAE,CAAC,CAAA;QAC9B,IAAI,IAAI,KAAK,SAAS,EAAE,CAAC;YACvB,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;YACd,SAAQ;QACV,CAAC;QACD,MAAM,EAAE,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC,CAAA;QACvB,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA,CAAC,MAAM;IACvB,CAAC;IACD,OAAO,GAAG,CAAA;AACZ,CAAC;AAED,wDAAwD;AACxD,MAAM,UAAU,WAAW,CAAC,IAAY,EAAE,IAAY,EAAE,IAAY;IAClE,MAAM,OAAO,GAAG,YAAY,CAAC,IAAI,CAAC,CAAA;IAClC,IAAI,KAAK,GAAG,CAAC,CAAA;IACb,KAAK,MAAM,IAAI,IAAI,SAAS,CAAC,IAAI,CAAC;QAAE,KAAK,IAAI,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,OAAO,CAAC,QAAQ,CAAA;IACrF,OAAO,CAAC,KAAK,GAAG,IAAI,CAAC,GAAG,IAAI,CAAA;AAC9B,CAAC;AAED,gFAAgF;AAEhF,gEAAgE;AAChE,MAAM,UAAU,GAA2B;IACzC,IAAI,EAAE,WAAW;IACjB,KAAK,EAAE,gBAAgB;IACvB,KAAK,EAAE,mBAAmB;IAC1B,MAAM,EAAE,uBAAuB;IAC/B,KAAK,EAAE,aAAa;IACpB,MAAM,EAAE,YAAY;IACpB,MAAM,EAAE,cAAc;IACtB,OAAO,EAAE,kBAAkB;IAC3B,IAAI,EAAE,SAAS;IACf,KAAK,EAAE,cAAc;CACtB,CAAA;AAED,MAAM,QAAQ,GAAG,MAAM,CAAC,IAAI,CAAC,UAAU,CAAa,CAAA;AAOpD;;;;;;;GAOG;AACH,MAAM,OAAO,OAAO;IAClB,8EAA8E;IAC9E,6EAA6E;IAC7E,8EAA8E;IAC9E,6EAA6E;IAC7E,2EAA2E;IAC3E,4EAA4E;IAC5E,iCAAiC;IACjC,KAAK,CAAQ;IACb,MAAM,CAAQ;IACN,GAAG,GAAa,EAAE,CAAA;IAClB,MAAM,GAAe,EAAE,CAAA;IACvB,KAAK,GAAgB,EAAE,CAAA;IAC/B,4EAA4E;IACpE,SAAS,GAAG,IAAI,GAAG,EAAU,CAAA;IAC7B,MAAM,CAAe;IAE7B,YAAY,KAAa,EAAE,MAAc,EAAE,MAAqB;QAC9D,IAAI,CAAC,KAAK,GAAG,KAAK,CAAA;QAClB,IAAI,CAAC,MAAM,GAAG,MAAM,CAAA;QACpB,IAAI,CAAC,MAAM,GAAG,MAAM,CAAA;IACtB,CAAC;IAED;;;OAGG;IACH,OAAO,CAAC,KAAa,EAAE,MAAc;QACnC,IAAI,IAAI,CAAC,GAAG,CAAC,MAAM,EAAE,CAAC;YACpB,MAAM,IAAI,KAAK,CAAC,kDAAkD,CAAC,CAAA;QACrE,CAAC;QACD,IAAI,CAAC,KAAK,GAAG,KAAK,CAAA;QAClB,IAAI,CAAC,MAAM,GAAG,MAAM,CAAA;IACtB,CAAC;IAED;;;;OAIG;IACH,QAAQ,CAAC,IAAY,EAAE,IAAsE;QAC3F,IAAI,CAAC,IAAI;YAAE,OAAM;QACjB,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;QAC7B,MAAM,KAAK,GAAG,SAAS,CAAC,IAAI,EAAE,IAAI,CAAC,MAAM,CAAC,CAAA;QAC1C,IAAI,CAAC,GAAG,CAAC,IAAI,CACX,IAAI,EACJ,GAAG,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,KAAK,EACnE,IAAI,IAAI,CAAC,IAAI,IAAI,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,EACpC,GAAG,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,KAAK,EAClC,GAAG,kBAAkB,CAAC,KAAK,CAAC,KAAK,EACjC,IAAI,CACL,CAAA;IACH,CAAC;IAED,QAAQ,CAAC,IAQR;QACC,IAAI,IAAI,CAAC,KAAK,IAAI,CAAC,IAAI,IAAI,CAAC,MAAM,IAAI,CAAC;YAAE,OAAM;QAC/C,MAAM,MAAM,GAAG,GAAG,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAA;QACxF,IAAI,IAAI,CAAC,KAAK;YAAE,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,CAAA;QAClG,IAAI,IAAI,CAAC,WAAW,EAAE,CAAC;YACrB,IAAI,CAAC,GAAG,CAAC,IAAI,CACX,GAAG,GAAG,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC,CAAC,KAAK,EACrF,GAAG,GAAG,CAAC,IAAI,CAAC,WAAW,IAAI,CAAC,CAAC,IAAI,CAClC,CAAA;QACH,CAAC;QACD,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,MAAM,EAAE,IAAI,CAAC,KAAK,IAAI,IAAI,CAAC,WAAW,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAA;IACtF,CAAC;IAED,SAAS,CAAC,KAAe,EAAE,IAA6D;QACtF,MAAM,KAAK,GAAG,IAAI,CAAC,MAAM,CAAC,MAAM,CAAA;QAChC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;QACvB,IAAI,CAAC,GAAG,CAAC,IAAI,CACX,GAAG,EACH,GAAG,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,QAAQ,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,KAAK,EAC7E,MAAM,KAAK,KAAK,EAChB,GAAG,CACJ,CAAA;IACH,CAAC;IAED;;;;OAIG;IACH,OAAO,CAAC,EAAU,EAAE,EAAU,EAAE,EAAU,EAAE,EAAU,EAAE,GAAW;QACjE,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,CAAC,EAAE,EAAE,EAAE,EAAE,EAAE,EAAE,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,CAAA;IAClD,CAAC;IAED,gBAAgB;IAChB,KAAK;QACH,OAAO;YACL,OAAO,EAAE,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,IAAI;YACnC,MAAM,EAAE,IAAI,CAAC,MAAM;YACnB,KAAK,EAAE,IAAI,CAAC,KAAK;YACjB,KAAK,EAAE,CAAC,GAAG,IAAI,CAAC,SAAS,CAAC;SAC3B,CAAA;IACH,CAAC;CACF;AASD,MAAM,OAAO,WAAW;IACd,KAAK,GAAc,EAAE,CAAA;IAC7B,4EAA4E;IACnE,MAAM,GAAiB,EAAE,OAAO,EAAE,IAAI,GAAG,EAAE,EAAE,CAAA;IAEtD,OAAO,CAAC,KAAa,EAAE,MAAc;QACnC,MAAM,IAAI,GAAG,IAAI,OAAO,CAAC,KAAK,EAAE,MAAM,EAAE,IAAI,CAAC,MAAM,CAAC,CAAA;QACpD,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;QACrB,OAAO,IAAI,CAAA;IACb,CAAC;IAED,IAAI,SAAS;QACX,OAAO,IAAI,CAAC,KAAK,CAAC,MAAM,CAAA;IAC1B,CAAC;IAED;;;;;;OAMG;IACH,IAAI,CAAC,OAAgB,EAAE;QACrB,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM;YAAE,MAAM,IAAI,KAAK,CAAC,gCAAgC,CAAC,CAAA;QAEzE,MAAM,MAAM,GAAiB,EAAE,CAAA;QAC/B,MAAM,OAAO,GAAa,EAAE,CAAA;QAC5B,IAAI,MAAM,GAAG,CAAC,CAAA;QACd,IAAI,UAAU,GAAG,CAAC,CAAA;QAElB,MAAM,IAAI,GAAG,CAAC,IAAyB,EAAE,EAAE;YACzC,MAAM,KAAK,GAAG,OAAO,IAAI,KAAK,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAA;YAC5D,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;YAClB,MAAM,IAAI,KAAK,CAAC,MAAM,CAAA;QACxB,CAAC,CAAA;QACD,MAAM,KAAK,GAAG,GAAG,EAAE,CAAC,UAAU,EAAE,CAAA;QAChC,MAAM,KAAK,GAAG,CAAC,CAAS,EAAE,EAAE;YAC1B,OAAO,CAAC,CAAC,CAAC,GAAG,MAAM,CAAA;YACnB,IAAI,CAAC,GAAG,CAAC,UAAU,CAAC,CAAA;QACtB,CAAC,CAAA;QACD,MAAM,GAAG,GAAG,GAAG,EAAE,CAAC,IAAI,CAAC,UAAU,CAAC,CAAA;QAElC,IAAI,CAAC,YAAY,CAAC,CAAA;QAClB,0EAA0E;QAC1E,2EAA2E;QAC3E,yDAAyD;QACzD,IAAI,CAAC,IAAI,UAAU,CAAC,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,CAAC,CAAC,CAAA;QAE1D,MAAM,SAAS,GAAG,KAAK,EAAE,CAAA;QACzB,MAAM,OAAO,GAAG,KAAK,EAAE,CAAA;QAEvB,oEAAoE;QACpE,uEAAuE;QACvE,MAAM,OAAO,GAAG,IAAI,GAAG,EAAkB,CAAA;QACzC,KAAK,MAAM,EAAE,IAAI,QAAQ,EAAE,CAAC;YAC1B,MAAM,QAAQ,GAAG,KAAK,EAAE,CAAA;YACxB,OAAO,CAAC,GAAG,CAAC,EAAE,EAAE,QAAQ,CAAC,CAAA;YACzB,KAAK,CAAC,QAAQ,CAAC,CAAA;YACf,IAAI,CACF,6CAA6C,UAAU,CAAC,EAAE,CAAC,GAAG;gBAC9D,iCAAiC,CAClC,CAAA;YACD,GAAG,EAAE,CAAA;QACP,CAAC;QAED,MAAM,OAAO,GAAa,EAAE,CAAA;QAE5B,KAAK,MAAM,IAAI,IAAI,IAAI,CAAC,KAAK,EAAE,CAAC;YAC9B,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,EAAE,CAAA;YAC1B,MAAM,MAAM,GAAG,KAAK,EAAE,CAAA;YACtB,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;YAEpB,MAAM,SAAS,GAAG,KAAK,EAAE,CAAA;YACzB,MAAM,QAAQ,GAAG,KAAK,CAAC,MAAM,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,KAAK,EAAE,CAAC,CAAA;YAChD,MAAM,OAAO,GAAG,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,KAAK,EAAE,CAAC,CAAA;YAE9C,MAAM,aAAa,GAAG,KAAK,CAAC,KAAK;iBAC9B,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC;iBACzC,IAAI,CAAC,GAAG,CAAC,CAAA;YACZ,MAAM,gBAAgB,GAAG,QAAQ;iBAC9B,GAAG,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,EAAE,CAAC,MAAM,CAAC,IAAI,EAAE,MAAM,CAAC;iBACnC,IAAI,CAAC,GAAG,CAAC,CAAA;YAEZ,MAAM,SAAS,GACb,MAAM,aAAa,CAAC,CAAC,CAAC,YAAY,aAAa,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE;gBAC5D,GAAG,gBAAgB,CAAC,CAAC,CAAC,eAAe,gBAAgB,MAAM,CAAC,CAAC,CAAC,EAAE,IAAI,CAAA;YAEtE,KAAK,CAAC,MAAM,CAAC,CAAA;YACb,IAAI,CACF,0BAA0B,OAAO,OAAO;gBACxC,kBAAkB,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI;gBACzD,cAAc,SAAS,cAAc,SAAS,MAAM;gBACpD,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,aAAa,OAAO,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,GAAG,EAAE,MAAM,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC;gBAClF,OAAO,CACR,CAAA;YACD,GAAG,EAAE,CAAA;YAEL,MAAM,YAAY,GAAG,MAAM,CAAC,KAAK,CAAC,OAAO,CAAC,CAAA;YAC1C,KAAK,CAAC,SAAS,CAAC,CAAA;YAChB,IAAI,CAAC,cAAc,YAAY,CAAC,MAAM,eAAe,CAAC,CAAA;YACtD,IAAI,CAAC,YAAY,CAAC,CAAA;YAClB,IAAI,CAAC,eAAe,CAAC,CAAA;YACrB,GAAG,EAAE,CAAA;YAEL,KAAK,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,CAAC,EAAE,EAAE;gBAChC,KAAK,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAA;gBAClB,IAAI,CACF,4CAA4C,KAAK,CAAC,KAAK,YAAY,KAAK,CAAC,MAAM,GAAG;oBAClF,gEAAgE;oBAChE,WAAW,KAAK,CAAC,IAAI,CAAC,MAAM,eAAe,CAC5C,CAAA;gBACD,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAA;gBAChB,IAAI,CAAC,eAAe,CAAC,CAAA;gBACrB,GAAG,EAAE,CAAA;YACP,CAAC,CAAC,CAAA;YAEF,KAAK,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,CAAC,EAAE,EAAE;gBAC9B,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,CAAA;gBACjB,IAAI,CACF,yCAAyC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI;oBACzE,oDAAoD,SAAS,CAAC,IAAI,CAAC,GAAG,CAAC,UAAU,CAClF,CAAA;gBACD,GAAG,EAAE,CAAA;YACP,CAAC,CAAC,CAAA;QACJ,CAAC;QAED,KAAK,CAAC,OAAO,CAAC,CAAA;QACd,IAAI,CACF,0BAA0B,OAAO,CAAC,MAAM,GAAG;YAC3C,UAAU,OAAO,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,GAAG,EAAE,MAAM,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,QAAQ,CAC7D,CAAA;QACD,GAAG,EAAE,CAAA;QAEL,KAAK,CAAC,SAAS,CAAC,CAAA;QAChB,IAAI,CAAC,4BAA4B,OAAO,WAAW,CAAC,CAAA;QACpD,GAAG,EAAE,CAAA;QAEL,MAAM,MAAM,GAAG,KAAK,EAAE,CAAA;QACtB,KAAK,CAAC,MAAM,CAAC,CAAA;QACb,IAAI,CACF,MAAM,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,UAAU,SAAS,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE;YAC5D,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,WAAW,SAAS,CAAC,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE;YAC5D,aAAa,SAAS,CAAC,IAAI,CAAC,QAAQ,IAAI,qBAAqB,CAAC,GAAG;YACjE,YAAY,SAAS,CAAC,IAAI,CAAC,QAAQ,IAAI,qBAAqB,CAAC,OAAO,CACrE,CAAA;QACD,GAAG,EAAE,CAAA;QAEL,MAAM,MAAM,GAAG,MAAM,CAAA;QACrB,MAAM,KAAK,GAAG,UAAU,CAAA;QACxB,IAAI,CAAC,WAAW,KAAK,IAAI,CAAC,CAAA;QAC1B,IAAI,CAAC,uBAAuB,CAAC,CAAA;QAC7B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;YAClC,IAAI,CAAC,GAAG,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,QAAQ,CAAC,EAAE,EAAE,GAAG,CAAC,aAAa,CAAC,CAAA;QACjE,CAAC;QACD,IAAI,CACF,qBAAqB,KAAK,UAAU,SAAS,cAAc,MAAM,WAAW;YAC5E,cAAc,MAAM,WAAW,CAChC,CAAA;QAED,OAAO,IAAI,IAAI,CAAC,MAAoB,EAAE,EAAE,IAAI,EAAE,iBAAiB,EAAE,CAAC,CAAA;IACpE,CAAC;CACF;AAED,gFAAgF;AAEhF;;;;GAIG;AACH,SAAS,GAAG,CAAC,KAAa;IACxB,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,KAAK,CAAC;QAAE,OAAO,GAAG,CAAA;IACvC,MAAM,KAAK,GAAG,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,CAAA;IAC9B,OAAO,KAAK,CAAC,OAAO,CAAC,QAAQ,EAAE,EAAE,CAAC,IAAI,GAAG,CAAA;AAC3C,CAAC;AAED,+DAA+D;AAC/D,SAAS,kBAAkB,CAAC,KAAe;IACzC,IAAI,GAAG,GAAG,GAAG,CAAA;IACb,KAAK,MAAM,CAAC,IAAI,KAAK,EAAE,CAAC;QACtB,IAAI,CAAC,KAAK,IAAI,IAAI,CAAC,KAAK,IAAI,IAAI,CAAC,KAAK,IAAI;YAAE,GAAG,IAAI,IAAI,CAAA;QACvD,GAAG,IAAI,MAAM,CAAC,YAAY,CAAC,CAAC,CAAC,CAAA;IAC/B,CAAC;IACD,OAAO,GAAG,GAAG,GAAG,CAAA;AAClB,CAAC;AAED,SAAS,SAAS,CAAC,IAAY;IAC7B,OAAO,kBAAkB,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC,CAAA;AAC5C,CAAC;AAED,iFAAiF;AACjF,SAAS,MAAM,CAAC,CAAS;IACvB,MAAM,GAAG,GAAG,IAAI,UAAU,CAAC,CAAC,CAAC,MAAM,CAAC,CAAA;IACpC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,CAAC,IAAI,CAAC;QAAE,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,GAAG,IAAI,CAAA;IACrE,OAAO,GAAG,CAAA;AACZ,CAAC"}

package/dist/read/csv.d.ts

@@ -0,0 +1,17 @@
+import { type Grid, type RichDoc } from '../model.js';
+export interface CsvResult {
+    doc: RichDoc;
+    grid: Grid;
+}
+export declare function readCsv(file: File): Promise<CsvResult>;
+export declare function parseCsvFile(source: string, filename?: string): CsvResult;
+/**
+ * Split a CSV into rows and fields.
+ *
+ * One pass, one state flag. `inQuotes` decides what every character means, and
+ * the doubled-quote rule is handled by peeking one ahead rather than by a
+ * second flag — inside quotes, `""` is a literal quote and a lone `"` ends the
+ * field.
+ */
+export declare function parseCsv(text: string, delimiter: string): string[][];
+//# sourceMappingURL=csv.d.ts.map

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

@@ -0,0 +1 @@
+{"version":3,"file":"csv.d.ts","sourceRoot":"","sources":["../../src/read/csv.ts"],"names":[],"mappings":"AAcA,OAAO,EAAoB,KAAK,IAAI,EAAE,KAAK,OAAO,EAAY,MAAM,aAAa,CAAA;AAEjF,MAAM,WAAW,SAAS;IACxB,GAAG,EAAE,OAAO,CAAA;IACZ,IAAI,EAAE,IAAI,CAAA;CACX;AAED,wBAAsB,OAAO,CAAC,IAAI,EAAE,IAAI,GAAG,OAAO,CAAC,SAAS,CAAC,CAE5D;AAED,wBAAgB,YAAY,CAAC,MAAM,EAAE,MAAM,EAAE,QAAQ,CAAC,EAAE,MAAM,GAAG,SAAS,CA0BzE;AAED;;;;;;;GAOG;AACH,wBAAgB,QAAQ,CAAC,IAAI,EAAE,MAAM,EAAE,SAAS,EAAE,MAAM,GAAG,MAAM,EAAE,EAAE,CAiDpE"}

package/dist/read/csv.js

@@ -0,0 +1,167 @@
+// ---------------------------------------------------------------------------
+// CSV → Grid (and a RichDoc table).
+//
+// A real RFC 4180 parser rather than `split(',')`, because the split version
+// works on every file you test it with and fails on the first real one: a
+// quoted field can contain the delimiter, a newline, and — the part that trips
+// most hand-rolled parsers — an escaped quote, written as TWO quotes inside a
+// quoted field. `"He said ""no"", then left"` is one field.
+//
+// The delimiter is SNIFFED rather than assumed. A "CSV" exported from Excel in
+// most of Europe is semicolon-separated, because the comma is the decimal point
+// there, and a tab-separated file is routinely named .csv too.
+// ---------------------------------------------------------------------------
+import { tidy } from '../model.js';
+export async function readCsv(file) {
+    return parseCsvFile(await file.text(), file.name);
+}
+export function parseCsvFile(source, filename) {
+    const text = source.replace(/^/, ''); // strip a UTF-8 BOM from Excel
+    const delimiter = sniffDelimiter(text);
+    const rows = parseCsv(text, delimiter);
+    if (!rows.length)
+        throw new Error('That CSV has no rows in it.');
+    const hasHeader = looksLikeHeader(rows);
+    const header = hasHeader ? rows[0] : [];
+    const body = hasHeader ? rows.slice(1) : rows;
+    const grid = { header, rows: body };
+    const blocks = [
+        {
+            kind: 'table',
+            header: hasHeader ? header.map((cell) => [{ text: cell }]) : null,
+            rows: body.map((row) => row.map((cell) => [{ text: cell }])),
+        },
+    ];
+    const doc = tidy({
+        blocks,
+        notices: [],
+        title: filename?.replace(/\.(csv|tsv|tab)$/i, ''),
+    });
+    return { doc, grid };
+}
+/**
+ * Split a CSV into rows and fields.
+ *
+ * One pass, one state flag. `inQuotes` decides what every character means, and
+ * the doubled-quote rule is handled by peeking one ahead rather than by a
+ * second flag — inside quotes, `""` is a literal quote and a lone `"` ends the
+ * field.
+ */
+export function parseCsv(text, delimiter) {
+    const rows = [];
+    let row = [];
+    let field = '';
+    let inQuotes = false;
+    let i = 0;
+    while (i < text.length) {
+        const ch = text[i];
+        if (inQuotes) {
+            if (ch === '"') {
+                if (text[i + 1] === '"') {
+                    field += '"';
+                    i += 2;
+                    continue;
+                }
+                inQuotes = false;
+                i += 1;
+                continue;
+            }
+            field += ch;
+            i += 1;
+            continue;
+        }
+        if (ch === '"' && field === '') {
+            inQuotes = true;
+            i += 1;
+            continue;
+        }
+        if (ch === delimiter) {
+            row.push(field);
+            field = '';
+            i += 1;
+            continue;
+        }
+        if (ch === '\r') {
+            i += 1;
+            continue;
+        }
+        if (ch === '\n') {
+            row.push(field);
+            rows.push(row);
+            row = [];
+            field = '';
+            i += 1;
+            continue;
+        }
+        field += ch;
+        i += 1;
+    }
+    if (field !== '' || row.length) {
+        row.push(field);
+        rows.push(row);
+    }
+    // A trailing newline gives a final row of one empty field. Dropping it is
+    // right; dropping a genuinely empty row in the MIDDLE would not be, so this
+    // only looks at the last.
+    const last = rows[rows.length - 1];
+    if (last && last.length === 1 && last[0] === '')
+        rows.pop();
+    return rows;
+}
+/**
+ * Which delimiter is this file using?
+ *
+ * Counted over the first few lines OUTSIDE quotes — a comma inside a quoted
+ * address would otherwise outvote the semicolons actually separating the
+ * fields. Whichever candidate divides the file into the most consistent number
+ * of columns wins, which beats a raw count when one character is simply common
+ * in the prose.
+ */
+function sniffDelimiter(text) {
+    const sample = text.split('\n').slice(0, 20).join('\n');
+    let best = ',';
+    let bestScore = -1;
+    for (const candidate of [',', ';', '\t', '|']) {
+        const rows = parseCsv(sample, candidate).filter((r) => r.length > 0);
+        if (rows.length < 1)
+            continue;
+        const columns = rows[0].length;
+        if (columns < 2)
+            continue;
+        const consistent = rows.filter((r) => r.length === columns).length / rows.length;
+        // Columns break ties: two candidates can both split every row evenly, and
+        // the one producing more fields is the one actually delimiting.
+        const score = consistent * 100 + Math.min(columns, 50);
+        if (score > bestScore) {
+            bestScore = score;
+            best = candidate;
+        }
+    }
+    return best;
+}
+/**
+ * Is row 0 a header?
+ *
+ * The signal that works: a header's cells are non-empty text, and the rows
+ * below it contain numbers or dates where the header has words. A file whose
+ * first row looks exactly like every other row probably has no header.
+ */
+function looksLikeHeader(rows) {
+    if (rows.length < 2)
+        return true;
+    const first = rows[0];
+    if (first.some((c) => !c.trim()))
+        return false;
+    if (first.every((c) => isNumeric(c)))
+        return false;
+    const numericBelow = rows.slice(1, 20).some((row) => row.some((c) => isNumeric(c)));
+    const numericInHeader = first.some((c) => isNumeric(c));
+    return numericBelow ? !numericInHeader : true;
+}
+function isNumeric(value) {
+    const trimmed = value.trim();
+    if (!trimmed)
+        return false;
+    return /^-?[\d.,]+%?$/.test(trimmed) && /\d/.test(trimmed);
+}
+//# sourceMappingURL=csv.js.map

package/dist/read/csv.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"csv.js","sourceRoot":"","sources":["../../src/read/csv.ts"],"names":[],"mappings":"AAAA,8EAA8E;AAC9E,oCAAoC;AACpC,EAAE;AACF,6EAA6E;AAC7E,0EAA0E;AAC1E,+EAA+E;AAC/E,8EAA8E;AAC9E,4DAA4D;AAC5D,EAAE;AACF,+EAA+E;AAC/E,gFAAgF;AAChF,+DAA+D;AAC/D,8EAA8E;AAE9E,OAAO,EAAE,IAAI,EAAiD,MAAM,aAAa,CAAA;AAOjF,MAAM,CAAC,KAAK,UAAU,OAAO,CAAC,IAAU;IACtC,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,IAAI,GAAG,MAAM,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,CAAA,CAAC,+BAA+B;IACrE,MAAM,SAAS,GAAG,cAAc,CAAC,IAAI,CAAC,CAAA;IACtC,MAAM,IAAI,GAAG,QAAQ,CAAC,IAAI,EAAE,SAAS,CAAC,CAAA;IAEtC,IAAI,CAAC,IAAI,CAAC,MAAM;QAAE,MAAM,IAAI,KAAK,CAAC,6BAA6B,CAAC,CAAA;IAEhE,MAAM,SAAS,GAAG,eAAe,CAAC,IAAI,CAAC,CAAA;IACvC,MAAM,MAAM,GAAG,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAA;IACvC,MAAM,IAAI,GAAG,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAA;IAC7C,MAAM,IAAI,GAAS,EAAE,MAAM,EAAE,IAAI,EAAE,IAAI,EAAE,CAAA;IAEzC,MAAM,MAAM,GAAY;QACtB;YACE,IAAI,EAAE,OAAO;YACb,MAAM,EAAE,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,IAAI,EAAS,EAAE,CAAC,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI;YACxE,IAAI,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,IAAI,EAAS,EAAE,CAAC,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC;SACpE;KACF,CAAA;IAED,MAAM,GAAG,GAAG,IAAI,CAAC;QACf,MAAM;QACN,OAAO,EAAE,EAAE;QACX,KAAK,EAAE,QAAQ,EAAE,OAAO,CAAC,mBAAmB,EAAE,EAAE,CAAC;KAClD,CAAC,CAAA;IACF,OAAO,EAAE,GAAG,EAAE,IAAI,EAAE,CAAA;AACtB,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,QAAQ,CAAC,IAAY,EAAE,SAAiB;IACtD,MAAM,IAAI,GAAe,EAAE,CAAA;IAC3B,IAAI,GAAG,GAAa,EAAE,CAAA;IACtB,IAAI,KAAK,GAAG,EAAE,CAAA;IACd,IAAI,QAAQ,GAAG,KAAK,CAAA;IACpB,IAAI,CAAC,GAAG,CAAC,CAAA;IAET,OAAO,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC;QACvB,MAAM,EAAE,GAAG,IAAI,CAAC,CAAC,CAAC,CAAA;QAElB,IAAI,QAAQ,EAAE,CAAC;YACb,IAAI,EAAE,KAAK,GAAG,EAAE,CAAC;gBACf,IAAI,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,GAAG,EAAE,CAAC;oBAAC,KAAK,IAAI,GAAG,CAAC;oBAAC,CAAC,IAAI,CAAC,CAAC;oBAAC,SAAQ;gBAAC,CAAC;gBAC3D,QAAQ,GAAG,KAAK,CAAA;gBAChB,CAAC,IAAI,CAAC,CAAA;gBACN,SAAQ;YACV,CAAC;YACD,KAAK,IAAI,EAAE,CAAA;YACX,CAAC,IAAI,CAAC,CAAA;YACN,SAAQ;QACV,CAAC;QAED,IAAI,EAAE,KAAK,GAAG,IAAI,KAAK,KAAK,EAAE,EAAE,CAAC;YAAC,QAAQ,GAAG,IAAI,CAAC;YAAC,CAAC,IAAI,CAAC,CAAC;YAAC,SAAQ;QAAC,CAAC;QACrE,IAAI,EAAE,KAAK,SAAS,EAAE,CAAC;YAAC,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YAAC,KAAK,GAAG,EAAE,CAAC;YAAC,CAAC,IAAI,CAAC,CAAC;YAAC,SAAQ;QAAC,CAAC;QACvE,IAAI,EAAE,KAAK,IAAI,EAAE,CAAC;YAAC,CAAC,IAAI,CAAC,CAAC;YAAC,SAAQ;QAAC,CAAC;QACrC,IAAI,EAAE,KAAK,IAAI,EAAE,CAAC;YAChB,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;YACf,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;YACd,GAAG,GAAG,EAAE,CAAA;YACR,KAAK,GAAG,EAAE,CAAA;YACV,CAAC,IAAI,CAAC,CAAA;YACN,SAAQ;QACV,CAAC;QACD,KAAK,IAAI,EAAE,CAAA;QACX,CAAC,IAAI,CAAC,CAAA;IACR,CAAC;IAED,IAAI,KAAK,KAAK,EAAE,IAAI,GAAG,CAAC,MAAM,EAAE,CAAC;QAC/B,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;QACf,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;IAChB,CAAC;IAED,0EAA0E;IAC1E,4EAA4E;IAC5E,0BAA0B;IAC1B,MAAM,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAA;IAClC,IAAI,IAAI,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC,IAAI,IAAI,CAAC,CAAC,CAAC,KAAK,EAAE;QAAE,IAAI,CAAC,GAAG,EAAE,CAAA;IAE3D,OAAO,IAAI,CAAA;AACb,CAAC;AAED;;;;;;;;GAQG;AACH,SAAS,cAAc,CAAC,IAAY;IAClC,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;IACvD,IAAI,IAAI,GAAG,GAAG,CAAA;IACd,IAAI,SAAS,GAAG,CAAC,CAAC,CAAA;IAElB,KAAK,MAAM,SAAS,IAAI,CAAC,GAAG,EAAE,GAAG,EAAE,IAAI,EAAE,GAAG,CAAC,EAAE,CAAC;QAC9C,MAAM,IAAI,GAAG,QAAQ,CAAC,MAAM,EAAE,SAAS,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAA;QACpE,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC;YAAE,SAAQ;QAC7B,MAAM,OAAO,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,MAAM,CAAA;QAC9B,IAAI,OAAO,GAAG,CAAC;YAAE,SAAQ;QACzB,MAAM,UAAU,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,OAAO,CAAC,CAAC,MAAM,GAAG,IAAI,CAAC,MAAM,CAAA;QAChF,0EAA0E;QAC1E,gEAAgE;QAChE,MAAM,KAAK,GAAG,UAAU,GAAG,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,OAAO,EAAE,EAAE,CAAC,CAAA;QACtD,IAAI,KAAK,GAAG,SAAS,EAAE,CAAC;YACtB,SAAS,GAAG,KAAK,CAAA;YACjB,IAAI,GAAG,SAAS,CAAA;QAClB,CAAC;IACH,CAAC;IACD,OAAO,IAAI,CAAA;AACb,CAAC;AAED;;;;;;GAMG;AACH,SAAS,eAAe,CAAC,IAAyB;IAChD,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC;QAAE,OAAO,IAAI,CAAA;IAChC,MAAM,KAAK,GAAG,IAAI,CAAC,CAAC,CAAC,CAAA;IACrB,IAAI,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QAAE,OAAO,KAAK,CAAA;IAC9C,IAAI,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC;QAAE,OAAO,KAAK,CAAA;IAElD,MAAM,YAAY,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,CAAA;IACnF,MAAM,eAAe,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,CAAA;IACvD,OAAO,YAAY,CAAC,CAAC,CAAC,CAAC,eAAe,CAAC,CAAC,CAAC,IAAI,CAAA;AAC/C,CAAC;AAED,SAAS,SAAS,CAAC,KAAa;IAC9B,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,EAAE,CAAA;IAC5B,IAAI,CAAC,OAAO;QAAE,OAAO,KAAK,CAAA;IAC1B,OAAO,eAAe,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;AAC5D,CAAC"}

package/dist/read/doc.d.ts

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

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

@@ -0,0 +1 @@
+{"version":3,"file":"doc.d.ts","sourceRoot":"","sources":["../../src/read/doc.ts"],"names":[],"mappings":"AAwBA,OAAO,EAA+B,KAAK,OAAO,EAAE,MAAM,aAAa,CAAA;AAiBvE,wBAAsB,OAAO,CAAC,IAAI,EAAE,IAAI,GAAG,OAAO,CAAC,OAAO,CAAC,CAmD1D"}