@nac3/forge-cli 0.2.0-alpha.1

package/dist/reader/parsers/epub.js

@@ -0,0 +1,265 @@
+/**
+ * EPUB parser for the document reader (V1.21 of HITO 1).
+ *
+ * Hand-rolled. No npm dep -- uses node:zlib for deflate.
+ * EPUB is a ZIP containing:
+ *
+ *   - mimetype file ("application/epub+zip").
+ *   - META-INF/container.xml -> points to the OPF.
+ *   - The OPF (XML) -> declares <manifest> (file list) +
+ *     <spine> (reading order) + <metadata>.
+ *   - XHTML files in the spine carry the actual content.
+ *
+ * Pipeline:
+ *   1. extractZipFiles(buffer) -> Map<path, Buffer>.
+ *   2. read container.xml -> resolve OPF path.
+ *   3. parse OPF -> manifest map (id -> href) + spine order +
+ *      title from <dc:title>.
+ *   4. concatenate the XHTML files in spine order.
+ *   5. parseHtml on the concatenated stream.
+ *   6. override title from OPF metadata.
+ *
+ * The XHTML reuse means images / tables / code blocks render
+ * with the same shape as native HTML docs. CSS is ignored
+ * (no styling matters for voice readback).
+ *
+ * Pathological inputs (corrupt ZIP, missing container.xml,
+ * malformed OPF) throw a loud Error so the caller surfaces
+ * the failure -- never silently returns an empty document.
+ *
+ * ASCII-only.
+ */
+import zlib from 'node:zlib';
+import { parseHtml } from './html.js';
+const EOCD_SIG = 0x06054b50;
+const CD_SIG = 0x02014b50;
+const LFH_SIG = 0x04034b50;
+function findEocd(buf) {
+    /* EOCD record is at most 65557 bytes from the end (max
+       comment length is 65535). Scan backwards. */
+    const max = Math.min(buf.length, 65557);
+    const start = buf.length - max;
+    for (let i = buf.length - 22; i >= start; i -= 1) {
+        if (buf.readUInt32LE(i) === EOCD_SIG)
+            return i;
+    }
+    throw new Error('epub: not a valid ZIP archive (EOCD signature not found)');
+}
+function readCentralDir(buf) {
+    const eocd = findEocd(buf);
+    const totalEntries = buf.readUInt16LE(eocd + 10);
+    const cdSize = buf.readUInt32LE(eocd + 12);
+    const cdOffset = buf.readUInt32LE(eocd + 16);
+    const entries = [];
+    let p = cdOffset;
+    const cdEnd = cdOffset + cdSize;
+    for (let n = 0; n < totalEntries && p + 46 <= cdEnd; n += 1) {
+        if (buf.readUInt32LE(p) !== CD_SIG) {
+            throw new Error('epub: malformed central directory at offset ' + p);
+        }
+        const method = buf.readUInt16LE(p + 10);
+        const compressedSize = buf.readUInt32LE(p + 20);
+        const uncompressedSize = buf.readUInt32LE(p + 24);
+        const fileNameLen = buf.readUInt16LE(p + 28);
+        const extraLen = buf.readUInt16LE(p + 30);
+        const commentLen = buf.readUInt16LE(p + 32);
+        const localHeaderOffset = buf.readUInt32LE(p + 42);
+        const name = buf.toString('utf-8', p + 46, p + 46 + fileNameLen);
+        entries.push({ name, method, compressedSize, uncompressedSize, localHeaderOffset });
+        p += 46 + fileNameLen + extraLen + commentLen;
+    }
+    return entries;
+}
+function readLocalFile(buf, entry) {
+    const p = entry.localHeaderOffset;
+    if (buf.readUInt32LE(p) !== LFH_SIG) {
+        throw new Error('epub: malformed local file header for ' + entry.name);
+    }
+    const fileNameLen = buf.readUInt16LE(p + 26);
+    const extraLen = buf.readUInt16LE(p + 28);
+    const dataStart = p + 30 + fileNameLen + extraLen;
+    const dataEnd = dataStart + entry.compressedSize;
+    const compressed = buf.subarray(dataStart, dataEnd);
+    if (entry.method === 0) {
+        return Buffer.from(compressed);
+    }
+    if (entry.method === 8) {
+        return zlib.inflateRawSync(compressed);
+    }
+    throw new Error('epub: unsupported compression method ' + entry.method + ' for ' + entry.name);
+}
+function extractZipFiles(buf) {
+    const out = new Map();
+    for (const entry of readCentralDir(buf)) {
+        /* Skip directory entries (trailing slash, zero size). */
+        if (entry.name.endsWith('/'))
+            continue;
+        try {
+            out.set(entry.name, readLocalFile(buf, entry));
+        }
+        catch (err) {
+            /* Skip unreadable entries -- but the must-haves throw
+               downstream when missing. */
+        }
+    }
+    return out;
+}
+/* ----- minimal XML helpers ----- */
+function decodeEntities(s) {
+    return s
+        .replace(/&#x([0-9a-fA-F]+);/g, (_, hex) => {
+        try {
+            return String.fromCodePoint(parseInt(hex, 16));
+        }
+        catch {
+            return '';
+        }
+    })
+        .replace(/&#(\d+);/g, (_, dec) => {
+        try {
+            return String.fromCodePoint(parseInt(dec, 10));
+        }
+        catch {
+            return '';
+        }
+    })
+        .replace(/&amp;/g, '&')
+        .replace(/&lt;/g, '<')
+        .replace(/&gt;/g, '>')
+        .replace(/&quot;/g, '"')
+        .replace(/&apos;/g, "'");
+}
+function xmlAttr(tag, name) {
+    const re = new RegExp('\\b' + name + '\\s*=\\s*(?:"([^"]*)"|\'([^\']*)\')', 'i');
+    const m = tag.match(re);
+    if (!m)
+        return null;
+    return decodeEntities(m[1] ?? m[2] ?? '');
+}
+function firstTag(xml, tagName) {
+    const re = new RegExp('<\\s*' + tagName + '\\b[^>]*>', 'i');
+    const m = xml.match(re);
+    return m ? m[0] : null;
+}
+function allTags(xml, tagName) {
+    const re = new RegExp('<\\s*' + tagName + '\\b[^>]*>', 'gi');
+    return xml.match(re) || [];
+}
+function tagInnerText(xml, tagName) {
+    const re = new RegExp('<\\s*' + tagName + '\\b[^>]*>([\\s\\S]*?)<\\s*/\\s*' + tagName + '\\s*>', 'i');
+    const m = xml.match(re);
+    if (!m || !m[1])
+        return null;
+    return decodeEntities(m[1].replace(/<[^>]+>/g, '').trim());
+}
+/* ----- EPUB-specific resolution ----- */
+function joinZipPath(base, rel) {
+    /* Strip filename from base. */
+    const idx = base.lastIndexOf('/');
+    const baseDir = idx >= 0 ? base.slice(0, idx + 1) : '';
+    /* Handle "../" / "./" / absolute references. */
+    const segs = (baseDir + rel).split('/');
+    const stack = [];
+    for (const seg of segs) {
+        if (!seg || seg === '.')
+            continue;
+        if (seg === '..') {
+            if (stack.length > 0)
+                stack.pop();
+            continue;
+        }
+        stack.push(seg);
+    }
+    return stack.join('/');
+}
+function resolveOpfPath(containerXml) {
+    const tag = firstTag(containerXml, 'rootfile');
+    if (!tag)
+        throw new Error('epub: container.xml missing <rootfile>');
+    const path = xmlAttr(tag, 'full-path');
+    if (!path)
+        throw new Error('epub: <rootfile> missing full-path attribute');
+    return path;
+}
+function parseOpf(opfXml) {
+    const title = tagInnerText(opfXml, 'dc:title') || tagInnerText(opfXml, 'title') || '';
+    const language = tagInnerText(opfXml, 'dc:language') || undefined;
+    /* Manifest. */
+    const manifest = new Map();
+    /* Pull only the <item ...> tags inside <manifest>. */
+    const manifestMatch = opfXml.match(/<manifest[^>]*>([\s\S]*?)<\/manifest\s*>/i);
+    if (manifestMatch && manifestMatch[1]) {
+        const items = allTags(manifestMatch[1], 'item');
+        for (const it of items) {
+            const id = xmlAttr(it, 'id');
+            const href = xmlAttr(it, 'href');
+            if (id && href)
+                manifest.set(id, href);
+        }
+    }
+    /* Spine. */
+    const spine = [];
+    const spineMatch = opfXml.match(/<spine[^>]*>([\s\S]*?)<\/spine\s*>/i);
+    if (spineMatch && spineMatch[1]) {
+        const items = allTags(spineMatch[1], 'itemref');
+        for (const it of items) {
+            const idref = xmlAttr(it, 'idref');
+            if (idref)
+                spine.push(idref);
+        }
+    }
+    const desc = { title, manifest, spine };
+    if (language)
+        desc.language = language;
+    return desc;
+}
+/* ----- main ----- */
+export function parseEpub(opts) {
+    const files = extractZipFiles(opts.buffer);
+    const containerBuf = files.get('META-INF/container.xml');
+    if (!containerBuf) {
+        throw new Error('epub: META-INF/container.xml not found in archive');
+    }
+    const opfPath = resolveOpfPath(containerBuf.toString('utf-8'));
+    const opfBuf = files.get(opfPath);
+    if (!opfBuf) {
+        throw new Error('epub: OPF file "' + opfPath + '" not found in archive');
+    }
+    const opf = parseOpf(opfBuf.toString('utf-8'));
+    /* Concatenate the spine XHTML files. Wrap each one in a
+       <section> so headings from later files do not merge with
+       the previous file's content. */
+    const parts = [];
+    for (const idref of opf.spine) {
+        const href = opf.manifest.get(idref);
+        if (!href)
+            continue;
+        const filePath = joinZipPath(opfPath, href);
+        const fileBuf = files.get(filePath);
+        if (!fileBuf)
+            continue;
+        const xhtml = fileBuf.toString('utf-8');
+        /* Pull only the <body>...</body> region; some EPUBs use
+           full XHTML documents with <html><head>...</head><body>...
+           so the wrap below would otherwise double-render. */
+        const bodyMatch = xhtml.match(/<body\b[^>]*>([\s\S]*)<\/body\s*>/i);
+        const body = bodyMatch && bodyMatch[1] ? bodyMatch[1] : xhtml;
+        parts.push(body);
+    }
+    /* Pass the joined body through the HTML parser. The doc gets
+       fed as a synthetic <html><body>...</body></html> so the
+       V1.19 walker's <body> shortcut still applies. */
+    const synthetic = '<html><body>' + parts.join('\n') + '</body></html>';
+    const inner = parseHtml({
+        filename: opts.filename,
+        buffer: Buffer.from(synthetic, 'utf-8'),
+        ...(opts.language ? { language: opts.language } : (opf.language ? { language: opf.language } : {})),
+    });
+    /* Override the format + title with EPUB-specific values. */
+    inner.format = 'epub';
+    inner.bytes = opts.buffer.length;
+    if (opf.title)
+        inner.title = opf.title;
+    return inner;
+}
+//# sourceMappingURL=epub.js.map

package/dist/reader/parsers/epub.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"epub.js","sourceRoot":"","sources":["../../../src/reader/parsers/epub.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA8BG;AAEH,OAAO,IAAI,MAAM,WAAW,CAAC;AAC7B,OAAO,EAAE,SAAS,EAAE,MAAM,WAAW,CAAC;AAmBtC,MAAM,QAAQ,GAAG,UAAU,CAAC;AAC5B,MAAM,MAAM,GAAG,UAAU,CAAC;AAC1B,MAAM,OAAO,GAAG,UAAU,CAAC;AAE3B,SAAS,QAAQ,CAAC,GAAW;IAC3B;mDAC+C;IAC/C,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,MAAM,EAAE,KAAK,CAAC,CAAC;IACxC,MAAM,KAAK,GAAG,GAAG,CAAC,MAAM,GAAG,GAAG,CAAC;IAC/B,KAAK,IAAI,CAAC,GAAG,GAAG,CAAC,MAAM,GAAG,EAAE,EAAE,CAAC,IAAI,KAAK,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;QACjD,IAAI,GAAG,CAAC,YAAY,CAAC,CAAC,CAAC,KAAK,QAAQ;YAAE,OAAO,CAAC,CAAC;IACjD,CAAC;IACD,MAAM,IAAI,KAAK,CAAC,0DAA0D,CAAC,CAAC;AAC9E,CAAC;AAED,SAAS,cAAc,CAAC,GAAW;IACjC,MAAM,IAAI,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAC;IAC3B,MAAM,YAAY,GAAG,GAAG,CAAC,YAAY,CAAC,IAAI,GAAG,EAAE,CAAC,CAAC;IACjD,MAAM,MAAM,GAAG,GAAG,CAAC,YAAY,CAAC,IAAI,GAAG,EAAE,CAAC,CAAC;IAC3C,MAAM,QAAQ,GAAG,GAAG,CAAC,YAAY,CAAC,IAAI,GAAG,EAAE,CAAC,CAAC;IAC7C,MAAM,OAAO,GAAsB,EAAE,CAAC;IACtC,IAAI,CAAC,GAAG,QAAQ,CAAC;IACjB,MAAM,KAAK,GAAG,QAAQ,GAAG,MAAM,CAAC;IAChC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,YAAY,IAAI,CAAC,GAAG,EAAE,IAAI,KAAK,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;QAC5D,IAAI,GAAG,CAAC,YAAY,CAAC,CAAC,CAAC,KAAK,MAAM,EAAE,CAAC;YACnC,MAAM,IAAI,KAAK,CAAC,8CAA8C,GAAG,CAAC,CAAC,CAAC;QACtE,CAAC;QACD,MAAM,MAAM,GAAG,GAAG,CAAC,YAAY,CAAC,CAAC,GAAG,EAAE,CAAC,CAAC;QACxC,MAAM,cAAc,GAAG,GAAG,CAAC,YAAY,CAAC,CAAC,GAAG,EAAE,CAAC,CAAC;QAChD,MAAM,gBAAgB,GAAG,GAAG,CAAC,YAAY,CAAC,CAAC,GAAG,EAAE,CAAC,CAAC;QAClD,MAAM,WAAW,GAAG,GAAG,CAAC,YAAY,CAAC,CAAC,GAAG,EAAE,CAAC,CAAC;QAC7C,MAAM,QAAQ,GAAG,GAAG,CAAC,YAAY,CAAC,CAAC,GAAG,EAAE,CAAC,CAAC;QAC1C,MAAM,UAAU,GAAG,GAAG,CAAC,YAAY,CAAC,CAAC,GAAG,EAAE,CAAC,CAAC;QAC5C,MAAM,iBAAiB,GAAG,GAAG,CAAC,YAAY,CAAC,CAAC,GAAG,EAAE,CAAC,CAAC;QACnD,MAAM,IAAI,GAAG,GAAG,CAAC,QAAQ,CAAC,OAAO,EAAE,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,EAAE,GAAG,WAAW,CAAC,CAAC;QACjE,OAAO,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,cAAc,EAAE,gBAAgB,EAAE,iBAAiB,EAAE,CAAC,CAAC;QACpF,CAAC,IAAI,EAAE,GAAG,WAAW,GAAG,QAAQ,GAAG,UAAU,CAAC;IAChD,CAAC;IACD,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,SAAS,aAAa,CAAC,GAAW,EAAE,KAAsB;IACxD,MAAM,CAAC,GAAG,KAAK,CAAC,iBAAiB,CAAC;IAClC,IAAI,GAAG,CAAC,YAAY,CAAC,CAAC,CAAC,KAAK,OAAO,EAAE,CAAC;QACpC,MAAM,IAAI,KAAK,CAAC,wCAAwC,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC;IACzE,CAAC;IACD,MAAM,WAAW,GAAG,GAAG,CAAC,YAAY,CAAC,CAAC,GAAG,EAAE,CAAC,CAAC;IAC7C,MAAM,QAAQ,GAAG,GAAG,CAAC,YAAY,CAAC,CAAC,GAAG,EAAE,CAAC,CAAC;IAC1C,MAAM,SAAS,GAAG,CAAC,GAAG,EAAE,GAAG,WAAW,GAAG,QAAQ,CAAC;IAClD,MAAM,OAAO,GAAG,SAAS,GAAG,KAAK,CAAC,cAAc,CAAC;IACjD,MAAM,UAAU,GAAG,GAAG,CAAC,QAAQ,CAAC,SAAS,EAAE,OAAO,CAAC,CAAC;IACpD,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACvB,OAAO,MAAM,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;IACjC,CAAC;IACD,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACvB,OAAO,IAAI,CAAC,cAAc,CAAC,UAAU,CAAC,CAAC;IACzC,CAAC;IACD,MAAM,IAAI,KAAK,CAAC,uCAAuC,GAAG,KAAK,CAAC,MAAM,GAAG,OAAO,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC;AACjG,CAAC;AAED,SAAS,eAAe,CAAC,GAAW;IAClC,MAAM,GAAG,GAAG,IAAI,GAAG,EAAkB,CAAC;IACtC,KAAK,MAAM,KAAK,IAAI,cAAc,CAAC,GAAG,CAAC,EAAE,CAAC;QACxC,yDAAyD;QACzD,IAAI,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC;YAAE,SAAS;QACvC,IAAI,CAAC;YACH,GAAG,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,EAAE,aAAa,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC,CAAC;QACjD,CAAC;QAAC,OAAO,GAAG,EAAE,CAAC;YACb;0CAC8B;QAChC,CAAC;IACH,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAED,qCAAqC;AAErC,SAAS,cAAc,CAAC,CAAS;IAC/B,OAAO,CAAC;SACL,OAAO,CAAC,qBAAqB,EAAE,CAAC,CAAC,EAAE,GAAG,EAAE,EAAE;QACzC,IAAI,CAAC;YAAC,OAAO,MAAM,CAAC,aAAa,CAAC,QAAQ,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,CAAC;QAAC,CAAC;QAAC,MAAM,CAAC;YAAC,OAAO,EAAE,CAAC;QAAC,CAAC;IAC9E,CAAC,CAAC;SACD,OAAO,CAAC,WAAW,EAAE,CAAC,CAAC,EAAE,GAAG,EAAE,EAAE;QAC/B,IAAI,CAAC;YAAC,OAAO,MAAM,CAAC,aAAa,CAAC,QAAQ,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,CAAC;QAAC,CAAC;QAAC,MAAM,CAAC;YAAC,OAAO,EAAE,CAAC;QAAC,CAAC;IAC9E,CAAC,CAAC;SACD,OAAO,CAAC,QAAQ,EAAE,GAAG,CAAC;SACtB,OAAO,CAAC,OAAO,EAAE,GAAG,CAAC;SACrB,OAAO,CAAC,OAAO,EAAE,GAAG,CAAC;SACrB,OAAO,CAAC,SAAS,EAAE,GAAG,CAAC;SACvB,OAAO,CAAC,SAAS,EAAE,GAAG,CAAC,CAAC;AAC7B,CAAC;AAED,SAAS,OAAO,CAAC,GAAW,EAAE,IAAY;IACxC,MAAM,EAAE,GAAG,IAAI,MAAM,CAAC,KAAK,GAAG,IAAI,GAAG,qCAAqC,EAAE,GAAG,CAAC,CAAC;IACjF,MAAM,CAAC,GAAG,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;IACxB,IAAI,CAAC,CAAC;QAAE,OAAO,IAAI,CAAC;IACpB,OAAO,cAAc,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;AAC5C,CAAC;AAED,SAAS,QAAQ,CAAC,GAAW,EAAE,OAAe;IAC5C,MAAM,EAAE,GAAG,IAAI,MAAM,CAAC,OAAO,GAAG,OAAO,GAAG,WAAW,EAAE,GAAG,CAAC,CAAC;IAC5D,MAAM,CAAC,GAAG,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;IACxB,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;AACzB,CAAC;AAED,SAAS,OAAO,CAAC,GAAW,EAAE,OAAe;IAC3C,MAAM,EAAE,GAAG,IAAI,MAAM,CAAC,OAAO,GAAG,OAAO,GAAG,WAAW,EAAE,IAAI,CAAC,CAAC;IAC7D,OAAO,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC;AAC7B,CAAC;AAED,SAAS,YAAY,CAAC,GAAW,EAAE,OAAe;IAChD,MAAM,EAAE,GAAG,IAAI,MAAM,CAAC,OAAO,GAAG,OAAO,GAAG,iCAAiC,GAAG,OAAO,GAAG,OAAO,EAAE,GAAG,CAAC,CAAC;IACtG,MAAM,CAAC,GAAG,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;IACxB,IAAI,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC;QAAE,OAAO,IAAI,CAAC;IAC7B,OAAO,cAAc,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,UAAU,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;AAC7D,CAAC;AAED,0CAA0C;AAE1C,SAAS,WAAW,CAAC,IAAY,EAAE,GAAW;IAC5C,+BAA+B;IAC/B,MAAM,GAAG,GAAG,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;IAClC,MAAM,OAAO,GAAG,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;IACvD,gDAAgD;IAChD,MAAM,IAAI,GAAG,CAAC,OAAO,GAAG,GAAG,CAAC,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;IACxC,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;QACvB,IAAI,CAAC,GAAG,IAAI,GAAG,KAAK,GAAG;YAAE,SAAS;QAClC,IAAI,GAAG,KAAK,IAAI,EAAE,CAAC;YACjB,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC;gBAAE,KAAK,CAAC,GAAG,EAAE,CAAC;YAClC,SAAS;QACX,CAAC;QACD,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IAClB,CAAC;IACD,OAAO,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;AACzB,CAAC;AAED,SAAS,cAAc,CAAC,YAAoB;IAC1C,MAAM,GAAG,GAAG,QAAQ,CAAC,YAAY,EAAE,UAAU,CAAC,CAAC;IAC/C,IAAI,CAAC,GAAG;QAAE,MAAM,IAAI,KAAK,CAAC,wCAAwC,CAAC,CAAC;IACpE,MAAM,IAAI,GAAG,OAAO,CAAC,GAAG,EAAE,WAAW,CAAC,CAAC;IACvC,IAAI,CAAC,IAAI;QAAE,MAAM,IAAI,KAAK,CAAC,8CAA8C,CAAC,CAAC;IAC3E,OAAO,IAAI,CAAC;AACd,CAAC;AASD,SAAS,QAAQ,CAAC,MAAc;IAC9B,MAAM,KAAK,GAAG,YAAY,CAAC,MAAM,EAAE,UAAU,CAAC,IAAI,YAAY,CAAC,MAAM,EAAE,OAAO,CAAC,IAAI,EAAE,CAAC;IACtF,MAAM,QAAQ,GAAG,YAAY,CAAC,MAAM,EAAE,aAAa,CAAC,IAAI,SAAS,CAAC;IAElE,eAAe;IACf,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAkB,CAAC;IAC3C,sDAAsD;IACtD,MAAM,aAAa,GAAG,MAAM,CAAC,KAAK,CAAC,2CAA2C,CAAC,CAAC;IAChF,IAAI,aAAa,IAAI,aAAa,CAAC,CAAC,CAAC,EAAE,CAAC;QACtC,MAAM,KAAK,GAAG,OAAO,CAAC,aAAa,CAAC,CAAC,CAAC,EAAE,MAAM,CAAC,CAAC;QAChD,KAAK,MAAM,EAAE,IAAI,KAAK,EAAE,CAAC;YACvB,MAAM,EAAE,GAAG,OAAO,CAAC,EAAE,EAAE,IAAI,CAAC,CAAC;YAC7B,MAAM,IAAI,GAAG,OAAO,CAAC,EAAE,EAAE,MAAM,CAAC,CAAC;YACjC,IAAI,EAAE,IAAI,IAAI;gBAAE,QAAQ,CAAC,GAAG,CAAC,EAAE,EAAE,IAAI,CAAC,CAAC;QACzC,CAAC;IACH,CAAC;IAED,YAAY;IACZ,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,MAAM,UAAU,GAAG,MAAM,CAAC,KAAK,CAAC,qCAAqC,CAAC,CAAC;IACvE,IAAI,UAAU,IAAI,UAAU,CAAC,CAAC,CAAC,EAAE,CAAC;QAChC,MAAM,KAAK,GAAG,OAAO,CAAC,UAAU,CAAC,CAAC,CAAC,EAAE,SAAS,CAAC,CAAC;QAChD,KAAK,MAAM,EAAE,IAAI,KAAK,EAAE,CAAC;YACvB,MAAM,KAAK,GAAG,OAAO,CAAC,EAAE,EAAE,OAAO,CAAC,CAAC;YACnC,IAAI,KAAK;gBAAE,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QAC/B,CAAC;IACH,CAAC;IACD,MAAM,IAAI,GAAkB,EAAE,KAAK,EAAE,QAAQ,EAAE,KAAK,EAAE,CAAC;IACvD,IAAI,QAAQ;QAAE,IAAI,CAAC,QAAQ,GAAG,QAAQ,CAAC;IACvC,OAAO,IAAI,CAAC;AACd,CAAC;AAED,sBAAsB;AAEtB,MAAM,UAAU,SAAS,CAAC,IAAsB;IAC9C,MAAM,KAAK,GAAG,eAAe,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;IAC3C,MAAM,YAAY,GAAG,KAAK,CAAC,GAAG,CAAC,wBAAwB,CAAC,CAAC;IACzD,IAAI,CAAC,YAAY,EAAE,CAAC;QAClB,MAAM,IAAI,KAAK,CAAC,mDAAmD,CAAC,CAAC;IACvE,CAAC;IACD,MAAM,OAAO,GAAG,cAAc,CAAC,YAAY,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC,CAAC;IAC/D,MAAM,MAAM,GAAG,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;IAClC,IAAI,CAAC,MAAM,EAAE,CAAC;QACZ,MAAM,IAAI,KAAK,CAAC,kBAAkB,GAAG,OAAO,GAAG,wBAAwB,CAAC,CAAC;IAC3E,CAAC;IACD,MAAM,GAAG,GAAG,QAAQ,CAAC,MAAM,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC,CAAC;IAE/C;;sCAEkC;IAClC,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,KAAK,MAAM,KAAK,IAAI,GAAG,CAAC,KAAK,EAAE,CAAC;QAC9B,MAAM,IAAI,GAAG,GAAG,CAAC,QAAQ,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;QACrC,IAAI,CAAC,IAAI;YAAE,SAAS;QACpB,MAAM,QAAQ,GAAG,WAAW,CAAC,OAAO,EAAE,IAAI,CAAC,CAAC;QAC5C,MAAM,OAAO,GAAG,KAAK,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;QACpC,IAAI,CAAC,OAAO;YAAE,SAAS;QACvB,MAAM,KAAK,GAAG,OAAO,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC;QACxC;;8DAEsD;QACtD,MAAM,SAAS,GAAG,KAAK,CAAC,KAAK,CAAC,oCAAoC,CAAC,CAAC;QACpE,MAAM,IAAI,GAAG,SAAS,IAAI,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC;QAC9D,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IACnB,CAAC;IAED;;uDAEmD;IACnD,MAAM,SAAS,GAAG,cAAc,GAAG,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,gBAAgB,CAAC;IACvE,MAAM,KAAK,GAAG,SAAS,CAAC;QACtB,QAAQ,EAAE,IAAI,CAAC,QAAQ;QACvB,MAAM,EAAE,MAAM,CAAC,IAAI,CAAC,SAAS,EAAE,OAAO,CAAC;QACvC,GAAG,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,IAAI,CAAC,QAAQ,EAAE,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,GAAG,CAAC,QAAQ,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;KACpG,CAAC,CAAC;IAEH,4DAA4D;IAC5D,KAAK,CAAC,MAAM,GAAG,MAAM,CAAC;IACtB,KAAK,CAAC,KAAK,GAAG,IAAI,CAAC,MAAM,CAAC,MAAM,CAAC;IACjC,IAAI,GAAG,CAAC,KAAK;QAAE,KAAK,CAAC,KAAK,GAAG,GAAG,CAAC,KAAK,CAAC;IACvC,OAAO,KAAK,CAAC;AACf,CAAC"}

package/dist/reader/parsers/html.d.ts

@@ -0,0 +1,40 @@
+/**
+ * HTML parser for the document reader (V1.19 of HITO 1).
+ *
+ * Tiny hand-rolled parser. No npm dep. Targets the subset of
+ * HTML that matters for voice readback of design/spec docs
+ * the ADIP team would feed Forge:
+ *
+ *   - <h1>..<h6> define sections; the heading text is the
+ *     section heading and the level matches the tag number.
+ *   - <p>, <div>, <li> -> paragraph blocks.
+ *   - <blockquote> -> quote blocks (kept as their own text
+ *     family so TTS can prosody-shape later).
+ *   - <ul> / <ol> children -> list_item blocks.
+ *   - <pre> / <code> -> non-text code block with content
+ *     preserved verbatim.
+ *   - <table> -> non-text 'table' with a summary "tabla con
+ *     N filas".
+ *   - <img alt="X" src="Y"> -> non-text 'image' with the alt
+ *     text as summary (caption fallback when no alt).
+ *   - <title> in <head> -> document title.
+ *   - <script> / <style> / <noscript> / HTML comments stripped
+ *     entirely before walking.
+ *
+ * The parser is regex-based. HTML is famously hostile to regex
+ * parsing, but the spec-doc use case is well-formed enough that
+ * a regex walker covers it. Pathological inputs (interleaved
+ * tags, unclosed elements with nested children) fall back to
+ * the surrounding paragraph at worst -- voice reader degrades
+ * gracefully instead of crashing.
+ *
+ * ASCII-only.
+ */
+import type { NormalisedDocument } from '../types.js';
+export interface ParseHtmlOptions {
+    buffer: Buffer;
+    filename: string;
+    language?: string;
+}
+export declare function parseHtml(opts: ParseHtmlOptions): NormalisedDocument;
+//# sourceMappingURL=html.d.ts.map

package/dist/reader/parsers/html.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"html.d.ts","sourceRoot":"","sources":["../../../src/reader/parsers/html.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA+BG;AAEH,OAAO,KAAK,EAAE,kBAAkB,EAAgC,MAAM,aAAa,CAAC;AAEpF,MAAM,WAAW,gBAAgB;IAC/B,MAAM,EAAE,MAAM,CAAC;IACf,QAAQ,EAAE,MAAM,CAAC;IACjB,QAAQ,CAAC,EAAE,MAAM,CAAC;CACnB;AAuSD,wBAAgB,SAAS,CAAC,IAAI,EAAE,gBAAgB,GAAG,kBAAkB,CAkGpE"}