@kywi-software/core 0.1.0

package/dist/ax/html-to-markdown.js

@@ -0,0 +1,295 @@
+/**
+ * Dependency-free HTML → Markdown converter for the AX layer.
+ *
+ * This is a deterministic projection of stored `body` HTML into Markdown, used
+ * by `llms-full.txt` and the per-page markdown negotiation surface. It is
+ * intentionally small and string-built (mirroring `feeds/rss.ts`) rather than
+ * pulling in a parser dependency; `isomorphic-dompurify` is stubbed in this repo
+ * so we do not lean on a DOM.
+ *
+ * Coverage: h1–h6, p, br, a, strong/b, em/i, code, pre, blockquote, ul/ol/li,
+ * img, hr. Unknown tags are unwrapped to their inner text. Output is stable for
+ * a given input (no timestamps, no randomness).
+ */
+const VOID_TAGS = new Set(['br', 'hr', 'img', 'meta', 'link', 'input']);
+// ─── Entity decoding ──────────────────────────────────────────────────────────
+const NAMED_ENTITIES = {
+    amp: '&', lt: '<', gt: '>', quot: '"', apos: "'", nbsp: ' ',
+    mdash: '—', ndash: '–', hellip: '…', copy: '©', reg: '®', trade: '™',
+};
+function decodeEntities(input) {
+    return input.replace(/&(#x?[0-9a-fA-F]+|[a-zA-Z]+);/g, (whole, body) => {
+        if (body[0] === '#') {
+            const isHex = body[1] === 'x' || body[1] === 'X';
+            const code = parseInt(isHex ? body.slice(2) : body.slice(1), isHex ? 16 : 10);
+            return Number.isNaN(code) ? whole : String.fromCodePoint(code);
+        }
+        const mapped = NAMED_ENTITIES[body.toLowerCase()];
+        return mapped ?? whole;
+    });
+}
+function parseAttrs(raw) {
+    const attrs = {};
+    const re = /([a-zA-Z_:][-a-zA-Z0-9_:.]*)(?:\s*=\s*("([^"]*)"|'([^']*)'|([^\s"'=<>`]+)))?/g;
+    let m;
+    while ((m = re.exec(raw)) !== null) {
+        const name = m[1].toLowerCase();
+        const value = m[3] ?? m[4] ?? m[5] ?? '';
+        attrs[name] = decodeEntities(value);
+    }
+    return attrs;
+}
+function tokenize(html) {
+    const tokens = [];
+    const tagRe = /<(\/?)([a-zA-Z][a-zA-Z0-9]*)((?:[^>"']|"[^"]*"|'[^']*')*?)(\/?)>/g;
+    let lastIndex = 0;
+    let m;
+    while ((m = tagRe.exec(html)) !== null) {
+        if (m.index > lastIndex) {
+            tokens.push({ type: 'text', text: html.slice(lastIndex, m.index) });
+        }
+        const isClose = m[1] === '/';
+        const tag = m[2].toLowerCase();
+        const selfClose = m[4] === '/';
+        if (isClose) {
+            tokens.push({ type: 'close', tag });
+        }
+        else if (selfClose || VOID_TAGS.has(tag)) {
+            tokens.push({ type: 'void', tag, attrs: parseAttrs(m[3] ?? '') });
+        }
+        else {
+            tokens.push({ type: 'open', tag, attrs: parseAttrs(m[3] ?? '') });
+        }
+        lastIndex = tagRe.lastIndex;
+    }
+    if (lastIndex < html.length) {
+        tokens.push({ type: 'text', text: html.slice(lastIndex) });
+    }
+    return tokens;
+}
+function buildTree(tokens) {
+    const root = { tag: 'root', text: '', attrs: {}, children: [] };
+    const stack = [root];
+    for (const tok of tokens) {
+        const parent = stack[stack.length - 1];
+        if (tok.type === 'text') {
+            if (tok.text)
+                parent.children.push({ tag: null, text: tok.text, attrs: {}, children: [] });
+        }
+        else if (tok.type === 'void') {
+            parent.children.push({ tag: tok.tag, text: '', attrs: tok.attrs ?? {}, children: [] });
+        }
+        else if (tok.type === 'open') {
+            const node = { tag: tok.tag, text: '', attrs: tok.attrs ?? {}, children: [] };
+            parent.children.push(node);
+            stack.push(node);
+        }
+        else {
+            // close: pop up to the matching open tag (tolerant of misnesting)
+            for (let i = stack.length - 1; i >= 1; i--) {
+                if (stack[i].tag === tok.tag) {
+                    stack.length = i;
+                    break;
+                }
+            }
+        }
+    }
+    return root;
+}
+// ─── Rendering ────────────────────────────────────────────────────────────────
+function collapseWhitespace(s) {
+    return s.replace(/\s+/g, ' ');
+}
+/** Render inline content (text + inline elements) to a single Markdown string. */
+function renderInline(nodes) {
+    let out = '';
+    for (const node of nodes) {
+        if (node.tag === null) {
+            out += collapseWhitespace(decodeEntities(node.text));
+            continue;
+        }
+        switch (node.tag) {
+            case 'br':
+                out += '  \n';
+                break;
+            case 'strong':
+            case 'b':
+                out += `**${renderInline(node.children).trim()}**`;
+                break;
+            case 'em':
+            case 'i':
+                out += `*${renderInline(node.children).trim()}*`;
+                break;
+            case 'code':
+                out += `\`${renderInline(node.children).trim()}\``;
+                break;
+            case 'a': {
+                const href = node.attrs['href'] ?? '';
+                const label = renderInline(node.children).trim();
+                out += href ? `[${label}](${href})` : label;
+                break;
+            }
+            case 'img': {
+                const src = node.attrs['src'] ?? '';
+                const alt = node.attrs['alt'] ?? '';
+                out += src ? `![${alt}](${src})` : '';
+                break;
+            }
+            default:
+                // Unknown/inline-ish element → unwrap to its rendered children.
+                out += renderInline(node.children);
+        }
+    }
+    return out;
+}
+function repeat(ch, n) {
+    return n > 0 ? ch.repeat(n) : '';
+}
+/** Render a list, tracking ordered/unordered and nesting depth. */
+function renderList(node, ordered, depth) {
+    const lines = [];
+    let index = 1;
+    for (const child of node.children) {
+        if (child.tag !== 'li')
+            continue;
+        const indent = repeat('  ', depth);
+        const marker = ordered ? `${index}.` : '-';
+        // Split the li into inline content vs nested lists.
+        const inlineChildren = child.children.filter(c => c.tag !== 'ul' && c.tag !== 'ol');
+        const nestedLists = child.children.filter(c => c.tag === 'ul' || c.tag === 'ol');
+        const text = renderInline(inlineChildren).trim();
+        lines.push(`${indent}${marker} ${text}`.replace(/\s+$/, ''));
+        for (const nested of nestedLists) {
+            lines.push(renderList(nested, nested.tag === 'ol', depth + 1));
+        }
+        index++;
+    }
+    return lines.join('\n');
+}
+/** Render block-level nodes, separating blocks with blank lines. */
+function renderBlocks(nodes) {
+    const blocks = [];
+    let inlineRun = [];
+    const flushInline = () => {
+        if (inlineRun.length === 0)
+            return;
+        const text = renderInline(inlineRun).trim();
+        if (text)
+            blocks.push(text);
+        inlineRun = [];
+    };
+    for (const node of nodes) {
+        const tag = node.tag;
+        if (tag === null) {
+            inlineRun.push(node);
+            continue;
+        }
+        switch (tag) {
+            case 'h1':
+            case 'h2':
+            case 'h3':
+            case 'h4':
+            case 'h5':
+            case 'h6': {
+                flushInline();
+                const level = Number(tag[1]);
+                blocks.push(`${repeat('#', level)} ${renderInline(node.children).trim()}`);
+                break;
+            }
+            case 'p': {
+                flushInline();
+                const text = renderInline(node.children).trim();
+                if (text)
+                    blocks.push(text);
+                break;
+            }
+            case 'br':
+                inlineRun.push(node);
+                break;
+            case 'hr':
+                flushInline();
+                blocks.push('---');
+                break;
+            case 'ul':
+            case 'ol': {
+                flushInline();
+                blocks.push(renderList(node, tag === 'ol', 0));
+                break;
+            }
+            case 'blockquote': {
+                flushInline();
+                const inner = renderBlocks(node.children).trim();
+                blocks.push(inner.split('\n').map(l => (l ? `> ${l}` : '>')).join('\n'));
+                break;
+            }
+            case 'pre': {
+                flushInline();
+                // Preserve pre text verbatim (decode entities, strip an inner <code>).
+                const codeText = extractText(node);
+                blocks.push('```\n' + codeText.replace(/\n+$/, '') + '\n```');
+                break;
+            }
+            case 'a':
+            case 'strong':
+            case 'b':
+            case 'em':
+            case 'i':
+            case 'code':
+            case 'img':
+            case 'span':
+                inlineRun.push(node);
+                break;
+            case 'div':
+            case 'section':
+            case 'article':
+            case 'main':
+            case 'header':
+            case 'footer':
+            case 'figure':
+            case 'figcaption': {
+                flushInline();
+                const inner = renderBlocks(node.children).trim();
+                if (inner)
+                    blocks.push(inner);
+                break;
+            }
+            default: {
+                // Unknown block → recurse into children as blocks.
+                flushInline();
+                const inner = renderBlocks(node.children).trim();
+                if (inner)
+                    blocks.push(inner);
+            }
+        }
+    }
+    flushInline();
+    return blocks.join('\n\n');
+}
+/** Recursively collect decoded text from a node (for <pre>). */
+function extractText(node) {
+    if (node.tag === null)
+        return decodeEntities(node.text);
+    if (node.tag === 'br')
+        return '\n';
+    return node.children.map(extractText).join('');
+}
+/**
+ * Convert an HTML string to Markdown. Returns '' for null/undefined/empty input.
+ */
+export function htmlToMarkdown(html) {
+    if (!html)
+        return '';
+    const tree = buildTree(tokenize(html));
+    return renderBlocks(tree.children).trim();
+}
+/**
+ * Strip all HTML tags to plain text (decoding entities, collapsing whitespace).
+ * Used by the deterministic summary fallback chain.
+ */
+export function stripHtml(html) {
+    if (!html)
+        return '';
+    const tree = buildTree(tokenize(html));
+    return collapseWhitespace(extractText(tree)).trim();
+}
+//# sourceMappingURL=html-to-markdown.js.map

package/dist/ax/html-to-markdown.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"html-to-markdown.js","sourceRoot":"","sources":["../../src/ax/html-to-markdown.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;GAYG;AASH,MAAM,SAAS,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,IAAI,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,OAAO,CAAC,CAAC,CAAA;AAEvE,iFAAiF;AAEjF,MAAM,cAAc,GAA2B;IAC7C,GAAG,EAAE,GAAG,EAAE,EAAE,EAAE,GAAG,EAAE,EAAE,EAAE,GAAG,EAAE,IAAI,EAAE,GAAG,EAAE,IAAI,EAAE,GAAG,EAAE,IAAI,EAAE,GAAG;IAC3D,KAAK,EAAE,GAAG,EAAE,KAAK,EAAE,GAAG,EAAE,MAAM,EAAE,GAAG,EAAE,IAAI,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,KAAK,EAAE,GAAG;CACrE,CAAA;AAED,SAAS,cAAc,CAAC,KAAa;IACnC,OAAO,KAAK,CAAC,OAAO,CAAC,gCAAgC,EAAE,CAAC,KAAK,EAAE,IAAY,EAAE,EAAE;QAC7E,IAAI,IAAI,CAAC,CAAC,CAAC,KAAK,GAAG,EAAE,CAAC;YACpB,MAAM,KAAK,GAAG,IAAI,CAAC,CAAC,CAAC,KAAK,GAAG,IAAI,IAAI,CAAC,CAAC,CAAC,KAAK,GAAG,CAAA;YAChD,MAAM,IAAI,GAAG,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAA;YAC7E,OAAO,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,MAAM,CAAC,aAAa,CAAC,IAAI,CAAC,CAAA;QAChE,CAAC;QACD,MAAM,MAAM,GAAG,cAAc,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,CAAA;QACjD,OAAO,MAAM,IAAI,KAAK,CAAA;IACxB,CAAC,CAAC,CAAA;AACJ,CAAC;AAWD,SAAS,UAAU,CAAC,GAAW;IAC7B,MAAM,KAAK,GAA2B,EAAE,CAAA;IACxC,MAAM,EAAE,GAAG,+EAA+E,CAAA;IAC1F,IAAI,CAAyB,CAAA;IAC7B,OAAO,CAAC,CAAC,GAAG,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;QACnC,MAAM,IAAI,GAAG,CAAC,CAAC,CAAC,CAAE,CAAC,WAAW,EAAE,CAAA;QAChC,MAAM,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAA;QACxC,KAAK,CAAC,IAAI,CAAC,GAAG,cAAc,CAAC,KAAK,CAAC,CAAA;IACrC,CAAC;IACD,OAAO,KAAK,CAAA;AACd,CAAC;AAED,SAAS,QAAQ,CAAC,IAAY;IAC5B,MAAM,MAAM,GAAY,EAAE,CAAA;IAC1B,MAAM,KAAK,GAAG,mEAAmE,CAAA;IACjF,IAAI,SAAS,GAAG,CAAC,CAAA;IACjB,IAAI,CAAyB,CAAA;IAC7B,OAAO,CAAC,CAAC,GAAG,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;QACvC,IAAI,CAAC,CAAC,KAAK,GAAG,SAAS,EAAE,CAAC;YACxB,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,IAAI,CAAC,KAAK,CAAC,SAAS,EAAE,CAAC,CAAC,KAAK,CAAC,EAAE,CAAC,CAAA;QACrE,CAAC;QACD,MAAM,OAAO,GAAG,CAAC,CAAC,CAAC,CAAC,KAAK,GAAG,CAAA;QAC5B,MAAM,GAAG,GAAG,CAAC,CAAC,CAAC,CAAE,CAAC,WAAW,EAAE,CAAA;QAC/B,MAAM,SAAS,GAAG,CAAC,CAAC,CAAC,CAAC,KAAK,GAAG,CAAA;QAC9B,IAAI,OAAO,EAAE,CAAC;YACZ,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,GAAG,EAAE,CAAC,CAAA;QACrC,CAAC;aAAM,IAAI,SAAS,IAAI,SAAS,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC;YAC3C,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,GAAG,EAAE,KAAK,EAAE,UAAU,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,EAAE,CAAC,CAAA;QACnE,CAAC;aAAM,CAAC;YACN,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,GAAG,EAAE,KAAK,EAAE,UAAU,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,EAAE,CAAC,CAAA;QACnE,CAAC;QACD,SAAS,GAAG,KAAK,CAAC,SAAS,CAAA;IAC7B,CAAC;IACD,IAAI,SAAS,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC;QAC5B,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,IAAI,CAAC,KAAK,CAAC,SAAS,CAAC,EAAE,CAAC,CAAA;IAC5D,CAAC;IACD,OAAO,MAAM,CAAA;AACf,CAAC;AAED,SAAS,SAAS,CAAC,MAAe;IAChC,MAAM,IAAI,GAAa,EAAE,GAAG,EAAE,MAAM,EAAE,IAAI,EAAE,EAAE,EAAE,KAAK,EAAE,EAAE,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAA;IACzE,MAAM,KAAK,GAAe,CAAC,IAAI,CAAC,CAAA;IAChC,KAAK,MAAM,GAAG,IAAI,MAAM,EAAE,CAAC;QACzB,MAAM,MAAM,GAAG,KAAK,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAE,CAAA;QACvC,IAAI,GAAG,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YACxB,IAAI,GAAG,CAAC,IAAI;gBAAE,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,EAAE,GAAG,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC,CAAA;QAC5F,CAAC;aAAM,IAAI,GAAG,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YAC/B,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,EAAE,GAAG,EAAE,GAAG,CAAC,GAAI,EAAE,IAAI,EAAE,EAAE,EAAE,KAAK,EAAE,GAAG,CAAC,KAAK,IAAI,EAAE,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC,CAAA;QACzF,CAAC;aAAM,IAAI,GAAG,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YAC/B,MAAM,IAAI,GAAa,EAAE,GAAG,EAAE,GAAG,CAAC,GAAI,EAAE,IAAI,EAAE,EAAE,EAAE,KAAK,EAAE,GAAG,CAAC,KAAK,IAAI,EAAE,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAA;YACxF,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;YAC1B,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;QAClB,CAAC;aAAM,CAAC;YACN,kEAAkE;YAClE,KAAK,IAAI,CAAC,GAAG,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;gBAC3C,IAAI,KAAK,CAAC,CAAC,CAAE,CAAC,GAAG,KAAK,GAAG,CAAC,GAAG,EAAE,CAAC;oBAC9B,KAAK,CAAC,MAAM,GAAG,CAAC,CAAA;oBAChB,MAAK;gBACP,CAAC;YACH,CAAC;QACH,CAAC;IACH,CAAC;IACD,OAAO,IAAI,CAAA;AACb,CAAC;AAED,iFAAiF;AAEjF,SAAS,kBAAkB,CAAC,CAAS;IACnC,OAAO,CAAC,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAA;AAC/B,CAAC;AAED,kFAAkF;AAClF,SAAS,YAAY,CAAC,KAAiB;IACrC,IAAI,GAAG,GAAG,EAAE,CAAA;IACZ,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,IAAI,CAAC,GAAG,KAAK,IAAI,EAAE,CAAC;YACtB,GAAG,IAAI,kBAAkB,CAAC,cAAc,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAA;YACpD,SAAQ;QACV,CAAC;QACD,QAAQ,IAAI,CAAC,GAAG,EAAE,CAAC;YACjB,KAAK,IAAI;gBACP,GAAG,IAAI,MAAM,CAAA;gBACb,MAAK;YACP,KAAK,QAAQ,CAAC;YACd,KAAK,GAAG;gBACN,GAAG,IAAI,KAAK,YAAY,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,IAAI,EAAE,IAAI,CAAA;gBAClD,MAAK;YACP,KAAK,IAAI,CAAC;YACV,KAAK,GAAG;gBACN,GAAG,IAAI,IAAI,YAAY,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,IAAI,EAAE,GAAG,CAAA;gBAChD,MAAK;YACP,KAAK,MAAM;gBACT,GAAG,IAAI,KAAK,YAAY,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,IAAI,EAAE,IAAI,CAAA;gBAClD,MAAK;YACP,KAAK,GAAG,CAAC,CAAC,CAAC;gBACT,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,IAAI,EAAE,CAAA;gBACrC,MAAM,KAAK,GAAG,YAAY,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,IAAI,EAAE,CAAA;gBAChD,GAAG,IAAI,IAAI,CAAC,CAAC,CAAC,IAAI,KAAK,KAAK,IAAI,GAAG,CAAC,CAAC,CAAC,KAAK,CAAA;gBAC3C,MAAK;YACP,CAAC;YACD,KAAK,KAAK,CAAC,CAAC,CAAC;gBACX,MAAM,GAAG,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,IAAI,EAAE,CAAA;gBACnC,MAAM,GAAG,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,IAAI,EAAE,CAAA;gBACnC,GAAG,IAAI,GAAG,CAAC,CAAC,CAAC,KAAK,GAAG,KAAK,GAAG,GAAG,CAAC,CAAC,CAAC,EAAE,CAAA;gBACrC,MAAK;YACP,CAAC;YACD;gBACE,gEAAgE;gBAChE,GAAG,IAAI,YAAY,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAA;QACtC,CAAC;IACH,CAAC;IACD,OAAO,GAAG,CAAA;AACZ,CAAC;AAED,SAAS,MAAM,CAAC,EAAU,EAAE,CAAS;IACnC,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAA;AAClC,CAAC;AAED,mEAAmE;AACnE,SAAS,UAAU,CAAC,IAAc,EAAE,OAAgB,EAAE,KAAa;IACjE,MAAM,KAAK,GAAa,EAAE,CAAA;IAC1B,IAAI,KAAK,GAAG,CAAC,CAAA;IACb,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;QAClC,IAAI,KAAK,CAAC,GAAG,KAAK,IAAI;YAAE,SAAQ;QAChC,MAAM,MAAM,GAAG,MAAM,CAAC,IAAI,EAAE,KAAK,CAAC,CAAA;QAClC,MAAM,MAAM,GAAG,OAAO,CAAC,CAAC,CAAC,GAAG,KAAK,GAAG,CAAC,CAAC,CAAC,GAAG,CAAA;QAC1C,oDAAoD;QACpD,MAAM,cAAc,GAAG,KAAK,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,KAAK,IAAI,IAAI,CAAC,CAAC,GAAG,KAAK,IAAI,CAAC,CAAA;QACnF,MAAM,WAAW,GAAG,KAAK,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,KAAK,IAAI,IAAI,CAAC,CAAC,GAAG,KAAK,IAAI,CAAC,CAAA;QAChF,MAAM,IAAI,GAAG,YAAY,CAAC,cAAc,CAAC,CAAC,IAAI,EAAE,CAAA;QAChD,KAAK,CAAC,IAAI,CAAC,GAAG,MAAM,GAAG,MAAM,IAAI,IAAI,EAAE,CAAC,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,CAAA;QAC5D,KAAK,MAAM,MAAM,IAAI,WAAW,EAAE,CAAC;YACjC,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,MAAM,EAAE,MAAM,CAAC,GAAG,KAAK,IAAI,EAAE,KAAK,GAAG,CAAC,CAAC,CAAC,CAAA;QAChE,CAAC;QACD,KAAK,EAAE,CAAA;IACT,CAAC;IACD,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;AACzB,CAAC;AAED,oEAAoE;AACpE,SAAS,YAAY,CAAC,KAAiB;IACrC,MAAM,MAAM,GAAa,EAAE,CAAA;IAC3B,IAAI,SAAS,GAAe,EAAE,CAAA;IAE9B,MAAM,WAAW,GAAG,GAAG,EAAE;QACvB,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC;YAAE,OAAM;QAClC,MAAM,IAAI,GAAG,YAAY,CAAC,SAAS,CAAC,CAAC,IAAI,EAAE,CAAA;QAC3C,IAAI,IAAI;YAAE,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;QAC3B,SAAS,GAAG,EAAE,CAAA;IAChB,CAAC,CAAA;IAED,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,CAAA;QACpB,IAAI,GAAG,KAAK,IAAI,EAAE,CAAC;YACjB,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;YACpB,SAAQ;QACV,CAAC;QACD,QAAQ,GAAG,EAAE,CAAC;YACZ,KAAK,IAAI,CAAC;YAAC,KAAK,IAAI,CAAC;YAAC,KAAK,IAAI,CAAC;YAAC,KAAK,IAAI,CAAC;YAAC,KAAK,IAAI,CAAC;YAAC,KAAK,IAAI,CAAC,CAAC,CAAC;gBACjE,WAAW,EAAE,CAAA;gBACb,MAAM,KAAK,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAA;gBAC5B,MAAM,CAAC,IAAI,CAAC,GAAG,MAAM,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,YAAY,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAA;gBAC1E,MAAK;YACP,CAAC;YACD,KAAK,GAAG,CAAC,CAAC,CAAC;gBACT,WAAW,EAAE,CAAA;gBACb,MAAM,IAAI,GAAG,YAAY,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,IAAI,EAAE,CAAA;gBAC/C,IAAI,IAAI;oBAAE,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;gBAC3B,MAAK;YACP,CAAC;YACD,KAAK,IAAI;gBACP,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;gBACpB,MAAK;YACP,KAAK,IAAI;gBACP,WAAW,EAAE,CAAA;gBACb,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;gBAClB,MAAK;YACP,KAAK,IAAI,CAAC;YAAC,KAAK,IAAI,CAAC,CAAC,CAAC;gBACrB,WAAW,EAAE,CAAA;gBACb,MAAM,CAAC,IAAI,CAAC,UAAU,CAAC,IAAI,EAAE,GAAG,KAAK,IAAI,EAAE,CAAC,CAAC,CAAC,CAAA;gBAC9C,MAAK;YACP,CAAC;YACD,KAAK,YAAY,CAAC,CAAC,CAAC;gBAClB,WAAW,EAAE,CAAA;gBACb,MAAM,KAAK,GAAG,YAAY,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,IAAI,EAAE,CAAA;gBAChD,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAA;gBACxE,MAAK;YACP,CAAC;YACD,KAAK,KAAK,CAAC,CAAC,CAAC;gBACX,WAAW,EAAE,CAAA;gBACb,uEAAuE;gBACvE,MAAM,QAAQ,GAAG,WAAW,CAAC,IAAI,CAAC,CAAA;gBAClC,MAAM,CAAC,IAAI,CAAC,OAAO,GAAG,QAAQ,CAAC,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC,GAAG,OAAO,CAAC,CAAA;gBAC7D,MAAK;YACP,CAAC;YACD,KAAK,GAAG,CAAC;YAAC,KAAK,QAAQ,CAAC;YAAC,KAAK,GAAG,CAAC;YAAC,KAAK,IAAI,CAAC;YAAC,KAAK,GAAG,CAAC;YAAC,KAAK,MAAM,CAAC;YAAC,KAAK,KAAK,CAAC;YAAC,KAAK,MAAM;gBAC1F,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;gBACpB,MAAK;YACP,KAAK,KAAK,CAAC;YAAC,KAAK,SAAS,CAAC;YAAC,KAAK,SAAS,CAAC;YAAC,KAAK,MAAM,CAAC;YAAC,KAAK,QAAQ,CAAC;YACvE,KAAK,QAAQ,CAAC;YAAC,KAAK,QAAQ,CAAC;YAAC,KAAK,YAAY,CAAC,CAAC,CAAC;gBAChD,WAAW,EAAE,CAAA;gBACb,MAAM,KAAK,GAAG,YAAY,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,IAAI,EAAE,CAAA;gBAChD,IAAI,KAAK;oBAAE,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;gBAC7B,MAAK;YACP,CAAC;YACD,OAAO,CAAC,CAAC,CAAC;gBACR,mDAAmD;gBACnD,WAAW,EAAE,CAAA;gBACb,MAAM,KAAK,GAAG,YAAY,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,IAAI,EAAE,CAAA;gBAChD,IAAI,KAAK;oBAAE,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;YAC/B,CAAC;QACH,CAAC;IACH,CAAC;IACD,WAAW,EAAE,CAAA;IACb,OAAO,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;AAC5B,CAAC;AAED,gEAAgE;AAChE,SAAS,WAAW,CAAC,IAAc;IACjC,IAAI,IAAI,CAAC,GAAG,KAAK,IAAI;QAAE,OAAO,cAAc,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;IACvD,IAAI,IAAI,CAAC,GAAG,KAAK,IAAI;QAAE,OAAO,IAAI,CAAA;IAClC,OAAO,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,WAAW,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAA;AAChD,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,cAAc,CAAC,IAA+B;IAC5D,IAAI,CAAC,IAAI;QAAE,OAAO,EAAE,CAAA;IACpB,MAAM,IAAI,GAAG,SAAS,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAA;IACtC,OAAO,YAAY,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,IAAI,EAAE,CAAA;AAC3C,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,SAAS,CAAC,IAA+B;IACvD,IAAI,CAAC,IAAI;QAAE,OAAO,EAAE,CAAA;IACpB,MAAM,IAAI,GAAG,SAAS,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAA;IACtC,OAAO,kBAAkB,CAAC,WAAW,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,EAAE,CAAA;AACrD,CAAC"}

package/dist/ax/index.d.ts

@@ -0,0 +1,25 @@
+/**
+ * @kywi-software/core/ax — the Agent Experience (Tier 1) layer.
+ *
+ * Server-only, deterministic generators that read one `AxContentSource` (the
+ * published content tree) and emit each agent-facing projection: llms.txt,
+ * llms-full.txt, sitemap.xml, robots.txt, per-page JSON-LD, and single-page
+ * markdown.
+ *
+ * BOUNDARY INVARIANT: nothing in this module may import a model-provider client.
+ * Tier 1 never calls a language model — missing data uses a deterministic
+ * fallback, so output is byte-stable across crawls.
+ */
+export type { AxContentNode, AxSiteMeta, AxContentSource, AxRenderContext, } from './types.js';
+export { createDbAxContentSource } from './content-source.js';
+export { resolveBaseUrl, absoluteUrl } from './url.js';
+export { htmlToMarkdown, stripHtml } from './html-to-markdown.js';
+export { summaryFor, truncate, oneLine } from './summary.js';
+export { isAxEligible, isIndexable, passesTypeFilter, INDEXABLE_BASE_TYPES } from './filters.js';
+export { generateLlmsTxt } from './llms-txt.js';
+export { generateLlmsFullTxt } from './llms-full-txt.js';
+export { generateSitemapXml } from './sitemap-xml.js';
+export { generateRobotsTxt, AI_CRAWLERS } from './robots-txt.js';
+export { generateJsonLd } from './jsonld.js';
+export { generatePageMarkdown } from './page-markdown.js';
+//# sourceMappingURL=index.d.ts.map

package/dist/ax/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/ax/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AACH,YAAY,EACV,aAAa,EACb,UAAU,EACV,eAAe,EACf,eAAe,GAChB,MAAM,YAAY,CAAA;AAEnB,OAAO,EAAE,uBAAuB,EAAE,MAAM,qBAAqB,CAAA;AAC7D,OAAO,EAAE,cAAc,EAAE,WAAW,EAAE,MAAM,UAAU,CAAA;AACtD,OAAO,EAAE,cAAc,EAAE,SAAS,EAAE,MAAM,uBAAuB,CAAA;AACjE,OAAO,EAAE,UAAU,EAAE,QAAQ,EAAE,OAAO,EAAE,MAAM,cAAc,CAAA;AAC5D,OAAO,EAAE,YAAY,EAAE,WAAW,EAAE,gBAAgB,EAAE,oBAAoB,EAAE,MAAM,cAAc,CAAA;AAChG,OAAO,EAAE,eAAe,EAAE,MAAM,eAAe,CAAA;AAC/C,OAAO,EAAE,mBAAmB,EAAE,MAAM,oBAAoB,CAAA;AACxD,OAAO,EAAE,kBAAkB,EAAE,MAAM,kBAAkB,CAAA;AACrD,OAAO,EAAE,iBAAiB,EAAE,WAAW,EAAE,MAAM,iBAAiB,CAAA;AAChE,OAAO,EAAE,cAAc,EAAE,MAAM,aAAa,CAAA;AAC5C,OAAO,EAAE,oBAAoB,EAAE,MAAM,oBAAoB,CAAA"}

package/dist/ax/index.js

@@ -0,0 +1,12 @@
+export { createDbAxContentSource } from './content-source.js';
+export { resolveBaseUrl, absoluteUrl } from './url.js';
+export { htmlToMarkdown, stripHtml } from './html-to-markdown.js';
+export { summaryFor, truncate, oneLine } from './summary.js';
+export { isAxEligible, isIndexable, passesTypeFilter, INDEXABLE_BASE_TYPES } from './filters.js';
+export { generateLlmsTxt } from './llms-txt.js';
+export { generateLlmsFullTxt } from './llms-full-txt.js';
+export { generateSitemapXml } from './sitemap-xml.js';
+export { generateRobotsTxt, AI_CRAWLERS } from './robots-txt.js';
+export { generateJsonLd } from './jsonld.js';
+export { generatePageMarkdown } from './page-markdown.js';
+//# sourceMappingURL=index.js.map

package/dist/ax/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/ax/index.ts"],"names":[],"mappings":"AAmBA,OAAO,EAAE,uBAAuB,EAAE,MAAM,qBAAqB,CAAA;AAC7D,OAAO,EAAE,cAAc,EAAE,WAAW,EAAE,MAAM,UAAU,CAAA;AACtD,OAAO,EAAE,cAAc,EAAE,SAAS,EAAE,MAAM,uBAAuB,CAAA;AACjE,OAAO,EAAE,UAAU,EAAE,QAAQ,EAAE,OAAO,EAAE,MAAM,cAAc,CAAA;AAC5D,OAAO,EAAE,YAAY,EAAE,WAAW,EAAE,gBAAgB,EAAE,oBAAoB,EAAE,MAAM,cAAc,CAAA;AAChG,OAAO,EAAE,eAAe,EAAE,MAAM,eAAe,CAAA;AAC/C,OAAO,EAAE,mBAAmB,EAAE,MAAM,oBAAoB,CAAA;AACxD,OAAO,EAAE,kBAAkB,EAAE,MAAM,kBAAkB,CAAA;AACrD,OAAO,EAAE,iBAAiB,EAAE,WAAW,EAAE,MAAM,iBAAiB,CAAA;AAChE,OAAO,EAAE,cAAc,EAAE,MAAM,aAAa,CAAA;AAC5C,OAAO,EAAE,oBAAoB,EAAE,MAAM,oBAAoB,CAAA"}

package/dist/ax/jsonld.d.ts

@@ -0,0 +1,18 @@
+import type { AxContentNode, AxRenderContext } from './types.js';
+/**
+ * Pure, config-driven content→schema.org JSON-LD mapping (Tier 1, deterministic).
+ *
+ * Resolution order for a node:
+ *   1. `mapping.transform(node, ctx)` — full escape hatch, wins over everything.
+ *   2. `mapping.schemaType` from `ax.jsonLd.types[contentTypeName]`, else a
+ *      built-in default keyed by baseType/name (WebPage / Article / Event /
+ *      FAQPage).
+ *   3. Built-in field defaults for the resolved type (unless
+ *      `ax.jsonLd.defaults === false`), then `mapping.fields` remaps applied on
+ *      top (schema.org prop -> content field name).
+ *
+ * Never calls a language model. FAQPage is a pure field read of a stored
+ * `[{ question, answer }]` array — see the FAQ-as-content-type design.
+ */
+export declare function generateJsonLd(node: AxContentNode, ctx: AxRenderContext): Record<string, unknown> | Record<string, unknown>[];
+//# sourceMappingURL=jsonld.d.ts.map

package/dist/ax/jsonld.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"jsonld.d.ts","sourceRoot":"","sources":["../../src/ax/jsonld.ts"],"names":[],"mappings":"AAKA,OAAO,KAAK,EAAE,aAAa,EAAE,eAAe,EAAE,MAAM,YAAY,CAAA;AAEhE;;;;;;;;;;;;;;GAcG;AACH,wBAAgB,cAAc,CAC5B,IAAI,EAAE,aAAa,EACnB,GAAG,EAAE,eAAe,GACnB,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,EAAE,CAqCrD"}

package/dist/ax/jsonld.js

@@ -0,0 +1,173 @@
+import { resolveAxConfig } from '../config/define.js';
+import { absoluteUrl } from './url.js';
+import { htmlToMarkdown } from './html-to-markdown.js';
+import { truncate } from './summary.js';
+/**
+ * Pure, config-driven content→schema.org JSON-LD mapping (Tier 1, deterministic).
+ *
+ * Resolution order for a node:
+ *   1. `mapping.transform(node, ctx)` — full escape hatch, wins over everything.
+ *   2. `mapping.schemaType` from `ax.jsonLd.types[contentTypeName]`, else a
+ *      built-in default keyed by baseType/name (WebPage / Article / Event /
+ *      FAQPage).
+ *   3. Built-in field defaults for the resolved type (unless
+ *      `ax.jsonLd.defaults === false`), then `mapping.fields` remaps applied on
+ *      top (schema.org prop -> content field name).
+ *
+ * Never calls a language model. FAQPage is a pure field read of a stored
+ * `[{ question, answer }]` array — see the FAQ-as-content-type design.
+ */
+export function generateJsonLd(node, ctx) {
+    const ax = resolveAxConfig(ctx.config.ax);
+    const mapping = ax.jsonLd?.types?.[node.contentTypeName];
+    // 1. Escape hatch.
+    if (mapping?.transform)
+        return mapping.transform(node, ctx);
+    const url = absoluteUrl(ctx.baseUrl, node.path);
+    const schemaType = mapping?.schemaType ?? defaultSchemaType(node);
+    const applyDefaults = ax.jsonLd?.defaults !== false;
+    // FAQPage is structurally different (a mainEntity array), handle it first.
+    if (schemaType === 'FAQPage') {
+        return buildFaqPage(node, mapping, url);
+    }
+    const name = str(node.metaTitle) || str(node.title);
+    const obj = {
+        '@context': 'https://schema.org',
+        '@type': schemaType,
+        url,
+        name,
+    };
+    if (applyDefaults)
+        applyTypeDefaults(obj, node, schemaType, ctx);
+    // 3. Field remap applied over defaults (schema.org prop -> content field name).
+    if (mapping?.fields) {
+        for (const [prop, fieldName] of Object.entries(mapping.fields)) {
+            const val = node[fieldName];
+            if (val !== undefined && val !== null && val !== '') {
+                obj[prop] = normalizeValue(val);
+            }
+        }
+    }
+    return obj;
+}
+// ─── Built-in type resolution ─────────────────────────────────────────────────
+const ARTICLE_NAME_RE = /(blog|article|news|post)/i;
+function defaultSchemaType(node) {
+    const name = node.contentTypeName ?? '';
+    if (node.baseType === 'event')
+        return 'Event';
+    if (/faq/i.test(name))
+        return 'FAQPage';
+    if (ARTICLE_NAME_RE.test(name))
+        return 'Article';
+    if (node.baseType === 'content')
+        return 'WebPage';
+    return 'WebPage';
+}
+function applyTypeDefaults(obj, node, schemaType, ctx) {
+    const description = str(node.metaDescription) || str(node.summary);
+    const published = toIso(node.publishedAt);
+    const modified = toIso(node.updatedAt);
+    const imageUrl = imageFor(node, ctx.baseUrl);
+    if (schemaType === 'Article') {
+        obj['headline'] = str(node.metaTitle) || str(node.title);
+        if (description)
+            obj['description'] = description;
+        if (published)
+            obj['datePublished'] = published;
+        if (modified)
+            obj['dateModified'] = modified;
+        if (imageUrl)
+            obj['image'] = imageUrl;
+        const body = htmlToMarkdown(str(node.body));
+        if (body)
+            obj['articleBody'] = truncate(body, 5000);
+        const author = str(node['author']);
+        if (author)
+            obj['author'] = { '@type': 'Person', name: author };
+        return;
+    }
+    if (schemaType === 'Event') {
+        if (description)
+            obj['description'] = description;
+        const start = toIso(node['startDate']);
+        const end = toIso(node['endDate']);
+        if (start)
+            obj['startDate'] = start;
+        if (end)
+            obj['endDate'] = end;
+        const location = str(node['location']);
+        if (location)
+            obj['location'] = { '@type': 'Place', name: location };
+        if (imageUrl)
+            obj['image'] = imageUrl;
+        return;
+    }
+    // WebPage (and any unmapped default).
+    if (description)
+        obj['description'] = description;
+    if (published)
+        obj['datePublished'] = published;
+    if (modified)
+        obj['dateModified'] = modified;
+    if (imageUrl)
+        obj['image'] = imageUrl;
+}
+function buildFaqPage(node, mapping, url) {
+    const sourceField = mapping?.faq?.source ?? 'questions';
+    const raw = node[sourceField];
+    const pairs = coerceQaPairs(raw);
+    return {
+        '@context': 'https://schema.org',
+        '@type': 'FAQPage',
+        url,
+        name: str(node.metaTitle) || str(node.title),
+        mainEntity: pairs.map((qa) => ({
+            '@type': 'Question',
+            name: qa.question,
+            acceptedAnswer: { '@type': 'Answer', text: qa.answer },
+        })),
+    };
+}
+function coerceQaPairs(raw) {
+    if (!Array.isArray(raw))
+        return [];
+    const out = [];
+    for (const item of raw) {
+        if (item && typeof item === 'object') {
+            const rec = item;
+            const question = str(rec['question'] ?? rec['q'] ?? rec['title']);
+            const answer = str(rec['answer'] ?? rec['a'] ?? rec['text']);
+            if (question && answer)
+                out.push({ question, answer });
+        }
+    }
+    return out;
+}
+// ─── Helpers ──────────────────────────────────────────────────────────────────
+/**
+ * Reference-app convention for the media file URL. Decoupled/headless
+ * deployments can override the image via a `fields.image` mapping that points at
+ * a stored absolute-URL field.
+ */
+function imageFor(node, baseUrl) {
+    const id = str(node.featuredImageId) || str(node.metaImageId);
+    if (!id)
+        return undefined;
+    return absoluteUrl(baseUrl, `/api/v1/media/${id}/file`);
+}
+function str(v) {
+    return typeof v === 'string' ? v.trim() : '';
+}
+function toIso(v) {
+    if (!v)
+        return undefined;
+    const d = v instanceof Date ? v : new Date(String(v));
+    return Number.isNaN(d.getTime()) ? undefined : d.toISOString();
+}
+function normalizeValue(v) {
+    if (v instanceof Date)
+        return v.toISOString();
+    return v;
+}
+//# sourceMappingURL=jsonld.js.map

package/dist/ax/jsonld.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"jsonld.js","sourceRoot":"","sources":["../../src/ax/jsonld.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,eAAe,EAAE,MAAM,qBAAqB,CAAA;AAErD,OAAO,EAAE,WAAW,EAAE,MAAM,UAAU,CAAA;AACtC,OAAO,EAAE,cAAc,EAAE,MAAM,uBAAuB,CAAA;AACtD,OAAO,EAAE,QAAQ,EAAE,MAAM,cAAc,CAAA;AAGvC;;;;;;;;;;;;;;GAcG;AACH,MAAM,UAAU,cAAc,CAC5B,IAAmB,EACnB,GAAoB;IAEpB,MAAM,EAAE,GAAG,eAAe,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC,CAAA;IACzC,MAAM,OAAO,GAAG,EAAE,CAAC,MAAM,EAAE,KAAK,EAAE,CAAC,IAAI,CAAC,eAAe,CAAC,CAAA;IAExD,mBAAmB;IACnB,IAAI,OAAO,EAAE,SAAS;QAAE,OAAO,OAAO,CAAC,SAAS,CAAC,IAAI,EAAE,GAAG,CAAC,CAAA;IAE3D,MAAM,GAAG,GAAG,WAAW,CAAC,GAAG,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,CAAC,CAAA;IAC/C,MAAM,UAAU,GAAG,OAAO,EAAE,UAAU,IAAI,iBAAiB,CAAC,IAAI,CAAC,CAAA;IACjE,MAAM,aAAa,GAAG,EAAE,CAAC,MAAM,EAAE,QAAQ,KAAK,KAAK,CAAA;IAEnD,2EAA2E;IAC3E,IAAI,UAAU,KAAK,SAAS,EAAE,CAAC;QAC7B,OAAO,YAAY,CAAC,IAAI,EAAE,OAAO,EAAE,GAAG,CAAC,CAAA;IACzC,CAAC;IAED,MAAM,IAAI,GAAG,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;IACnD,MAAM,GAAG,GAA4B;QACnC,UAAU,EAAE,oBAAoB;QAChC,OAAO,EAAE,UAAU;QACnB,GAAG;QACH,IAAI;KACL,CAAA;IAED,IAAI,aAAa;QAAE,iBAAiB,CAAC,GAAG,EAAE,IAAI,EAAE,UAAU,EAAE,GAAG,CAAC,CAAA;IAEhE,gFAAgF;IAChF,IAAI,OAAO,EAAE,MAAM,EAAE,CAAC;QACpB,KAAK,MAAM,CAAC,IAAI,EAAE,SAAS,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,CAAC;YAC/D,MAAM,GAAG,GAAG,IAAI,CAAC,SAAS,CAAC,CAAA;YAC3B,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,IAAI,IAAI,GAAG,KAAK,EAAE,EAAE,CAAC;gBACpD,GAAG,CAAC,IAAI,CAAC,GAAG,cAAc,CAAC,GAAG,CAAC,CAAA;YACjC,CAAC;QACH,CAAC;IACH,CAAC;IAED,OAAO,GAAG,CAAA;AACZ,CAAC;AAED,iFAAiF;AAEjF,MAAM,eAAe,GAAG,2BAA2B,CAAA;AAEnD,SAAS,iBAAiB,CAAC,IAAmB;IAC5C,MAAM,IAAI,GAAG,IAAI,CAAC,eAAe,IAAI,EAAE,CAAA;IACvC,IAAI,IAAI,CAAC,QAAQ,KAAK,OAAO;QAAE,OAAO,OAAO,CAAA;IAC7C,IAAI,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC;QAAE,OAAO,SAAS,CAAA;IACvC,IAAI,eAAe,CAAC,IAAI,CAAC,IAAI,CAAC;QAAE,OAAO,SAAS,CAAA;IAChD,IAAI,IAAI,CAAC,QAAQ,KAAK,SAAS;QAAE,OAAO,SAAS,CAAA;IACjD,OAAO,SAAS,CAAA;AAClB,CAAC;AAED,SAAS,iBAAiB,CACxB,GAA4B,EAC5B,IAAmB,EACnB,UAAkB,EAClB,GAAoB;IAEpB,MAAM,WAAW,GAAG,GAAG,CAAC,IAAI,CAAC,eAAe,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;IAClE,MAAM,SAAS,GAAG,KAAK,CAAC,IAAI,CAAC,WAAW,CAAC,CAAA;IACzC,MAAM,QAAQ,GAAG,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC,CAAA;IACtC,MAAM,QAAQ,GAAG,QAAQ,CAAC,IAAI,EAAE,GAAG,CAAC,OAAO,CAAC,CAAA;IAE5C,IAAI,UAAU,KAAK,SAAS,EAAE,CAAC;QAC7B,GAAG,CAAC,UAAU,CAAC,GAAG,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;QACxD,IAAI,WAAW;YAAE,GAAG,CAAC,aAAa,CAAC,GAAG,WAAW,CAAA;QACjD,IAAI,SAAS;YAAE,GAAG,CAAC,eAAe,CAAC,GAAG,SAAS,CAAA;QAC/C,IAAI,QAAQ;YAAE,GAAG,CAAC,cAAc,CAAC,GAAG,QAAQ,CAAA;QAC5C,IAAI,QAAQ;YAAE,GAAG,CAAC,OAAO,CAAC,GAAG,QAAQ,CAAA;QACrC,MAAM,IAAI,GAAG,cAAc,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAA;QAC3C,IAAI,IAAI;YAAE,GAAG,CAAC,aAAa,CAAC,GAAG,QAAQ,CAAC,IAAI,EAAE,IAAI,CAAC,CAAA;QACnD,MAAM,MAAM,GAAG,GAAG,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAA;QAClC,IAAI,MAAM;YAAE,GAAG,CAAC,QAAQ,CAAC,GAAG,EAAE,OAAO,EAAE,QAAQ,EAAE,IAAI,EAAE,MAAM,EAAE,CAAA;QAC/D,OAAM;IACR,CAAC;IAED,IAAI,UAAU,KAAK,OAAO,EAAE,CAAC;QAC3B,IAAI,WAAW;YAAE,GAAG,CAAC,aAAa,CAAC,GAAG,WAAW,CAAA;QACjD,MAAM,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC,CAAA;QACtC,MAAM,GAAG,GAAG,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,CAAA;QAClC,IAAI,KAAK;YAAE,GAAG,CAAC,WAAW,CAAC,GAAG,KAAK,CAAA;QACnC,IAAI,GAAG;YAAE,GAAG,CAAC,SAAS,CAAC,GAAG,GAAG,CAAA;QAC7B,MAAM,QAAQ,GAAG,GAAG,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC,CAAA;QACtC,IAAI,QAAQ;YAAE,GAAG,CAAC,UAAU,CAAC,GAAG,EAAE,OAAO,EAAE,OAAO,EAAE,IAAI,EAAE,QAAQ,EAAE,CAAA;QACpE,IAAI,QAAQ;YAAE,GAAG,CAAC,OAAO,CAAC,GAAG,QAAQ,CAAA;QACrC,OAAM;IACR,CAAC;IAED,sCAAsC;IACtC,IAAI,WAAW;QAAE,GAAG,CAAC,aAAa,CAAC,GAAG,WAAW,CAAA;IACjD,IAAI,SAAS;QAAE,GAAG,CAAC,eAAe,CAAC,GAAG,SAAS,CAAA;IAC/C,IAAI,QAAQ;QAAE,GAAG,CAAC,cAAc,CAAC,GAAG,QAAQ,CAAA;IAC5C,IAAI,QAAQ;QAAE,GAAG,CAAC,OAAO,CAAC,GAAG,QAAQ,CAAA;AACvC,CAAC;AASD,SAAS,YAAY,CACnB,IAAmB,EACnB,OAAsC,EACtC,GAAW;IAEX,MAAM,WAAW,GAAG,OAAO,EAAE,GAAG,EAAE,MAAM,IAAI,WAAW,CAAA;IACvD,MAAM,GAAG,GAAG,IAAI,CAAC,WAAW,CAAC,CAAA;IAC7B,MAAM,KAAK,GAAG,aAAa,CAAC,GAAG,CAAC,CAAA;IAChC,OAAO;QACL,UAAU,EAAE,oBAAoB;QAChC,OAAO,EAAE,SAAS;QAClB,GAAG;QACH,IAAI,EAAE,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC;QAC5C,UAAU,EAAE,KAAK,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC;YAC7B,OAAO,EAAE,UAAU;YACnB,IAAI,EAAE,EAAE,CAAC,QAAQ;YACjB,cAAc,EAAE,EAAE,OAAO,EAAE,QAAQ,EAAE,IAAI,EAAE,EAAE,CAAC,MAAM,EAAE;SACvD,CAAC,CAAC;KACJ,CAAA;AACH,CAAC;AAED,SAAS,aAAa,CAAC,GAAY;IACjC,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,GAAG,CAAC;QAAE,OAAO,EAAE,CAAA;IAClC,MAAM,GAAG,GAAa,EAAE,CAAA;IACxB,KAAK,MAAM,IAAI,IAAI,GAAG,EAAE,CAAC;QACvB,IAAI,IAAI,IAAI,OAAO,IAAI,KAAK,QAAQ,EAAE,CAAC;YACrC,MAAM,GAAG,GAAG,IAA+B,CAAA;YAC3C,MAAM,QAAQ,GAAG,GAAG,CAAC,GAAG,CAAC,UAAU,CAAC,IAAI,GAAG,CAAC,GAAG,CAAC,IAAI,GAAG,CAAC,OAAO,CAAC,CAAC,CAAA;YACjE,MAAM,MAAM,GAAG,GAAG,CAAC,GAAG,CAAC,QAAQ,CAAC,IAAI,GAAG,CAAC,GAAG,CAAC,IAAI,GAAG,CAAC,MAAM,CAAC,CAAC,CAAA;YAC5D,IAAI,QAAQ,IAAI,MAAM;gBAAE,GAAG,CAAC,IAAI,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,CAAC,CAAA;QACxD,CAAC;IACH,CAAC;IACD,OAAO,GAAG,CAAA;AACZ,CAAC;AAED,iFAAiF;AAEjF;;;;GAIG;AACH,SAAS,QAAQ,CAAC,IAAmB,EAAE,OAAe;IACpD,MAAM,EAAE,GAAG,GAAG,CAAC,IAAI,CAAC,eAAe,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,WAAW,CAAC,CAAA;IAC7D,IAAI,CAAC,EAAE;QAAE,OAAO,SAAS,CAAA;IACzB,OAAO,WAAW,CAAC,OAAO,EAAE,iBAAiB,EAAE,OAAO,CAAC,CAAA;AACzD,CAAC;AAED,SAAS,GAAG,CAAC,CAAU;IACrB,OAAO,OAAO,CAAC,KAAK,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,EAAE,CAAA;AAC9C,CAAC;AAED,SAAS,KAAK,CAAC,CAAU;IACvB,IAAI,CAAC,CAAC;QAAE,OAAO,SAAS,CAAA;IACxB,MAAM,CAAC,GAAG,CAAC,YAAY,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAA;IACrD,OAAO,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,WAAW,EAAE,CAAA;AAChE,CAAC;AAED,SAAS,cAAc,CAAC,CAAU;IAChC,IAAI,CAAC,YAAY,IAAI;QAAE,OAAO,CAAC,CAAC,WAAW,EAAE,CAAA;IAC7C,OAAO,CAAC,CAAA;AACV,CAAC"}

package/dist/ax/llms-full-txt.d.ts

@@ -0,0 +1,11 @@
+import type { AxContentSource, AxRenderContext } from './types.js';
+/**
+ * Generate `llms-full.txt` — the full-text agent corpus: a deterministic
+ * HTML→Markdown projection of every AX-eligible published node's `body`.
+ *
+ * Preamble matches the llms.txt header, then one `## {title}` section per node
+ * (ordered by path), each with a URL/Type/Updated header line and the
+ * Markdown-rendered body. Drafts and `excludeFromSearch` nodes never appear.
+ */
+export declare function generateLlmsFullTxt(source: AxContentSource, ctx: AxRenderContext): Promise<string>;
+//# sourceMappingURL=llms-full-txt.d.ts.map

package/dist/ax/llms-full-txt.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"llms-full-txt.d.ts","sourceRoot":"","sources":["../../src/ax/llms-full-txt.ts"],"names":[],"mappings":"AAIA,OAAO,KAAK,EAAiB,eAAe,EAAE,eAAe,EAAE,MAAM,YAAY,CAAA;AAEjF;;;;;;;GAOG;AACH,wBAAsB,mBAAmB,CAAC,MAAM,EAAE,eAAe,EAAE,GAAG,EAAE,eAAe,GAAG,OAAO,CAAC,MAAM,CAAC,CA4BxG"}

package/dist/ax/llms-full-txt.js

@@ -0,0 +1,60 @@
+import { resolveAxConfig } from '../config/define.js';
+import { absoluteUrl } from './url.js';
+import { htmlToMarkdown } from './html-to-markdown.js';
+import { isAxEligible } from './filters.js';
+/**
+ * Generate `llms-full.txt` — the full-text agent corpus: a deterministic
+ * HTML→Markdown projection of every AX-eligible published node's `body`.
+ *
+ * Preamble matches the llms.txt header, then one `## {title}` section per node
+ * (ordered by path), each with a URL/Type/Updated header line and the
+ * Markdown-rendered body. Drafts and `excludeFromSearch` nodes never appear.
+ */
+export async function generateLlmsFullTxt(source, ctx) {
+    const ax = resolveAxConfig(ctx.config.ax);
+    const opts = {
+        ...(ax.llmsFullTxt?.includeTypes !== undefined ? { includeTypes: ax.llmsFullTxt.includeTypes } : {}),
+        ...(ax.llmsFullTxt?.excludeTypes !== undefined ? { excludeTypes: ax.llmsFullTxt.excludeTypes } : {}),
+    };
+    const maxItems = ax.llmsFullTxt?.maxItems;
+    const [nodes, site] = await Promise.all([
+        source.listForSite(ctx.siteId),
+        source.getSiteMeta(ctx.siteId),
+    ]);
+    let eligible = nodes.filter(n => isAxEligible(n, opts));
+    if (typeof maxItems === 'number' && maxItems >= 0) {
+        eligible = eligible.slice(0, maxItems);
+    }
+    const parts = [];
+    parts.push(`# ${site?.name ?? 'Site'}`);
+    const intro = ax.llmsTxt?.intro ?? site?.description;
+    if (intro)
+        parts.push(`> ${intro}`);
+    for (const node of eligible) {
+        parts.push(section(node, ctx.baseUrl));
+    }
+    return `${parts.join('\n\n')}\n`;
+}
+function section(node, baseUrl) {
+    const url = absoluteUrl(baseUrl, node.path);
+    const title = (node.title ?? '').trim() || node.slug;
+    const updated = toIso(node.updatedAt ?? node.publishedAt);
+    const header = [
+        `## ${title}`,
+        `URL: ${url}`,
+        `Type: ${node.contentTypeName}${updated ? `   Updated: ${updated}` : ''}`,
+    ].join('\n');
+    const bodySource = strOrEmpty(node.body) || strOrEmpty(node.summary);
+    const markdown = htmlToMarkdown(bodySource);
+    return markdown ? `${header}\n\n${markdown}` : header;
+}
+function strOrEmpty(v) {
+    return typeof v === 'string' ? v : '';
+}
+function toIso(v) {
+    if (!v)
+        return '';
+    const d = v instanceof Date ? v : new Date(String(v));
+    return Number.isNaN(d.getTime()) ? '' : d.toISOString();
+}
+//# sourceMappingURL=llms-full-txt.js.map

package/dist/ax/llms-full-txt.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"llms-full-txt.js","sourceRoot":"","sources":["../../src/ax/llms-full-txt.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,eAAe,EAAE,MAAM,qBAAqB,CAAA;AACrD,OAAO,EAAE,WAAW,EAAE,MAAM,UAAU,CAAA;AACtC,OAAO,EAAE,cAAc,EAAE,MAAM,uBAAuB,CAAA;AACtD,OAAO,EAAE,YAAY,EAAE,MAAM,cAAc,CAAA;AAG3C;;;;;;;GAOG;AACH,MAAM,CAAC,KAAK,UAAU,mBAAmB,CAAC,MAAuB,EAAE,GAAoB;IACrF,MAAM,EAAE,GAAG,eAAe,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC,CAAA;IACzC,MAAM,IAAI,GAAG;QACX,GAAG,CAAC,EAAE,CAAC,WAAW,EAAE,YAAY,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,YAAY,EAAE,EAAE,CAAC,WAAW,CAAC,YAAY,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QACpG,GAAG,CAAC,EAAE,CAAC,WAAW,EAAE,YAAY,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,YAAY,EAAE,EAAE,CAAC,WAAW,CAAC,YAAY,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;KACrG,CAAA;IACD,MAAM,QAAQ,GAAG,EAAE,CAAC,WAAW,EAAE,QAAQ,CAAA;IAEzC,MAAM,CAAC,KAAK,EAAE,IAAI,CAAC,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC;QACtC,MAAM,CAAC,WAAW,CAAC,GAAG,CAAC,MAAM,CAAC;QAC9B,MAAM,CAAC,WAAW,CAAC,GAAG,CAAC,MAAM,CAAC;KAC/B,CAAC,CAAA;IAEF,IAAI,QAAQ,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,YAAY,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,CAAA;IACvD,IAAI,OAAO,QAAQ,KAAK,QAAQ,IAAI,QAAQ,IAAI,CAAC,EAAE,CAAC;QAClD,QAAQ,GAAG,QAAQ,CAAC,KAAK,CAAC,CAAC,EAAE,QAAQ,CAAC,CAAA;IACxC,CAAC;IAED,MAAM,KAAK,GAAa,EAAE,CAAA;IAC1B,KAAK,CAAC,IAAI,CAAC,KAAK,IAAI,EAAE,IAAI,IAAI,MAAM,EAAE,CAAC,CAAA;IACvC,MAAM,KAAK,GAAG,EAAE,CAAC,OAAO,EAAE,KAAK,IAAI,IAAI,EAAE,WAAW,CAAA;IACpD,IAAI,KAAK;QAAE,KAAK,CAAC,IAAI,CAAC,KAAK,KAAK,EAAE,CAAC,CAAA;IAEnC,KAAK,MAAM,IAAI,IAAI,QAAQ,EAAE,CAAC;QAC5B,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,EAAE,GAAG,CAAC,OAAO,CAAC,CAAC,CAAA;IACxC,CAAC;IAED,OAAO,GAAG,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAA;AAClC,CAAC;AAED,SAAS,OAAO,CAAC,IAAmB,EAAE,OAAe;IACnD,MAAM,GAAG,GAAG,WAAW,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,CAAC,CAAA;IAC3C,MAAM,KAAK,GAAG,CAAC,IAAI,CAAC,KAAK,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,IAAI,IAAI,CAAC,IAAI,CAAA;IACpD,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,CAAC,SAAS,IAAI,IAAI,CAAC,WAAW,CAAC,CAAA;IACzD,MAAM,MAAM,GAAG;QACb,MAAM,KAAK,EAAE;QACb,QAAQ,GAAG,EAAE;QACb,SAAS,IAAI,CAAC,eAAe,GAAG,OAAO,CAAC,CAAC,CAAC,eAAe,OAAO,EAAE,CAAC,CAAC,CAAC,EAAE,EAAE;KAC1E,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;IACZ,MAAM,UAAU,GAAG,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,UAAU,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;IACpE,MAAM,QAAQ,GAAG,cAAc,CAAC,UAAU,CAAC,CAAA;IAC3C,OAAO,QAAQ,CAAC,CAAC,CAAC,GAAG,MAAM,OAAO,QAAQ,EAAE,CAAC,CAAC,CAAC,MAAM,CAAA;AACvD,CAAC;AAED,SAAS,UAAU,CAAC,CAAU;IAC5B,OAAO,OAAO,CAAC,KAAK,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAA;AACvC,CAAC;AAED,SAAS,KAAK,CAAC,CAAU;IACvB,IAAI,CAAC,CAAC;QAAE,OAAO,EAAE,CAAA;IACjB,MAAM,CAAC,GAAG,CAAC,YAAY,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAA;IACrD,OAAO,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,WAAW,EAAE,CAAA;AACzD,CAAC"}