cli-jaw 2.0.5 → 2.0.6

package/dist/src/browser/adaptive-fetch/metadata.js

@@ -0,0 +1,169 @@
+// @ts-nocheck
+// Mirrored from agbrowse adaptive-fetch v2; keep runtime behavior aligned while cli-jaw mirror remains experimental.
+import { extractTitleFromHtml, htmlToReadableText, normalizeWhitespace } from './transforms.js';
+/**
+ * @param {string} html
+ * @param {string} finalUrl
+ */
+export function extractMetadataFromHtml(html = '', finalUrl = '') {
+    const title = firstNonEmpty(getMetaContent(html, 'property', 'og:title'), getMetaContent(html, 'name', 'twitter:title'), extractTitleFromHtml(html));
+    const description = firstNonEmpty(getMetaContent(html, 'name', 'description'), getMetaContent(html, 'property', 'og:description'), getMetaContent(html, 'name', 'twitter:description'));
+    const canonicalUrl = resolveMaybeUrl(getLinkHref(html, 'canonical'), finalUrl);
+    const feedUrls = extractFeedUrls(html, finalUrl);
+    const oEmbedUrls = extractOembedUrls(html, finalUrl);
+    const jsonLd = extractJsonLdBlocks(html);
+    const text = htmlToReadableText(html);
+    return {
+        source: 'metadata',
+        finalUrl,
+        title,
+        text,
+        metadata: {
+            canonicalUrl,
+            description,
+            feedUrls,
+            oEmbedUrls,
+            openGraph: extractOpenGraph(html),
+            jsonLd,
+        },
+        evidence: [
+            title ? 'title' : null,
+            description ? 'description' : null,
+            canonicalUrl ? 'canonical' : null,
+            feedUrls.length > 0 ? 'feed-link' : null,
+            oEmbedUrls.length > 0 ? 'oembed-link' : null,
+            jsonLd.length > 0 ? 'json-ld' : null,
+        ].filter(Boolean),
+        warnings: [],
+    };
+}
+/**
+ * @param {string} html
+ */
+export function extractJsonLdBlocks(html = '') {
+    const blocks = [];
+    const re = /<script[^>]+type=["']application\/ld\+json["'][^>]*>([\s\S]*?)<\/script>/gi;
+    let match;
+    while ((match = re.exec(html))) {
+        const raw = match[1].trim();
+        if (!raw)
+            continue;
+        try {
+            blocks.push(JSON.parse(raw));
+        }
+        catch {
+            blocks.push({ raw, parseError: true });
+        }
+    }
+    return blocks;
+}
+/**
+ * @param {string} html
+ */
+function extractOpenGraph(html) {
+    /** @type {Record<string, string>} */
+    const og = {};
+    const re = /<meta\s+[^>]*property=["']og:([^"']+)["'][^>]*content=["']([^"']*)["'][^>]*>/gi;
+    let match;
+    while ((match = re.exec(html)))
+        og[match[1]] = normalizeWhitespace(match[2]);
+    return og;
+}
+/**
+ * @param {string} html
+ * @param {string} base
+ */
+export function extractFeedUrls(html = '', base = '') {
+    const urls = [];
+    const re = /<link\b[^>]*>/gi;
+    let match;
+    while ((match = re.exec(html))) {
+        const tag = match[0];
+        const rel = getTagAttr(tag, 'rel').toLowerCase();
+        const type = getTagAttr(tag, 'type').toLowerCase();
+        const href = getTagAttr(tag, 'href');
+        if (!href || !/\balternate\b/.test(rel))
+            continue;
+        if (!/(application\/rss\+xml|application\/atom\+xml|application\/feed\+json|text\/xml|application\/xml)/i.test(type))
+            continue;
+        const resolved = resolveMaybeUrl(href, base);
+        if (resolved && !urls.includes(resolved))
+            urls.push(resolved);
+    }
+    return urls;
+}
+/**
+ * @param {string} html
+ * @param {string} base
+ */
+export function extractOembedUrls(html = '', base = '') {
+    const urls = [];
+    const re = /<link\b[^>]*>/gi;
+    let match;
+    while ((match = re.exec(html))) {
+        const tag = match[0];
+        const rel = getTagAttr(tag, 'rel').toLowerCase();
+        const type = getTagAttr(tag, 'type').toLowerCase();
+        const href = getTagAttr(tag, 'href');
+        if (!href || !/\balternate\b/.test(rel))
+            continue;
+        if (!/(application\/json\+oembed|text\/xml\+oembed|application\/xml\+oembed)/i.test(type))
+            continue;
+        const resolved = resolveMaybeUrl(href, base);
+        if (resolved && !urls.includes(resolved))
+            urls.push(resolved);
+    }
+    return urls;
+}
+/**
+ * @param {string} html
+ * @param {'name'|'property'} attr
+ * @param {string} key
+ */
+function getMetaContent(html, attr, key) {
+    const re = new RegExp(`<meta\\s+[^>]*${attr}=["']${escapeRegExp(key)}["'][^>]*content=["']([^"']*)["'][^>]*>`, 'i');
+    const match = html.match(re);
+    return match ? normalizeWhitespace(match[1]) : '';
+}
+/**
+ * @param {string} html
+ * @param {string} rel
+ */
+function getLinkHref(html, rel) {
+    const re = new RegExp(`<link\\s+[^>]*rel=["']${escapeRegExp(rel)}["'][^>]*href=["']([^"']*)["'][^>]*>`, 'i');
+    const match = html.match(re);
+    return match ? normalizeWhitespace(match[1]) : '';
+}
+/**
+ * @param {string} tag
+ * @param {string} attr
+ */
+function getTagAttr(tag, attr) {
+    const re = new RegExp(`\\b${escapeRegExp(attr)}=["']([^"']*)["']`, 'i');
+    const match = tag.match(re);
+    return match ? normalizeWhitespace(match[1]) : '';
+}
+/**
+ * @param {string} raw
+ * @param {string} base
+ */
+function resolveMaybeUrl(raw, base) {
+    if (!raw)
+        return '';
+    try {
+        return new URL(raw, base || undefined).href;
+    }
+    catch {
+        return raw;
+    }
+}
+function firstNonEmpty(...values) {
+    return values.find(v => typeof v === 'string' && v.trim()) || '';
+}
+/**
+ * @param {string} text
+ */
+function escapeRegExp(text) {
+    return text.replace(/[.*+?^${}()|[\]\\]/g, '\\$&');
+}
+//# sourceMappingURL=metadata.js.map

package/dist/src/browser/adaptive-fetch/metadata.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"metadata.js","sourceRoot":"","sources":["../../../../src/browser/adaptive-fetch/metadata.ts"],"names":[],"mappings":"AAAA,cAAc;AACd,qHAAqH;AAErH,OAAO,EAAE,oBAAoB,EAAE,kBAAkB,EAAE,mBAAmB,EAAE,MAAM,iBAAiB,CAAC;AAEhG;;;GAGG;AACH,MAAM,UAAU,uBAAuB,CAAC,IAAI,GAAG,EAAE,EAAE,QAAQ,GAAG,EAAE;IAC5D,MAAM,KAAK,GAAG,aAAa,CACvB,cAAc,CAAC,IAAI,EAAE,UAAU,EAAE,UAAU,CAAC,EAC5C,cAAc,CAAC,IAAI,EAAE,MAAM,EAAE,eAAe,CAAC,EAC7C,oBAAoB,CAAC,IAAI,CAAC,CAC7B,CAAC;IACF,MAAM,WAAW,GAAG,aAAa,CAC7B,cAAc,CAAC,IAAI,EAAE,MAAM,EAAE,aAAa,CAAC,EAC3C,cAAc,CAAC,IAAI,EAAE,UAAU,EAAE,gBAAgB,CAAC,EAClD,cAAc,CAAC,IAAI,EAAE,MAAM,EAAE,qBAAqB,CAAC,CACtD,CAAC;IACF,MAAM,YAAY,GAAG,eAAe,CAAC,WAAW,CAAC,IAAI,EAAE,WAAW,CAAC,EAAE,QAAQ,CAAC,CAAC;IAC/E,MAAM,QAAQ,GAAG,eAAe,CAAC,IAAI,EAAE,QAAQ,CAAC,CAAC;IACjD,MAAM,UAAU,GAAG,iBAAiB,CAAC,IAAI,EAAE,QAAQ,CAAC,CAAC;IACrD,MAAM,MAAM,GAAG,mBAAmB,CAAC,IAAI,CAAC,CAAC;IACzC,MAAM,IAAI,GAAG,kBAAkB,CAAC,IAAI,CAAC,CAAC;IACtC,OAAO;QACH,MAAM,EAAE,UAAU;QAClB,QAAQ;QACR,KAAK;QACL,IAAI;QACJ,QAAQ,EAAE;YACN,YAAY;YACZ,WAAW;YACX,QAAQ;YACR,UAAU;YACV,SAAS,EAAE,gBAAgB,CAAC,IAAI,CAAC;YACjC,MAAM;SACT;QACD,QAAQ,EAAE;YACN,KAAK,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI;YACtB,WAAW,CAAC,CAAC,CAAC,aAAa,CAAC,CAAC,CAAC,IAAI;YAClC,YAAY,CAAC,CAAC,CAAC,WAAW,CAAC,CAAC,CAAC,IAAI;YACjC,QAAQ,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,WAAW,CAAC,CAAC,CAAC,IAAI;YACxC,UAAU,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,aAAa,CAAC,CAAC,CAAC,IAAI;YAC5C,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI;SACvC,CAAC,MAAM,CAAC,OAAO,CAAC;QACjB,QAAQ,EAAE,EAAE;KACf,CAAC;AACN,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,mBAAmB,CAAC,IAAI,GAAG,EAAE;IACzC,MAAM,MAAM,GAAG,EAAE,CAAC;IAClB,MAAM,EAAE,GAAG,4EAA4E,CAAC;IACxF,IAAI,KAAK,CAAC;IACV,OAAO,CAAC,KAAK,GAAG,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,EAAE,CAAC;QAC7B,MAAM,GAAG,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QAC5B,IAAI,CAAC,GAAG;YAAE,SAAS;QACnB,IAAI,CAAC;YACD,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC;QACjC,CAAC;QAAC,MAAM,CAAC;YACL,MAAM,CAAC,IAAI,CAAC,EAAE,GAAG,EAAE,UAAU,EAAE,IAAI,EAAE,CAAC,CAAC;QAC3C,CAAC;IACL,CAAC;IACD,OAAO,MAAM,CAAC;AAClB,CAAC;AAED;;GAEG;AACH,SAAS,gBAAgB,CAAC,IAAI;IAC1B,qCAAqC;IACrC,MAAM,EAAE,GAAG,EAAE,CAAC;IACd,MAAM,EAAE,GAAG,gFAAgF,CAAC;IAC5F,IAAI,KAAK,CAAC;IACV,OAAO,CAAC,KAAK,GAAG,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAAE,EAAE,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,GAAG,mBAAmB,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;IAC7E,OAAO,EAAE,CAAC;AACd,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,eAAe,CAAC,IAAI,GAAG,EAAE,EAAE,IAAI,GAAG,EAAE;IAChD,MAAM,IAAI,GAAG,EAAE,CAAC;IAChB,MAAM,EAAE,GAAG,iBAAiB,CAAC;IAC7B,IAAI,KAAK,CAAC;IACV,OAAO,CAAC,KAAK,GAAG,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,EAAE,CAAC;QAC7B,MAAM,GAAG,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;QACrB,MAAM,GAAG,GAAG,UAAU,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC,WAAW,EAAE,CAAC;QACjD,MAAM,IAAI,GAAG,UAAU,CAAC,GAAG,EAAE,MAAM,CAAC,CAAC,WAAW,EAAE,CAAC;QACnD,MAAM,IAAI,GAAG,UAAU,CAAC,GAAG,EAAE,MAAM,CAAC,CAAC;QACrC,IAAI,CAAC,IAAI,IAAI,CAAC,eAAe,CAAC,IAAI,CAAC,GAAG,CAAC;YAAE,SAAS;QAClD,IAAI,CAAC,oGAAoG,CAAC,IAAI,CAAC,IAAI,CAAC;YAAE,SAAS;QAC/H,MAAM,QAAQ,GAAG,eAAe,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;QAC7C,IAAI,QAAQ,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC;YAAE,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;IAClE,CAAC;IACD,OAAO,IAAI,CAAC;AAChB,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,iBAAiB,CAAC,IAAI,GAAG,EAAE,EAAE,IAAI,GAAG,EAAE;IAClD,MAAM,IAAI,GAAG,EAAE,CAAC;IAChB,MAAM,EAAE,GAAG,iBAAiB,CAAC;IAC7B,IAAI,KAAK,CAAC;IACV,OAAO,CAAC,KAAK,GAAG,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,EAAE,CAAC;QAC7B,MAAM,GAAG,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;QACrB,MAAM,GAAG,GAAG,UAAU,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC,WAAW,EAAE,CAAC;QACjD,MAAM,IAAI,GAAG,UAAU,CAAC,GAAG,EAAE,MAAM,CAAC,CAAC,WAAW,EAAE,CAAC;QACnD,MAAM,IAAI,GAAG,UAAU,CAAC,GAAG,EAAE,MAAM,CAAC,CAAC;QACrC,IAAI,CAAC,IAAI,IAAI,CAAC,eAAe,CAAC,IAAI,CAAC,GAAG,CAAC;YAAE,SAAS;QAClD,IAAI,CAAC,yEAAyE,CAAC,IAAI,CAAC,IAAI,CAAC;YAAE,SAAS;QACpG,MAAM,QAAQ,GAAG,eAAe,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;QAC7C,IAAI,QAAQ,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC;YAAE,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;IAClE,CAAC;IACD,OAAO,IAAI,CAAC;AAChB,CAAC;AAED;;;;GAIG;AACH,SAAS,cAAc,CAAC,IAAI,EAAE,IAAI,EAAE,GAAG;IACnC,MAAM,EAAE,GAAG,IAAI,MAAM,CAAC,iBAAiB,IAAI,QAAQ,YAAY,CAAC,GAAG,CAAC,yCAAyC,EAAE,GAAG,CAAC,CAAC;IACpH,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;IAC7B,OAAO,KAAK,CAAC,CAAC,CAAC,mBAAmB,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;AACtD,CAAC;AAED;;;GAGG;AACH,SAAS,WAAW,CAAC,IAAI,EAAE,GAAG;IAC1B,MAAM,EAAE,GAAG,IAAI,MAAM,CAAC,yBAAyB,YAAY,CAAC,GAAG,CAAC,sCAAsC,EAAE,GAAG,CAAC,CAAC;IAC7G,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;IAC7B,OAAO,KAAK,CAAC,CAAC,CAAC,mBAAmB,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;AACtD,CAAC;AAED;;;GAGG;AACH,SAAS,UAAU,CAAC,GAAG,EAAE,IAAI;IACzB,MAAM,EAAE,GAAG,IAAI,MAAM,CAAC,MAAM,YAAY,CAAC,IAAI,CAAC,mBAAmB,EAAE,GAAG,CAAC,CAAC;IACxE,MAAM,KAAK,GAAG,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;IAC5B,OAAO,KAAK,CAAC,CAAC,CAAC,mBAAmB,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;AACtD,CAAC;AAED;;;GAGG;AACH,SAAS,eAAe,CAAC,GAAG,EAAE,IAAI;IAC9B,IAAI,CAAC,GAAG;QAAE,OAAO,EAAE,CAAC;IACpB,IAAI,CAAC;QACD,OAAO,IAAI,GAAG,CAAC,GAAG,EAAE,IAAI,IAAI,SAAS,CAAC,CAAC,IAAI,CAAC;IAChD,CAAC;IAAC,MAAM,CAAC;QACL,OAAO,GAAG,CAAC;IACf,CAAC;AACL,CAAC;AAED,SAAS,aAAa,CAAC,GAAG,MAAM;IAC5B,OAAO,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,OAAO,CAAC,KAAK,QAAQ,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,CAAC;AACrE,CAAC;AAED;;GAEG;AACH,SAAS,YAAY,CAAC,IAAI;IACtB,OAAO,IAAI,CAAC,OAAO,CAAC,qBAAqB,EAAE,MAAM,CAAC,CAAC;AACvD,CAAC"}

package/dist/src/browser/adaptive-fetch/output.js

@@ -0,0 +1,71 @@
+// @ts-nocheck
+// Mirrored from agbrowse adaptive-fetch v2; keep runtime behavior aligned while cli-jaw mirror remains experimental.
+export const DEFAULT_OUTPUT_CONTENT_BYTES = 64 * 1024;
+/**
+ * @param {Record<string, any>} result
+ * @param {{ contentLimitBytes?: number }} [options]
+ */
+export function compactAdaptiveFetchResult(result, options = {}) {
+    const limit = positiveInteger(options.contentLimitBytes, DEFAULT_OUTPUT_CONTENT_BYTES);
+    const compacted = truncateTextToUtf8Bytes(result.content || '', limit);
+    return {
+        ...result,
+        content: compacted.text,
+        contentBytes: compacted.bytes,
+        contentLimitBytes: compacted.limit,
+        contentTruncated: compacted.truncated,
+    };
+}
+/**
+ * @param {string} text
+ * @param {number} limit
+ */
+export function truncateTextToUtf8Bytes(text, limit) {
+    const value = String(text || '');
+    const safeLimit = positiveInteger(limit, DEFAULT_OUTPUT_CONTENT_BYTES);
+    const bytes = Buffer.byteLength(value, 'utf8');
+    if (bytes <= safeLimit) {
+        return { text: value, bytes, limit: safeLimit, truncated: false };
+    }
+    let truncated = value.slice(0, safeLimit);
+    while (truncated.length > 0 && Buffer.byteLength(truncated, 'utf8') > safeLimit) {
+        truncated = truncated.slice(0, -1);
+    }
+    return { text: truncated, bytes, limit: safeLimit, truncated: true };
+}
+/**
+ * @param {string} text
+ * @param {{ write: Function }} [stdout]
+ */
+export function writeStdoutLine(text, stdout = process.stdout) {
+    const chunk = text.endsWith('\n') ? text : `${text}\n`;
+    return new Promise((resolve, reject) => {
+        let settled = false;
+        const done = (error) => {
+            if (settled)
+                return;
+            settled = true;
+            if (error)
+                reject(error);
+            else
+                resolve(undefined);
+        };
+        try {
+            const accepted = stdout.write(chunk, done);
+            if (accepted && stdout !== process.stdout)
+                done();
+        }
+        catch (error) {
+            done(error);
+        }
+    });
+}
+/**
+ * @param {unknown} value
+ * @param {number} fallback
+ */
+function positiveInteger(value, fallback) {
+    const n = Number(value);
+    return Number.isFinite(n) && n > 0 ? Math.floor(n) : fallback;
+}
+//# sourceMappingURL=output.js.map

package/dist/src/browser/adaptive-fetch/output.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"output.js","sourceRoot":"","sources":["../../../../src/browser/adaptive-fetch/output.ts"],"names":[],"mappings":"AAAA,cAAc;AACd,qHAAqH;AAErH,MAAM,CAAC,MAAM,4BAA4B,GAAG,EAAE,GAAG,IAAI,CAAC;AAEtD;;;GAGG;AACH,MAAM,UAAU,0BAA0B,CAAC,MAAM,EAAE,OAAO,GAAG,EAAE;IAC3D,MAAM,KAAK,GAAG,eAAe,CAAC,OAAO,CAAC,iBAAiB,EAAE,4BAA4B,CAAC,CAAC;IACvF,MAAM,SAAS,GAAG,uBAAuB,CAAC,MAAM,CAAC,OAAO,IAAI,EAAE,EAAE,KAAK,CAAC,CAAC;IACvE,OAAO;QACH,GAAG,MAAM;QACT,OAAO,EAAE,SAAS,CAAC,IAAI;QACvB,YAAY,EAAE,SAAS,CAAC,KAAK;QAC7B,iBAAiB,EAAE,SAAS,CAAC,KAAK;QAClC,gBAAgB,EAAE,SAAS,CAAC,SAAS;KACxC,CAAC;AACN,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,uBAAuB,CAAC,IAAI,EAAE,KAAK;IAC/C,MAAM,KAAK,GAAG,MAAM,CAAC,IAAI,IAAI,EAAE,CAAC,CAAC;IACjC,MAAM,SAAS,GAAG,eAAe,CAAC,KAAK,EAAE,4BAA4B,CAAC,CAAC;IACvE,MAAM,KAAK,GAAG,MAAM,CAAC,UAAU,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC;IAC/C,IAAI,KAAK,IAAI,SAAS,EAAE,CAAC;QACrB,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,SAAS,EAAE,SAAS,EAAE,KAAK,EAAE,CAAC;IACtE,CAAC;IACD,IAAI,SAAS,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,SAAS,CAAC,CAAC;IAC1C,OAAO,SAAS,CAAC,MAAM,GAAG,CAAC,IAAI,MAAM,CAAC,UAAU,CAAC,SAAS,EAAE,MAAM,CAAC,GAAG,SAAS,EAAE,CAAC;QAC9E,SAAS,GAAG,SAAS,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;IACvC,CAAC;IACD,OAAO,EAAE,IAAI,EAAE,SAAS,EAAE,KAAK,EAAE,KAAK,EAAE,SAAS,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC;AACzE,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,eAAe,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAAC,MAAM;IACzD,MAAM,KAAK,GAAG,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,IAAI,IAAI,CAAC;IACvD,OAAO,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE;QACnC,IAAI,OAAO,GAAG,KAAK,CAAC;QACpB,MAAM,IAAI,GAAG,CAAC,KAAK,EAAE,EAAE;YACnB,IAAI,OAAO;gBAAE,OAAO;YACpB,OAAO,GAAG,IAAI,CAAC;YACf,IAAI,KAAK;gBAAE,MAAM,CAAC,KAAK,CAAC,CAAC;;gBACpB,OAAO,CAAC,SAAS,CAAC,CAAC;QAC5B,CAAC,CAAC;QACF,IAAI,CAAC;YACD,MAAM,QAAQ,GAAG,MAAM,CAAC,KAAK,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC;YAC3C,IAAI,QAAQ,IAAI,MAAM,KAAK,OAAO,CAAC,MAAM;gBAAE,IAAI,EAAE,CAAC;QACtD,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACb,IAAI,CAAC,KAAK,CAAC,CAAC;QAChB,CAAC;IACL,CAAC,CAAC,CAAC;AACP,CAAC;AAED;;;GAGG;AACH,SAAS,eAAe,CAAC,KAAK,EAAE,QAAQ;IACpC,MAAM,CAAC,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC;IACxB,OAAO,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC;AAClE,CAAC"}

package/dist/src/browser/adaptive-fetch/reader-adapters.js

@@ -0,0 +1,135 @@
+// @ts-nocheck
+// Mirrored from agbrowse adaptive-fetch v2; keep runtime behavior aligned while cli-jaw mirror remains experimental.
+import { extractMetadataFromHtml } from './metadata.js';
+import { htmlToReadableText, isHtmlContentType, normalizeWhitespace } from './transforms.js';
+/**
+ * @param {any} fetched
+ * @param {{ source?: string, label?: string }} [context]
+ */
+export function fromFetchResult(fetched, context = {}) {
+    const source = context.source || 'fetch';
+    const isHtml = isHtmlContentType(fetched.contentType || '');
+    const metadata = isHtml ? extractMetadataFromHtml(fetched.text || '', fetched.finalUrl || '') : null;
+    const text = isHtml ? metadata.text : normalizeWhitespace(fetched.text || '');
+    return {
+        source,
+        label: context.label || source,
+        finalUrl: fetched.finalUrl || '',
+        title: metadata?.title || '',
+        text,
+        contentType: fetched.contentType || '',
+        status: Number(fetched.status || 0),
+        ok: Boolean(fetched.ok),
+        metadata: metadata?.metadata || null,
+        evidence: [...(fetched.evidence || []), ...(metadata?.evidence || [])],
+        warnings: fetched.warnings || [],
+        rawTextLength: String(fetched.text || '').length,
+    };
+}
+/**
+ * @param {any} result
+ */
+export function fromMetadataResult(result) {
+    return normalizeReaderCandidate({
+        source: 'metadata',
+        label: 'metadata',
+        finalUrl: result.finalUrl,
+        title: result.title,
+        text: result.text,
+        contentType: 'text/html',
+        status: 200,
+        ok: true,
+        metadata: result.metadata || null,
+        evidence: result.evidence || [],
+        warnings: result.warnings || [],
+    });
+}
+/**
+ * @param {any} result
+ */
+export function fromPublicEndpointResult(result) {
+    return normalizeReaderCandidate({ ...result, source: 'public_endpoint', label: result.label || 'public_endpoint' });
+}
+/**
+ * @param {any} result
+ */
+export function fromBrowserResult(result) {
+    return normalizeReaderCandidate({ ...result, source: 'browser', label: result.label || 'browser' });
+}
+/**
+ * @param {any} result
+ */
+export function fromNetworkCandidate(result) {
+    return normalizeReaderCandidate({ ...result, source: 'network_api', label: result.label || 'network_api' });
+}
+/**
+ * @param {any[]} results
+ */
+export function normalizeReaderCandidates(results = []) {
+    return results.map(normalizeReaderCandidate).filter(candidate => candidate.finalUrl || candidate.text);
+}
+/**
+ * @param {any} result
+ */
+export function normalizeReaderCandidate(result = {}) {
+    return {
+        source: result.source || 'reader',
+        label: result.label || result.source || 'reader',
+        finalUrl: result.finalUrl || '',
+        title: normalizeWhitespace(result.title || ''),
+        text: normalizeWhitespace(result.text || ''),
+        contentType: result.contentType || '',
+        status: Number(result.status || 0),
+        ok: result.ok !== false,
+        metadata: result.metadata || null,
+        evidence: Array.isArray(result.evidence) ? result.evidence.filter(Boolean) : [],
+        warnings: Array.isArray(result.warnings) ? result.warnings.filter(Boolean) : [],
+        safetyFlags: Array.isArray(result.safetyFlags) ? result.safetyFlags : [],
+        rawTextLength: Number(result.rawTextLength || String(result.text || '').length),
+    };
+}
+/**
+ * @param {any} result
+ */
+export function fromUserSessionResult(result) {
+    return normalizeReaderCandidate({
+        ...result,
+        source: 'browser_user',
+        label: result.label || 'browser-user-session',
+        safetyFlags: [...(result.safetyFlags || [])],
+    });
+}
+/**
+ * @param {any} result
+ */
+export function fromHumanResolvedResult(result) {
+    return normalizeReaderCandidate({
+        source: 'human_resolved',
+        label: 'human-resolved',
+        finalUrl: result.finalUrl ?? result.url ?? '',
+        title: result.title ?? '',
+        text: result.text ?? '',
+        contentType: result.contentType ?? 'text/html',
+        status: result.status ?? 200,
+        ok: true,
+        evidence: ['human-resolved-challenge'],
+        warnings: [],
+        safetyFlags: ['user_session_used', 'human_action_taken'],
+    });
+}
+/**
+ * @param {string} html
+ * @param {string} finalUrl
+ */
+export function fromHtmlText(html, finalUrl) {
+    return normalizeReaderCandidate({
+        source: 'reader',
+        finalUrl,
+        title: '',
+        text: htmlToReadableText(html),
+        contentType: 'text/html',
+        status: 200,
+        ok: true,
+    });
+}
+//# sourceMappingURL=reader-adapters.js.map

package/dist/src/browser/adaptive-fetch/reader-adapters.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"reader-adapters.js","sourceRoot":"","sources":["../../../../src/browser/adaptive-fetch/reader-adapters.ts"],"names":[],"mappings":"AAAA,cAAc;AACd,qHAAqH;AAErH,OAAO,EAAE,uBAAuB,EAAE,MAAM,eAAe,CAAC;AACxD,OAAO,EAAE,kBAAkB,EAAE,iBAAiB,EAAE,mBAAmB,EAAE,MAAM,iBAAiB,CAAC;AAE7F;;;GAGG;AACH,MAAM,UAAU,eAAe,CAAC,OAAO,EAAE,OAAO,GAAG,EAAE;IACjD,MAAM,MAAM,GAAG,OAAO,CAAC,MAAM,IAAI,OAAO,CAAC;IACzC,MAAM,MAAM,GAAG,iBAAiB,CAAC,OAAO,CAAC,WAAW,IAAI,EAAE,CAAC,CAAC;IAC5D,MAAM,QAAQ,GAAG,MAAM,CAAC,CAAC,CAAC,uBAAuB,CAAC,OAAO,CAAC,IAAI,IAAI,EAAE,EAAE,OAAO,CAAC,QAAQ,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;IACrG,MAAM,IAAI,GAAG,MAAM,CAAC,CAAC,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,mBAAmB,CAAC,OAAO,CAAC,IAAI,IAAI,EAAE,CAAC,CAAC;IAC9E,OAAO;QACH,MAAM;QACN,KAAK,EAAE,OAAO,CAAC,KAAK,IAAI,MAAM;QAC9B,QAAQ,EAAE,OAAO,CAAC,QAAQ,IAAI,EAAE;QAChC,KAAK,EAAE,QAAQ,EAAE,KAAK,IAAI,EAAE;QAC5B,IAAI;QACJ,WAAW,EAAE,OAAO,CAAC,WAAW,IAAI,EAAE;QACtC,MAAM,EAAE,MAAM,CAAC,OAAO,CAAC,MAAM,IAAI,CAAC,CAAC;QACnC,EAAE,EAAE,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;QACvB,QAAQ,EAAE,QAAQ,EAAE,QAAQ,IAAI,IAAI;QACpC,QAAQ,EAAE,CAAC,GAAG,CAAC,OAAO,CAAC,QAAQ,IAAI,EAAE,CAAC,EAAE,GAAG,CAAC,QAAQ,EAAE,QAAQ,IAAI,EAAE,CAAC,CAAC;QACtE,QAAQ,EAAE,OAAO,CAAC,QAAQ,IAAI,EAAE;QAChC,aAAa,EAAE,MAAM,CAAC,OAAO,CAAC,IAAI,IAAI,EAAE,CAAC,CAAC,MAAM;KACnD,CAAC;AACN,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,kBAAkB,CAAC,MAAM;IACrC,OAAO,wBAAwB,CAAC;QAC5B,MAAM,EAAE,UAAU;QAClB,KAAK,EAAE,UAAU;QACjB,QAAQ,EAAE,MAAM,CAAC,QAAQ;QACzB,KAAK,EAAE,MAAM,CAAC,KAAK;QACnB,IAAI,EAAE,MAAM,CAAC,IAAI;QACjB,WAAW,EAAE,WAAW;QACxB,MAAM,EAAE,GAAG;QACX,EAAE,EAAE,IAAI;QACR,QAAQ,EAAE,MAAM,CAAC,QAAQ,IAAI,IAAI;QACjC,QAAQ,EAAE,MAAM,CAAC,QAAQ,IAAI,EAAE;QAC/B,QAAQ,EAAE,MAAM,CAAC,QAAQ,IAAI,EAAE;KAClC,CAAC,CAAC;AACP,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,wBAAwB,CAAC,MAAM;IAC3C,OAAO,wBAAwB,CAAC,EAAE,GAAG,MAAM,EAAE,MAAM,EAAE,iBAAiB,EAAE,KAAK,EAAE,MAAM,CAAC,KAAK,IAAI,iBAAiB,EAAE,CAAC,CAAC;AACxH,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,iBAAiB,CAAC,MAAM;IACpC,OAAO,wBAAwB,CAAC,EAAE,GAAG,MAAM,EAAE,MAAM,EAAE,SAAS,EAAE,KAAK,EAAE,MAAM,CAAC,KAAK,IAAI,SAAS,EAAE,CAAC,CAAC;AACxG,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,oBAAoB,CAAC,MAAM;IACvC,OAAO,wBAAwB,CAAC,EAAE,GAAG,MAAM,EAAE,MAAM,EAAE,aAAa,EAAE,KAAK,EAAE,MAAM,CAAC,KAAK,IAAI,aAAa,EAAE,CAAC,CAAC;AAChH,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,yBAAyB,CAAC,OAAO,GAAG,EAAE;IAClD,OAAO,OAAO,CAAC,GAAG,CAAC,wBAAwB,CAAC,CAAC,MAAM,CAAC,SAAS,CAAC,EAAE,CAAC,SAAS,CAAC,QAAQ,IAAI,SAAS,CAAC,IAAI,CAAC,CAAC;AAC3G,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,wBAAwB,CAAC,MAAM,GAAG,EAAE;IAChD,OAAO;QACH,MAAM,EAAE,MAAM,CAAC,MAAM,IAAI,QAAQ;QACjC,KAAK,EAAE,MAAM,CAAC,KAAK,IAAI,MAAM,CAAC,MAAM,IAAI,QAAQ;QAChD,QAAQ,EAAE,MAAM,CAAC,QAAQ,IAAI,EAAE;QAC/B,KAAK,EAAE,mBAAmB,CAAC,MAAM,CAAC,KAAK,IAAI,EAAE,CAAC;QAC9C,IAAI,EAAE,mBAAmB,CAAC,MAAM,CAAC,IAAI,IAAI,EAAE,CAAC;QAC5C,WAAW,EAAE,MAAM,CAAC,WAAW,IAAI,EAAE;QACrC,MAAM,EAAE,MAAM,CAAC,MAAM,CAAC,MAAM,IAAI,CAAC,CAAC;QAClC,EAAE,EAAE,MAAM,CAAC,EAAE,KAAK,KAAK;QACvB,QAAQ,EAAE,MAAM,CAAC,QAAQ,IAAI,IAAI;QACjC,QAAQ,EAAE,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,EAAE;QAC/E,QAAQ,EAAE,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,EAAE;QAC/E,WAAW,EAAE,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,WAAW,CAAC,CAAC,CAAC,EAAE;QACxE,aAAa,EAAE,MAAM,CAAC,MAAM,CAAC,aAAa,IAAI,MAAM,CAAC,MAAM,CAAC,IAAI,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC;KAClF,CAAC;AACN,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,qBAAqB,CAAC,MAAM;IACxC,OAAO,wBAAwB,CAAC;QAC5B,GAAG,MAAM;QACT,MAAM,EAAE,cAAc;QACtB,KAAK,EAAE,MAAM,CAAC,KAAK,IAAI,sBAAsB;QAC7C,WAAW,EAAE,CAAC,GAAG,CAAC,MAAM,CAAC,WAAW,IAAI,EAAE,CAAC,CAAC;KAC/C,CAAC,CAAC;AACP,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,uBAAuB,CAAC,MAAM;IAC1C,OAAO,wBAAwB,CAAC;QAC5B,MAAM,EAAE,gBAAgB;QACxB,KAAK,EAAE,gBAAgB;QACvB,QAAQ,EAAE,MAAM,CAAC,QAAQ,IAAI,MAAM,CAAC,GAAG,IAAI,EAAE;QAC7C,KAAK,EAAE,MAAM,CAAC,KAAK,IAAI,EAAE;QACzB,IAAI,EAAE,MAAM,CAAC,IAAI,IAAI,EAAE;QACvB,WAAW,EAAE,MAAM,CAAC,WAAW,IAAI,WAAW;QAC9C,MAAM,EAAE,MAAM,CAAC,MAAM,IAAI,GAAG;QAC5B,EAAE,EAAE,IAAI;QACR,QAAQ,EAAE,CAAC,0BAA0B,CAAC;QACtC,QAAQ,EAAE,EAAE;QACZ,WAAW,EAAE,CAAC,mBAAmB,EAAE,oBAAoB,CAAC;KAC3D,CAAC,CAAC;AACP,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,YAAY,CAAC,IAAI,EAAE,QAAQ;IACvC,OAAO,wBAAwB,CAAC;QAC5B,MAAM,EAAE,QAAQ;QAChB,QAAQ;QACR,KAAK,EAAE,EAAE;QACT,IAAI,EAAE,kBAAkB,CAAC,IAAI,CAAC;QAC9B,WAAW,EAAE,WAAW;QACxB,MAAM,EAAE,GAAG;QACX,EAAE,EAAE,IAAI;KACX,CAAC,CAAC;AACP,CAAC"}