@olib-ai/owl-browser-sdk 2.0.5 → 2.0.7

package/dist/extraction/transforms.js

@@ -0,0 +1,277 @@
+/**
+ * Pure value transform pipeline for extraction.
+ *
+ * No browser dependency. All functions are stateless and side-effect free.
+ *
+ * Built-in transforms:
+ * - trim, lowercase, uppercase, number, clean, slug (from playwright extractor)
+ * - date: parse various date formats to ISO-8601 (including relative dates)
+ * - price: "$1,299.99" → 1299.99
+ * - url: resolve relative URLs to absolute
+ * - email: extract email address from text
+ * - stripHtml: remove HTML tags and decode all HTML entities
+ * - boolean: normalize truthy/falsy strings to "true"/"false"
+ * - compact: remove ALL whitespace (including newlines, tabs)
+ */
+/**
+ * Apply a single named transform to a string value.
+ */
+export function applyTransform(value, transform) {
+    switch (transform) {
+        case 'trim':
+            return value.trim();
+        case 'lowercase':
+            return value.toLowerCase();
+        case 'uppercase':
+            return value.toUpperCase();
+        case 'number': {
+            const cleaned = value.replace(/[^0-9.\-,]/g, '').replace(/,/g, '');
+            const n = parseFloat(cleaned);
+            return isNaN(n) ? null : String(n);
+        }
+        case 'clean':
+            return value.replace(/\s+/g, ' ').trim();
+        case 'slug':
+            return value
+                .toLowerCase()
+                .replace(/[^a-z0-9]+/g, '-')
+                .replace(/^-|-$/g, '');
+        case 'date':
+            return parseDate(value);
+        case 'price':
+            return parsePriceToString(value);
+        case 'url':
+            // URL resolution needs a base — just clean the value here
+            return value.trim();
+        case 'email':
+            return extractEmail(value);
+        case 'stripHtml':
+            return stripHtml(value);
+        case 'boolean': {
+            const lower = value.trim().toLowerCase();
+            const truthy = new Set(['yes', 'true', '1', 'on', 'enabled', '✓', '✔', '☑']);
+            const falsy = new Set(['no', 'false', '0', 'off', 'disabled', '✗', '✘', '☐']);
+            if (truthy.has(lower))
+                return 'true';
+            if (falsy.has(lower))
+                return 'false';
+            return value ? 'true' : 'false';
+        }
+        case 'compact':
+            return value.replace(/\s+/g, '').trim();
+        default:
+            return value;
+    }
+}
+/**
+ * Apply a pipeline of transforms to a string value.
+ */
+export function applyTransforms(value, transforms) {
+    if (!transforms || value === null)
+        return value;
+    const pipeline = Array.isArray(transforms) ? transforms : [transforms];
+    let result = value;
+    for (const t of pipeline) {
+        if (result === null)
+            break;
+        result = applyTransform(result, t);
+    }
+    return result;
+}
+/**
+ * Apply a regex pattern to extract a substring.
+ */
+export function applyPattern(value, pattern, group) {
+    if (!pattern || value === null)
+        return value;
+    try {
+        const re = new RegExp(pattern);
+        const match = re.exec(value);
+        if (!match)
+            return null;
+        return match[group ?? 0] ?? null;
+    }
+    catch {
+        return value;
+    }
+}
+/**
+ * Coerce a string to the specified type.
+ */
+export function coerceType(value, type) {
+    if (value === null)
+        return null;
+    if (type === 'number') {
+        const cleaned = value.replace(/[^0-9.\-,]/g, '').replace(/,/g, '');
+        const n = parseFloat(cleaned);
+        return isNaN(n) ? null : n;
+    }
+    return value;
+}
+/**
+ * Parse a price string to a number.
+ * Handles: "$1,299.99", "EUR 1.299,99", "1 299,99 kr", etc.
+ */
+export function parsePrice(value) {
+    // Remove currency symbols and whitespace
+    let cleaned = value.replace(/[^0-9.,\-\s]/g, '').trim();
+    if (!cleaned)
+        return null;
+    // Detect European format: "1.299,99" or "1 299,99"
+    const commaPos = cleaned.lastIndexOf(',');
+    const dotPos = cleaned.lastIndexOf('.');
+    if (commaPos > dotPos && commaPos > 0) {
+        // Comma is the decimal separator (European)
+        cleaned = cleaned.replace(/[.\s]/g, '').replace(',', '.');
+    }
+    else {
+        // Dot is the decimal separator (US/UK)
+        cleaned = cleaned.replace(/[,\s]/g, '');
+    }
+    const n = parseFloat(cleaned);
+    return isNaN(n) ? null : n;
+}
+/**
+ * Parse a date string to ISO-8601 format.
+ * Handles common formats: "Jan 15, 2024", "15/01/2024", "2024-01-15", etc.
+ */
+export function parseDate(value) {
+    const trimmed = value.trim();
+    if (!trimmed)
+        return null;
+    // Relative date patterns
+    const relativeResult = parseRelativeDate(trimmed);
+    if (relativeResult)
+        return relativeResult;
+    // Try native Date parsing first
+    const d = new Date(trimmed);
+    if (!isNaN(d.getTime())) {
+        return d.toISOString();
+    }
+    // Try DD/MM/YYYY or DD-MM-YYYY
+    const dmyMatch = /^(\d{1,2})[/\-.](\d{1,2})[/\-.](\d{4})$/.exec(trimmed);
+    if (dmyMatch) {
+        const [, day, month, year] = dmyMatch;
+        const parsed = new Date(`${year}-${month.padStart(2, '0')}-${day.padStart(2, '0')}`);
+        if (!isNaN(parsed.getTime()))
+            return parsed.toISOString();
+    }
+    // Try "15 Jan 2024" or "15 January 2024"
+    const dayMonthYear = /^(\d{1,2})\s+(Jan(?:uary)?|Feb(?:ruary)?|Mar(?:ch)?|Apr(?:il)?|May|Jun(?:e)?|Jul(?:y)?|Aug(?:ust)?|Sep(?:tember)?|Oct(?:ober)?|Nov(?:ember)?|Dec(?:ember)?)\s+(\d{4})$/i.exec(trimmed);
+    if (dayMonthYear) {
+        const parsed = new Date(`${dayMonthYear[2]} ${dayMonthYear[1]}, ${dayMonthYear[3]}`);
+        if (!isNaN(parsed.getTime()))
+            return parsed.toISOString();
+    }
+    return null;
+}
+/**
+ * Resolve a relative URL against a base URL.
+ */
+export function resolveUrl(relative, baseUrl) {
+    try {
+        return new URL(relative, baseUrl).href;
+    }
+    catch {
+        return relative;
+    }
+}
+// ==================== Internal Helpers ====================
+function parseRelativeDate(text) {
+    const lower = text.toLowerCase().trim();
+    const now = new Date();
+    if (lower === 'just now' || lower === 'now') {
+        return now.toISOString();
+    }
+    if (lower === 'today') {
+        return new Date(now.getFullYear(), now.getMonth(), now.getDate()).toISOString();
+    }
+    if (lower === 'yesterday') {
+        const d = new Date(now);
+        d.setDate(d.getDate() - 1);
+        return new Date(d.getFullYear(), d.getMonth(), d.getDate()).toISOString();
+    }
+    const agoMatch = /^(\d+)\s+(second|minute|hour|day|week|month|year)s?\s+ago$/i.exec(lower);
+    if (agoMatch) {
+        const amount = parseInt(agoMatch[1], 10);
+        const unit = agoMatch[2].toLowerCase();
+        const d = new Date(now);
+        switch (unit) {
+            case 'second':
+                d.setSeconds(d.getSeconds() - amount);
+                break;
+            case 'minute':
+                d.setMinutes(d.getMinutes() - amount);
+                break;
+            case 'hour':
+                d.setHours(d.getHours() - amount);
+                break;
+            case 'day':
+                d.setDate(d.getDate() - amount);
+                break;
+            case 'week':
+                d.setDate(d.getDate() - amount * 7);
+                break;
+            case 'month':
+                d.setMonth(d.getMonth() - amount);
+                break;
+            case 'year':
+                d.setFullYear(d.getFullYear() - amount);
+                break;
+        }
+        return d.toISOString();
+    }
+    return null;
+}
+function parsePriceToString(value) {
+    const n = parsePrice(value);
+    return n !== null ? String(n) : null;
+}
+function extractEmail(value) {
+    const match = /[a-zA-Z0-9._%+\-]+@[a-zA-Z0-9.\-]+\.[a-zA-Z]{2,}/.exec(value);
+    return match ? match[0] : null;
+}
+function stripHtml(value) {
+    let result = value
+        .replace(/<script[^>]*>[\s\S]*?<\/script>/gi, '')
+        .replace(/<style[^>]*>[\s\S]*?<\/style>/gi, '')
+        .replace(/<[^>]+>/g, '');
+    // Decode numeric HTML entities (decimal and hex)
+    result = result.replace(/&#x([0-9a-fA-F]+);/g, (_m, hex) => {
+        try {
+            return String.fromCodePoint(parseInt(hex, 16));
+        }
+        catch {
+            return '';
+        }
+    });
+    result = result.replace(/&#(\d+);/g, (_m, dec) => {
+        try {
+            return String.fromCodePoint(parseInt(dec, 10));
+        }
+        catch {
+            return '';
+        }
+    });
+    // Named entities — common ones
+    const entities = {
+        '&nbsp;': ' ', '&amp;': '&', '&lt;': '<', '&gt;': '>',
+        '&quot;': '"', '&#39;': "'", '&apos;': "'",
+        '&mdash;': '\u2014', '&ndash;': '\u2013', '&hellip;': '\u2026',
+        '&lsquo;': '\u2018', '&rsquo;': '\u2019',
+        '&ldquo;': '\u201C', '&rdquo;': '\u201D',
+        '&bull;': '\u2022', '&middot;': '\u00B7',
+        '&copy;': '\u00A9', '&reg;': '\u00AE', '&trade;': '\u2122',
+        '&deg;': '\u00B0', '&plusmn;': '\u00B1', '&times;': '\u00D7', '&divide;': '\u00F7',
+        '&frac12;': '\u00BD', '&frac14;': '\u00BC', '&frac34;': '\u00BE',
+        '&euro;': '\u20AC', '&pound;': '\u00A3', '&yen;': '\u00A5', '&cent;': '\u00A2',
+        '&laquo;': '\u00AB', '&raquo;': '\u00BB',
+        '&iexcl;': '\u00A1', '&iquest;': '\u00BF',
+        '&sect;': '\u00A7', '&para;': '\u00B6', '&dagger;': '\u2020', '&Dagger;': '\u2021',
+    };
+    for (const [entity, char] of Object.entries(entities)) {
+        result = result.replaceAll(entity, char);
+    }
+    return result.replace(/\s+/g, ' ').trim();
+}
+//# sourceMappingURL=transforms.js.map

package/dist/extraction/transforms.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"transforms.js","sourceRoot":"","sources":["../../src/extraction/transforms.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;GAcG;AAIH;;GAEG;AACH,MAAM,UAAU,cAAc,CAAC,KAAa,EAAE,SAAoB;IAChE,QAAQ,SAAS,EAAE,CAAC;QAClB,KAAK,MAAM;YACT,OAAO,KAAK,CAAC,IAAI,EAAE,CAAC;QACtB,KAAK,WAAW;YACd,OAAO,KAAK,CAAC,WAAW,EAAE,CAAC;QAC7B,KAAK,WAAW;YACd,OAAO,KAAK,CAAC,WAAW,EAAE,CAAC;QAC7B,KAAK,QAAQ,CAAC,CAAC,CAAC;YACd,MAAM,OAAO,GAAG,KAAK,CAAC,OAAO,CAAC,aAAa,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;YACnE,MAAM,CAAC,GAAG,UAAU,CAAC,OAAO,CAAC,CAAC;YAC9B,OAAO,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC;QACrC,CAAC;QACD,KAAK,OAAO;YACV,OAAO,KAAK,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;QAC3C,KAAK,MAAM;YACT,OAAO,KAAK;iBACT,WAAW,EAAE;iBACb,OAAO,CAAC,aAAa,EAAE,GAAG,CAAC;iBAC3B,OAAO,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC;QAC3B,KAAK,MAAM;YACT,OAAO,SAAS,CAAC,KAAK,CAAC,CAAC;QAC1B,KAAK,OAAO;YACV,OAAO,kBAAkB,CAAC,KAAK,CAAC,CAAC;QACnC,KAAK,KAAK;YACR,0DAA0D;YAC1D,OAAO,KAAK,CAAC,IAAI,EAAE,CAAC;QACtB,KAAK,OAAO;YACV,OAAO,YAAY,CAAC,KAAK,CAAC,CAAC;QAC7B,KAAK,WAAW;YACd,OAAO,SAAS,CAAC,KAAK,CAAC,CAAC;QAC1B,KAAK,SAAS,CAAC,CAAC,CAAC;YACf,MAAM,KAAK,GAAG,KAAK,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;YACzC,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,CAAC,KAAK,EAAE,MAAM,EAAE,GAAG,EAAE,IAAI,EAAE,SAAS,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;YAC7E,MAAM,KAAK,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,OAAO,EAAE,GAAG,EAAE,KAAK,EAAE,UAAU,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;YAC9E,IAAI,MAAM,CAAC,GAAG,CAAC,KAAK,CAAC;gBAAE,OAAO,MAAM,CAAC;YACrC,IAAI,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC;gBAAE,OAAO,OAAO,CAAC;YACrC,OAAO,KAAK,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,OAAO,CAAC;QAClC,CAAC;QACD,KAAK,SAAS;YACZ,OAAO,KAAK,CAAC,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;QAC1C;YACE,OAAO,KAAK,CAAC;IACjB,CAAC;AACH,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,eAAe,CAC7B,KAAoB,EACpB,UAAoC;IAEpC,IAAI,CAAC,UAAU,IAAI,KAAK,KAAK,IAAI;QAAE,OAAO,KAAK,CAAC;IAChD,MAAM,QAAQ,GAAG,KAAK,CAAC,OAAO,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC;IACvE,IAAI,MAAM,GAAkB,KAAK,CAAC;IAClC,KAAK,MAAM,CAAC,IAAI,QAAQ,EAAE,CAAC;QACzB,IAAI,MAAM,KAAK,IAAI;YAAE,MAAM;QAC3B,MAAM,GAAG,cAAc,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC;IACrC,CAAC;IACD,OAAO,MAAM,CAAC;AAChB,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,YAAY,CAC1B,KAAoB,EACpB,OAAgB,EAChB,KAAc;IAEd,IAAI,CAAC,OAAO,IAAI,KAAK,KAAK,IAAI;QAAE,OAAO,KAAK,CAAC;IAC7C,IAAI,CAAC;QACH,MAAM,EAAE,GAAG,IAAI,MAAM,CAAC,OAAO,CAAC,CAAC;QAC/B,MAAM,KAAK,GAAG,EAAE,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QAC7B,IAAI,CAAC,KAAK;YAAE,OAAO,IAAI,CAAC;QACxB,OAAO,KAAK,CAAC,KAAK,IAAI,CAAC,CAAC,IAAI,IAAI,CAAC;IACnC,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,KAAK,CAAC;IACf,CAAC;AACH,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,UAAU,CAAC,KAAoB,EAAE,IAAa;IAC5D,IAAI,KAAK,KAAK,IAAI;QAAE,OAAO,IAAI,CAAC;IAChC,IAAI,IAAI,KAAK,QAAQ,EAAE,CAAC;QACtB,MAAM,OAAO,GAAG,KAAK,CAAC,OAAO,CAAC,aAAa,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;QACnE,MAAM,CAAC,GAAG,UAAU,CAAC,OAAO,CAAC,CAAC;QAC9B,OAAO,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC;IAC7B,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,UAAU,CAAC,KAAa;IACtC,yCAAyC;IACzC,IAAI,OAAO,GAAG,KAAK,CAAC,OAAO,CAAC,eAAe,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;IACxD,IAAI,CAAC,OAAO;QAAE,OAAO,IAAI,CAAC;IAE1B,mDAAmD;IACnD,MAAM,QAAQ,GAAG,OAAO,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;IAC1C,MAAM,MAAM,GAAG,OAAO,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;IAExC,IAAI,QAAQ,GAAG,MAAM,IAAI,QAAQ,GAAG,CAAC,EAAE,CAAC;QACtC,4CAA4C;QAC5C,OAAO,GAAG,OAAO,CAAC,OAAO,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,GAAG,EAAE,GAAG,CAAC,CAAC;IAC5D,CAAC;SAAM,CAAC;QACN,uCAAuC;QACvC,OAAO,GAAG,OAAO,CAAC,OAAO,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC;IAC1C,CAAC;IAED,MAAM,CAAC,GAAG,UAAU,CAAC,OAAO,CAAC,CAAC;IAC9B,OAAO,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC;AAC7B,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,SAAS,CAAC,KAAa;IACrC,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,EAAE,CAAC;IAC7B,IAAI,CAAC,OAAO;QAAE,OAAO,IAAI,CAAC;IAE1B,yBAAyB;IACzB,MAAM,cAAc,GAAG,iBAAiB,CAAC,OAAO,CAAC,CAAC;IAClD,IAAI,cAAc;QAAE,OAAO,cAAc,CAAC;IAE1C,gCAAgC;IAChC,MAAM,CAAC,GAAG,IAAI,IAAI,CAAC,OAAO,CAAC,CAAC;IAC5B,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC,EAAE,CAAC;QACxB,OAAO,CAAC,CAAC,WAAW,EAAE,CAAC;IACzB,CAAC;IAED,+BAA+B;IAC/B,MAAM,QAAQ,GAAG,yCAAyC,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;IACzE,IAAI,QAAQ,EAAE,CAAC;QACb,MAAM,CAAC,EAAE,GAAG,EAAE,KAAK,EAAE,IAAI,CAAC,GAAG,QAAQ,CAAC;QACtC,MAAM,MAAM,GAAG,IAAI,IAAI,CAAC,GAAG,IAAI,IAAI,KAAM,CAAC,QAAQ,CAAC,CAAC,EAAE,GAAG,CAAC,IAAI,GAAI,CAAC,QAAQ,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE,CAAC,CAAC;QACvF,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,OAAO,EAAE,CAAC;YAAE,OAAO,MAAM,CAAC,WAAW,EAAE,CAAC;IAC5D,CAAC;IAED,yCAAyC;IACzC,MAAM,YAAY,GAChB,yKAAyK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;IAC1L,IAAI,YAAY,EAAE,CAAC;QACjB,MAAM,MAAM,GAAG,IAAI,IAAI,CAAC,GAAG,YAAY,CAAC,CAAC,CAAC,IAAI,YAAY,CAAC,CAAC,CAAC,KAAK,YAAY,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;QACrF,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,OAAO,EAAE,CAAC;YAAE,OAAO,MAAM,CAAC,WAAW,EAAE,CAAC;IAC5D,CAAC;IAED,OAAO,IAAI,CAAC;AACd,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,UAAU,CAAC,QAAgB,EAAE,OAAe;IAC1D,IAAI,CAAC;QACH,OAAO,IAAI,GAAG,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC,IAAI,CAAC;IACzC,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,QAAQ,CAAC;IAClB,CAAC;AACH,CAAC;AAED,6DAA6D;AAE7D,SAAS,iBAAiB,CAAC,IAAY;IACrC,MAAM,KAAK,GAAG,IAAI,CAAC,WAAW,EAAE,CAAC,IAAI,EAAE,CAAC;IACxC,MAAM,GAAG,GAAG,IAAI,IAAI,EAAE,CAAC;IAEvB,IAAI,KAAK,KAAK,UAAU,IAAI,KAAK,KAAK,KAAK,EAAE,CAAC;QAC5C,OAAO,GAAG,CAAC,WAAW,EAAE,CAAC;IAC3B,CAAC;IACD,IAAI,KAAK,KAAK,OAAO,EAAE,CAAC;QACtB,OAAO,IAAI,IAAI,CAAC,GAAG,CAAC,WAAW,EAAE,EAAE,GAAG,CAAC,QAAQ,EAAE,EAAE,GAAG,CAAC,OAAO,EAAE,CAAC,CAAC,WAAW,EAAE,CAAC;IAClF,CAAC;IACD,IAAI,KAAK,KAAK,WAAW,EAAE,CAAC;QAC1B,MAAM,CAAC,GAAG,IAAI,IAAI,CAAC,GAAG,CAAC,CAAC;QACxB,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,OAAO,EAAE,GAAG,CAAC,CAAC,CAAC;QAC3B,OAAO,IAAI,IAAI,CAAC,CAAC,CAAC,WAAW,EAAE,EAAE,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC,OAAO,EAAE,CAAC,CAAC,WAAW,EAAE,CAAC;IAC5E,CAAC;IAED,MAAM,QAAQ,GAAG,6DAA6D,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;IAC3F,IAAI,QAAQ,EAAE,CAAC;QACb,MAAM,MAAM,GAAG,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAAE,EAAE,EAAE,CAAC,CAAC;QAC1C,MAAM,IAAI,GAAG,QAAQ,CAAC,CAAC,CAAE,CAAC,WAAW,EAAE,CAAC;QACxC,MAAM,CAAC,GAAG,IAAI,IAAI,CAAC,GAAG,CAAC,CAAC;QACxB,QAAQ,IAAI,EAAE,CAAC;YACb,KAAK,QAAQ;gBAAE,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,UAAU,EAAE,GAAG,MAAM,CAAC,CAAC;gBAAC,MAAM;YAC5D,KAAK,QAAQ;gBAAE,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,UAAU,EAAE,GAAG,MAAM,CAAC,CAAC;gBAAC,MAAM;YAC5D,KAAK,MAAM;gBAAE,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,QAAQ,EAAE,GAAG,MAAM,CAAC,CAAC;gBAAC,MAAM;YACtD,KAAK,KAAK;gBAAE,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,OAAO,EAAE,GAAG,MAAM,CAAC,CAAC;gBAAC,MAAM;YACnD,KAAK,MAAM;gBAAE,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,OAAO,EAAE,GAAG,MAAM,GAAG,CAAC,CAAC,CAAC;gBAAC,MAAM;YACxD,KAAK,OAAO;gBAAE,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,QAAQ,EAAE,GAAG,MAAM,CAAC,CAAC;gBAAC,MAAM;YACvD,KAAK,MAAM;gBAAE,CAAC,CAAC,WAAW,CAAC,CAAC,CAAC,WAAW,EAAE,GAAG,MAAM,CAAC,CAAC;gBAAC,MAAM;QAC9D,CAAC;QACD,OAAO,CAAC,CAAC,WAAW,EAAE,CAAC;IACzB,CAAC;IAED,OAAO,IAAI,CAAC;AACd,CAAC;AAED,SAAS,kBAAkB,CAAC,KAAa;IACvC,MAAM,CAAC,GAAG,UAAU,CAAC,KAAK,CAAC,CAAC;IAC5B,OAAO,CAAC,KAAK,IAAI,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;AACvC,CAAC;AAED,SAAS,YAAY,CAAC,KAAa;IACjC,MAAM,KAAK,GAAG,kDAAkD,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;IAC7E,OAAO,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;AACjC,CAAC;AAED,SAAS,SAAS,CAAC,KAAa;IAC9B,IAAI,MAAM,GAAG,KAAK;SACf,OAAO,CAAC,mCAAmC,EAAE,EAAE,CAAC;SAChD,OAAO,CAAC,iCAAiC,EAAE,EAAE,CAAC;SAC9C,OAAO,CAAC,UAAU,EAAE,EAAE,CAAC,CAAC;IAE3B,iDAAiD;IACjD,MAAM,GAAG,MAAM,CAAC,OAAO,CAAC,qBAAqB,EAAE,CAAC,EAAE,EAAE,GAAG,EAAE,EAAE;QACzD,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,CAAC;IACH,MAAM,GAAG,MAAM,CAAC,OAAO,CAAC,WAAW,EAAE,CAAC,EAAE,EAAE,GAAG,EAAE,EAAE;QAC/C,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,CAAC;IAEH,+BAA+B;IAC/B,MAAM,QAAQ,GAA2B;QACvC,QAAQ,EAAE,GAAG,EAAE,OAAO,EAAE,GAAG,EAAE,MAAM,EAAE,GAAG,EAAE,MAAM,EAAE,GAAG;QACrD,QAAQ,EAAE,GAAG,EAAE,OAAO,EAAE,GAAG,EAAE,QAAQ,EAAE,GAAG;QAC1C,SAAS,EAAE,QAAQ,EAAE,SAAS,EAAE,QAAQ,EAAE,UAAU,EAAE,QAAQ;QAC9D,SAAS,EAAE,QAAQ,EAAE,SAAS,EAAE,QAAQ;QACxC,SAAS,EAAE,QAAQ,EAAE,SAAS,EAAE,QAAQ;QACxC,QAAQ,EAAE,QAAQ,EAAE,UAAU,EAAE,QAAQ;QACxC,QAAQ,EAAE,QAAQ,EAAE,OAAO,EAAE,QAAQ,EAAE,SAAS,EAAE,QAAQ;QAC1D,OAAO,EAAE,QAAQ,EAAE,UAAU,EAAE,QAAQ,EAAE,SAAS,EAAE,QAAQ,EAAE,UAAU,EAAE,QAAQ;QAClF,UAAU,EAAE,QAAQ,EAAE,UAAU,EAAE,QAAQ,EAAE,UAAU,EAAE,QAAQ;QAChE,QAAQ,EAAE,QAAQ,EAAE,SAAS,EAAE,QAAQ,EAAE,OAAO,EAAE,QAAQ,EAAE,QAAQ,EAAE,QAAQ;QAC9E,SAAS,EAAE,QAAQ,EAAE,SAAS,EAAE,QAAQ;QACxC,SAAS,EAAE,QAAQ,EAAE,UAAU,EAAE,QAAQ;QACzC,QAAQ,EAAE,QAAQ,EAAE,QAAQ,EAAE,QAAQ,EAAE,UAAU,EAAE,QAAQ,EAAE,UAAU,EAAE,QAAQ;KACnF,CAAC;IACF,KAAK,MAAM,CAAC,MAAM,EAAE,IAAI,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,QAAQ,CAAC,EAAE,CAAC;QACtD,MAAM,GAAG,MAAM,CAAC,UAAU,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC;IAC3C,CAAC;IAED,OAAO,MAAM,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;AAC5C,CAAC"}

package/dist/extraction/types.d.ts

@@ -0,0 +1,199 @@
+/**
+ * Shared type definitions for the extraction module.
+ */
+/** Built-in transform names. */
+export type Transform = 'trim' | 'lowercase' | 'uppercase' | 'number' | 'clean' | 'slug' | 'date' | 'price' | 'url' | 'email' | 'stripHtml' | 'boolean' | 'compact';
+/** Object-based field spec for advanced extraction. */
+export interface ObjectFieldSpec {
+    /** CSS selector (default: container itself). */
+    selector?: string;
+    /** Attribute to extract (default: textContent). */
+    attribute?: string;
+    /** Return string[] of ALL matches, not just first. */
+    all?: boolean;
+    /** Type coercion: 'text' (default), 'number', 'html', 'innerHtml'. */
+    type?: 'text' | 'number' | 'html' | 'innerHtml';
+    /** Regex pattern to apply on extracted text. */
+    pattern?: string;
+    /** Capture group index (default: 0 = full match). */
+    group?: number;
+    /** Transform pipeline applied after extraction. */
+    transform?: Transform | Transform[];
+    /** Fallback value when extraction returns null. */
+    default?: string | number | null;
+    /** Nested extraction: extract sub-items within each container. */
+    nested?: {
+        selector: string;
+        fields: Record<string, FieldSpec>;
+        limit?: number;
+    };
+}
+/** A field spec is either a string shorthand or an object spec. */
+export type FieldSpec = string | ObjectFieldSpec;
+/** Extracted record — values can be string, number, null, string[], or nested records. */
+export type ExtractedRecord = Record<string, unknown>;
+/** Result of an extraction operation. */
+export interface ExtractionResult {
+    /** Extracted items. */
+    items: ExtractedRecord[];
+    /** Total number of items extracted. */
+    totalItems: number;
+    /** Number of pages scraped (for multi-page). */
+    pagesScraped: number;
+    /** Page URLs visited. */
+    pageUrls: string[];
+    /** Whether the scrape completed or was aborted/limited. */
+    complete: boolean;
+    /** Duration in milliseconds. */
+    durationMs: number;
+}
+/** Meta tag extraction result. */
+export interface MetaData {
+    title: string | null;
+    description: string | null;
+    canonical: string | null;
+    og: Record<string, string>;
+    twitter: Record<string, string>;
+    jsonLd: object[];
+    microdata: ExtractedRecord[];
+    feeds: FeedLink[];
+    favicon: string | null;
+    other: Record<string, string>;
+}
+/** RSS/Atom feed link. */
+export interface FeedLink {
+    type: 'rss' | 'atom' | 'unknown';
+    href: string;
+    title: string | null;
+}
+/** Detected repeating DOM pattern. */
+export interface DetectedPattern {
+    /** CSS selector for the container. */
+    containerSelector: string;
+    /** CSS selector for individual items. */
+    itemSelector: string;
+    /** Number of items found. */
+    itemCount: number;
+    /** Confidence score (0-1). */
+    confidence: number;
+    /** Inferred field names and selectors. */
+    fields: Record<string, FieldSpec>;
+    /** Sample extracted items. */
+    sample: ExtractedRecord[];
+}
+/** Options for pattern detection. */
+export interface DetectOptions {
+    /** Minimum number of repeating items to consider (default: 3). */
+    minCount?: number;
+    /** Minimum confidence score (default: 0.5). */
+    minConfidence?: number;
+    /** Hint about what you're looking for. */
+    hint?: string;
+    /** Limit detection to a specific region selector. */
+    region?: string;
+}
+/** Options for table extraction. */
+export interface TableOptions {
+    /** Override header names. */
+    headers?: string[];
+    /** Number of rows to skip after header. */
+    skipRows?: number;
+    /** Maximum rows to extract. */
+    maxRows?: number;
+    /** Transpose the table (swap rows/columns). */
+    transpose?: boolean;
+    /** Column type coercion map. */
+    columnTypes?: Record<string, 'text' | 'number'>;
+}
+/** Options for content cleaning. */
+export interface CleanOptions {
+    /** Remove cookie consent banners (default: true). */
+    cookieBanners?: boolean;
+    /** Close modal dialogs (default: true). */
+    modals?: boolean;
+    /** Remove position:fixed/sticky elements (default: true). */
+    fixedElements?: boolean;
+    /** Remove ad elements (default: false). */
+    ads?: boolean;
+    /** Trigger lazy-loaded content (default: false). */
+    lazyLoad?: boolean;
+    /** Number of scrolls for lazy loading (default: 3). */
+    lazyLoadScrolls?: number;
+    /** Wait time between lazy load scrolls in ms (default: 500). */
+    lazyLoadWait?: number;
+}
+/** Result of content cleaning. */
+export interface CleanResult {
+    /** Number of elements removed. */
+    removedCount: number;
+    /** Types of elements removed. */
+    removedTypes: string[];
+}
+/** Pagination configuration. */
+export interface PaginationConfig {
+    /** Pagination type. */
+    type: 'click-next' | 'url-pattern' | 'infinite-scroll' | 'load-more' | 'url-list';
+    /** CSS selector for the "next" button (for click-next, load-more). */
+    nextSelector?: string;
+    /** URL pattern with {page} placeholder (for url-pattern). */
+    urlPattern?: string;
+    /** Starting page number (for url-pattern, default: 1). */
+    startPage?: number;
+    /** Explicit list of URLs to visit (for url-list). */
+    urls?: string[];
+    /** CSS selector for the scroll sentinel (for infinite-scroll). */
+    scrollSelector?: string;
+    /** Wait time after navigation/scroll in ms (default: 1000). */
+    waitAfter?: number;
+}
+/** Options for multi-page scraping. */
+export interface ScrapeOptions {
+    /** Field specs for extraction. */
+    fields?: Record<string, FieldSpec>;
+    /** Pagination config (auto-detected if not provided). */
+    pagination?: PaginationConfig;
+    /** Maximum pages to scrape (default: 10). */
+    maxPages?: number;
+    /** Maximum items to collect (default: unlimited). */
+    maxItems?: number;
+    /** Clean obstructions before extraction (default: true). */
+    clean?: boolean;
+    /** Delay between pages in ms (default: 500). */
+    pageDelay?: number;
+    /** Deduplicate items by field name. */
+    deduplicateBy?: string;
+    /** Follow detail page links. */
+    follow?: FollowConfig;
+    /** Callback fired after each page. */
+    onPage?: (pageNum: number, items: ExtractedRecord[]) => void;
+    /** Callback fired for each item. */
+    onItem?: (item: ExtractedRecord) => void;
+}
+/** Configuration for following detail page links. */
+export interface FollowConfig {
+    /** CSS selector for the detail link (relative to item container). */
+    selector: string;
+    /** Attribute containing the URL (default: 'href'). */
+    attribute?: string;
+    /** Fields to extract from the detail page. */
+    fields: Record<string, FieldSpec>;
+    /** Container selector on the detail page (default: 'body'). */
+    containerSelector?: string;
+    /** Wait time after navigating to detail page in ms (default: 1000). */
+    waitAfter?: number;
+}
+/** Page info returned by browser. */
+export interface PageInfo {
+    url: string;
+    title: string;
+}
+/** Options for list extraction. */
+export interface ListOptions {
+    /** CSS selector for individual items within the container. */
+    itemSelector?: string;
+    /** Field specs (auto-detected if not provided). */
+    fields?: Record<string, FieldSpec>;
+    /** Maximum items to extract. */
+    limit?: number;
+}
+//# sourceMappingURL=types.d.ts.map

package/dist/extraction/types.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"types.d.ts","sourceRoot":"","sources":["../../src/extraction/types.ts"],"names":[],"mappings":"AAAA;;GAEG;AAIH,gCAAgC;AAChC,MAAM,MAAM,SAAS,GACjB,MAAM,GACN,WAAW,GACX,WAAW,GACX,QAAQ,GACR,OAAO,GACP,MAAM,GACN,MAAM,GACN,OAAO,GACP,KAAK,GACL,OAAO,GACP,WAAW,GACX,SAAS,GACT,SAAS,CAAC;AAId,uDAAuD;AACvD,MAAM,WAAW,eAAe;IAC9B,gDAAgD;IAChD,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,mDAAmD;IACnD,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,sDAAsD;IACtD,GAAG,CAAC,EAAE,OAAO,CAAC;IACd,sEAAsE;IACtE,IAAI,CAAC,EAAE,MAAM,GAAG,QAAQ,GAAG,MAAM,GAAG,WAAW,CAAC;IAChD,gDAAgD;IAChD,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,qDAAqD;IACrD,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,mDAAmD;IACnD,SAAS,CAAC,EAAE,SAAS,GAAG,SAAS,EAAE,CAAC;IACpC,mDAAmD;IACnD,OAAO,CAAC,EAAE,MAAM,GAAG,MAAM,GAAG,IAAI,CAAC;IACjC,kEAAkE;IAClE,MAAM,CAAC,EAAE;QACP,QAAQ,EAAE,MAAM,CAAC;QACjB,MAAM,EAAE,MAAM,CAAC,MAAM,EAAE,SAAS,CAAC,CAAC;QAClC,KAAK,CAAC,EAAE,MAAM,CAAC;KAChB,CAAC;CACH;AAED,mEAAmE;AACnE,MAAM,MAAM,SAAS,GAAG,MAAM,GAAG,eAAe,CAAC;AAIjD,0FAA0F;AAC1F,MAAM,MAAM,eAAe,GAAG,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;AAEtD,yCAAyC;AACzC,MAAM,WAAW,gBAAgB;IAC/B,uBAAuB;IACvB,KAAK,EAAE,eAAe,EAAE,CAAC;IACzB,uCAAuC;IACvC,UAAU,EAAE,MAAM,CAAC;IACnB,gDAAgD;IAChD,YAAY,EAAE,MAAM,CAAC;IACrB,yBAAyB;IACzB,QAAQ,EAAE,MAAM,EAAE,CAAC;IACnB,2DAA2D;IAC3D,QAAQ,EAAE,OAAO,CAAC;IAClB,gCAAgC;IAChC,UAAU,EAAE,MAAM,CAAC;CACpB;AAID,kCAAkC;AAClC,MAAM,WAAW,QAAQ;IACvB,KAAK,EAAE,MAAM,GAAG,IAAI,CAAC;IACrB,WAAW,EAAE,MAAM,GAAG,IAAI,CAAC;IAC3B,SAAS,EAAE,MAAM,GAAG,IAAI,CAAC;IACzB,EAAE,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IAC3B,OAAO,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IAChC,MAAM,EAAE,MAAM,EAAE,CAAC;IACjB,SAAS,EAAE,eAAe,EAAE,CAAC;IAC7B,KAAK,EAAE,QAAQ,EAAE,CAAC;IAClB,OAAO,EAAE,MAAM,GAAG,IAAI,CAAC;IACvB,KAAK,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;CAC/B;AAED,0BAA0B;AAC1B,MAAM,WAAW,QAAQ;IACvB,IAAI,EAAE,KAAK,GAAG,MAAM,GAAG,SAAS,CAAC;IACjC,IAAI,EAAE,MAAM,CAAC;IACb,KAAK,EAAE,MAAM,GAAG,IAAI,CAAC;CACtB;AAID,sCAAsC;AACtC,MAAM,WAAW,eAAe;IAC9B,sCAAsC;IACtC,iBAAiB,EAAE,MAAM,CAAC;IAC1B,yCAAyC;IACzC,YAAY,EAAE,MAAM,CAAC;IACrB,6BAA6B;IAC7B,SAAS,EAAE,MAAM,CAAC;IAClB,8BAA8B;IAC9B,UAAU,EAAE,MAAM,CAAC;IACnB,0CAA0C;IAC1C,MAAM,EAAE,MAAM,CAAC,MAAM,EAAE,SAAS,CAAC,CAAC;IAClC,8BAA8B;IAC9B,MAAM,EAAE,eAAe,EAAE,CAAC;CAC3B;AAED,qCAAqC;AACrC,MAAM,WAAW,aAAa;IAC5B,kEAAkE;IAClE,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,+CAA+C;IAC/C,aAAa,CAAC,EAAE,MAAM,CAAC;IACvB,0CAA0C;IAC1C,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,qDAAqD;IACrD,MAAM,CAAC,EAAE,MAAM,CAAC;CACjB;AAID,oCAAoC;AACpC,MAAM,WAAW,YAAY;IAC3B,6BAA6B;IAC7B,OAAO,CAAC,EAAE,MAAM,EAAE,CAAC;IACnB,2CAA2C;IAC3C,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,+BAA+B;IAC/B,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,+CAA+C;IAC/C,SAAS,CAAC,EAAE,OAAO,CAAC;IACpB,gCAAgC;IAChC,WAAW,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,GAAG,QAAQ,CAAC,CAAC;CACjD;AAID,oCAAoC;AACpC,MAAM,WAAW,YAAY;IAC3B,qDAAqD;IACrD,aAAa,CAAC,EAAE,OAAO,CAAC;IACxB,2CAA2C;IAC3C,MAAM,CAAC,EAAE,OAAO,CAAC;IACjB,6DAA6D;IAC7D,aAAa,CAAC,EAAE,OAAO,CAAC;IACxB,2CAA2C;IAC3C,GAAG,CAAC,EAAE,OAAO,CAAC;IACd,oDAAoD;IACpD,QAAQ,CAAC,EAAE,OAAO,CAAC;IACnB,uDAAuD;IACvD,eAAe,CAAC,EAAE,MAAM,CAAC;IACzB,gEAAgE;IAChE,YAAY,CAAC,EAAE,MAAM,CAAC;CACvB;AAED,kCAAkC;AAClC,MAAM,WAAW,WAAW;IAC1B,kCAAkC;IAClC,YAAY,EAAE,MAAM,CAAC;IACrB,iCAAiC;IACjC,YAAY,EAAE,MAAM,EAAE,CAAC;CACxB;AAID,gCAAgC;AAChC,MAAM,WAAW,gBAAgB;IAC/B,uBAAuB;IACvB,IAAI,EAAE,YAAY,GAAG,aAAa,GAAG,iBAAiB,GAAG,WAAW,GAAG,UAAU,CAAC;IAClF,sEAAsE;IACtE,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,6DAA6D;IAC7D,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,0DAA0D;IAC1D,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,qDAAqD;IACrD,IAAI,CAAC,EAAE,MAAM,EAAE,CAAC;IAChB,kEAAkE;IAClE,cAAc,CAAC,EAAE,MAAM,CAAC;IACxB,+DAA+D;IAC/D,SAAS,CAAC,EAAE,MAAM,CAAC;CACpB;AAED,uCAAuC;AACvC,MAAM,WAAW,aAAa;IAC5B,kCAAkC;IAClC,MAAM,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,SAAS,CAAC,CAAC;IACnC,yDAAyD;IACzD,UAAU,CAAC,EAAE,gBAAgB,CAAC;IAC9B,6CAA6C;IAC7C,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,qDAAqD;IACrD,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,4DAA4D;IAC5D,KAAK,CAAC,EAAE,OAAO,CAAC;IAChB,gDAAgD;IAChD,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,uCAAuC;IACvC,aAAa,CAAC,EAAE,MAAM,CAAC;IACvB,gCAAgC;IAChC,MAAM,CAAC,EAAE,YAAY,CAAC;IACtB,sCAAsC;IACtC,MAAM,CAAC,EAAE,CAAC,OAAO,EAAE,MAAM,EAAE,KAAK,EAAE,eAAe,EAAE,KAAK,IAAI,CAAC;IAC7D,oCAAoC;IACpC,MAAM,CAAC,EAAE,CAAC,IAAI,EAAE,eAAe,KAAK,IAAI,CAAC;CAC1C;AAED,qDAAqD;AACrD,MAAM,WAAW,YAAY;IAC3B,qEAAqE;IACrE,QAAQ,EAAE,MAAM,CAAC;IACjB,sDAAsD;IACtD,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,8CAA8C;IAC9C,MAAM,EAAE,MAAM,CAAC,MAAM,EAAE,SAAS,CAAC,CAAC;IAClC,+DAA+D;IAC/D,iBAAiB,CAAC,EAAE,MAAM,CAAC;IAC3B,uEAAuE;IACvE,SAAS,CAAC,EAAE,MAAM,CAAC;CACpB;AAED,qCAAqC;AACrC,MAAM,WAAW,QAAQ;IACvB,GAAG,EAAE,MAAM,CAAC;IACZ,KAAK,EAAE,MAAM,CAAC;CACf;AAID,mCAAmC;AACnC,MAAM,WAAW,WAAW;IAC1B,8DAA8D;IAC9D,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,mDAAmD;IACnD,MAAM,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,SAAS,CAAC,CAAC;IACnC,gCAAgC;IAChC,KAAK,CAAC,EAAE,MAAM,CAAC;CAChB"}

package/dist/extraction/types.js

@@ -0,0 +1,5 @@
+/**
+ * Shared type definitions for the extraction module.
+ */
+export {};
+//# sourceMappingURL=types.js.map

package/dist/extraction/types.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"types.js","sourceRoot":"","sources":["../../src/extraction/types.ts"],"names":[],"mappings":"AAAA;;GAEG"}

package/dist/index.d.ts

@@ -36,5 +36,6 @@ export { HTTPTransport } from './transport/http.js';
 export { WebSocketTransport } from './transport/websocket.js';
 export { TokenAuth } from './auth/token.js';
 export { JWTAuth, decodeJWT, isJWTExpired } from './auth/jwt.js';
+export { Extractor } from './extraction/index.js';
 export declare const VERSION = "2.0.0";
 //# sourceMappingURL=index.d.ts.map

package/dist/index.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA4BG;AAGH,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AAGzC,OAAO,EAEL,QAAQ,EACR,aAAa,EACb,iBAAiB,EACjB,cAAc,EACd,YAAY,EAEZ,YAAY,EACZ,SAAS,EACT,WAAW,EAEX,cAAc,EACd,YAAY,EAEZ,IAAI,EACJ,QAAQ,EACR,aAAa,EACb,UAAU,EACV,UAAU,EACV,eAAe,EACf,iBAAiB,EAEjB,YAAY,EACZ,cAAc,EACd,UAAU,GACX,MAAM,YAAY,CAAC;AAGpB,OAAO,EACL,eAAe,EACf,eAAe,EACf,mBAAmB,EACnB,kBAAkB,EAClB,YAAY,EACZ,cAAc,EACd,cAAc,EACd,iBAAiB,EACjB,oBAAoB,EACpB,eAAe,EACf,kBAAkB,EAClB,gBAAgB,EAChB,kBAAkB,EAClB,cAAc,EACd,oBAAoB,GACrB,MAAM,aAAa,CAAC;AAGrB,OAAO,EAAE,aAAa,EAAE,gBAAgB,EAAE,mBAAmB,EAAE,MAAM,cAAc,CAAC;AAGpF,OAAO,EACL,YAAY,EACZ,gBAAgB,EAChB,cAAc,EACd,gBAAgB,EAChB,iBAAiB,GAClB,MAAM,iBAAiB,CAAC;AAGzB,OAAO,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AACpD,OAAO,EAAE,kBAAkB,EAAE,MAAM,0BAA0B,CAAC;AAG9D,OAAO,EAAE,SAAS,EAAE,MAAM,iBAAiB,CAAC;AAC5C,OAAO,EAAE,OAAO,EAAE,SAAS,EAAE,YAAY,EAAE,MAAM,eAAe,CAAC;AAGjE,eAAO,MAAM,OAAO,UAAU,CAAC"}
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA4BG;AAGH,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AAGzC,OAAO,EAEL,QAAQ,EACR,aAAa,EACb,iBAAiB,EACjB,cAAc,EACd,YAAY,EAEZ,YAAY,EACZ,SAAS,EACT,WAAW,EAEX,cAAc,EACd,YAAY,EAEZ,IAAI,EACJ,QAAQ,EACR,aAAa,EACb,UAAU,EACV,UAAU,EACV,eAAe,EACf,iBAAiB,EAEjB,YAAY,EACZ,cAAc,EACd,UAAU,GACX,MAAM,YAAY,CAAC;AAGpB,OAAO,EACL,eAAe,EACf,eAAe,EACf,mBAAmB,EACnB,kBAAkB,EAClB,YAAY,EACZ,cAAc,EACd,cAAc,EACd,iBAAiB,EACjB,oBAAoB,EACpB,eAAe,EACf,kBAAkB,EAClB,gBAAgB,EAChB,kBAAkB,EAClB,cAAc,EACd,oBAAoB,GACrB,MAAM,aAAa,CAAC;AAGrB,OAAO,EAAE,aAAa,EAAE,gBAAgB,EAAE,mBAAmB,EAAE,MAAM,cAAc,CAAC;AAGpF,OAAO,EACL,YAAY,EACZ,gBAAgB,EAChB,cAAc,EACd,gBAAgB,EAChB,iBAAiB,GAClB,MAAM,iBAAiB,CAAC;AAGzB,OAAO,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AACpD,OAAO,EAAE,kBAAkB,EAAE,MAAM,0BAA0B,CAAC;AAG9D,OAAO,EAAE,SAAS,EAAE,MAAM,iBAAiB,CAAC;AAC5C,OAAO,EAAE,OAAO,EAAE,SAAS,EAAE,YAAY,EAAE,MAAM,eAAe,CAAC;AAGjE,OAAO,EAAE,SAAS,EAAE,MAAM,uBAAuB,CAAC;AAGlD,eAAO,MAAM,OAAO,UAAU,CAAC"}

package/dist/index.js

@@ -45,6 +45,8 @@ export { WebSocketTransport } from './transport/websocket.js';
 // Auth (for advanced use)
 export { TokenAuth } from './auth/token.js';
 export { JWTAuth, decodeJWT, isJWTExpired } from './auth/jwt.js';
+// Extraction
+export { Extractor } from './extraction/index.js';
 // Version
 export const VERSION = '2.0.0';
 //# sourceMappingURL=index.js.map

package/dist/index.js.map

@@ -1 +1 @@
-{"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA4BG;AAEH,cAAc;AACd,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AAEzC,QAAQ;AACR,OAAO;AACL,QAAQ;AACR,QAAQ,EACR,aAAa,EACb,iBAAiB,EACjB,cAAc,EACd,YAAY,GAoBb,MAAM,YAAY,CAAC;AAEpB,SAAS;AACT,OAAO,EACL,eAAe,EACf,eAAe,EACf,mBAAmB,EACnB,kBAAkB,EAClB,YAAY,EACZ,cAAc,EACd,cAAc,EACd,iBAAiB,EACjB,oBAAoB,EACpB,eAAe,EACf,kBAAkB,EAClB,gBAAgB,EAChB,kBAAkB,EAClB,cAAc,EACd,oBAAoB,GACrB,MAAM,aAAa,CAAC;AAErB,UAAU;AACV,OAAO,EAAE,aAAa,EAAE,gBAAgB,EAAE,mBAAmB,EAAE,MAAM,cAAc,CAAC;AAEpF,iBAAiB;AACjB,OAAO,EACL,YAAY,EACZ,gBAAgB,EAChB,cAAc,EACd,gBAAgB,EAChB,iBAAiB,GAClB,MAAM,iBAAiB,CAAC;AAEzB,+BAA+B;AAC/B,OAAO,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AACpD,OAAO,EAAE,kBAAkB,EAAE,MAAM,0BAA0B,CAAC;AAE9D,0BAA0B;AAC1B,OAAO,EAAE,SAAS,EAAE,MAAM,iBAAiB,CAAC;AAC5C,OAAO,EAAE,OAAO,EAAE,SAAS,EAAE,YAAY,EAAE,MAAM,eAAe,CAAC;AAEjE,UAAU;AACV,MAAM,CAAC,MAAM,OAAO,GAAG,OAAO,CAAC"}
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA4BG;AAEH,cAAc;AACd,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AAEzC,QAAQ;AACR,OAAO;AACL,QAAQ;AACR,QAAQ,EACR,aAAa,EACb,iBAAiB,EACjB,cAAc,EACd,YAAY,GAoBb,MAAM,YAAY,CAAC;AAEpB,SAAS;AACT,OAAO,EACL,eAAe,EACf,eAAe,EACf,mBAAmB,EACnB,kBAAkB,EAClB,YAAY,EACZ,cAAc,EACd,cAAc,EACd,iBAAiB,EACjB,oBAAoB,EACpB,eAAe,EACf,kBAAkB,EAClB,gBAAgB,EAChB,kBAAkB,EAClB,cAAc,EACd,oBAAoB,GACrB,MAAM,aAAa,CAAC;AAErB,UAAU;AACV,OAAO,EAAE,aAAa,EAAE,gBAAgB,EAAE,mBAAmB,EAAE,MAAM,cAAc,CAAC;AAEpF,iBAAiB;AACjB,OAAO,EACL,YAAY,EACZ,gBAAgB,EAChB,cAAc,EACd,gBAAgB,EAChB,iBAAiB,GAClB,MAAM,iBAAiB,CAAC;AAEzB,+BAA+B;AAC/B,OAAO,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AACpD,OAAO,EAAE,kBAAkB,EAAE,MAAM,0BAA0B,CAAC;AAE9D,0BAA0B;AAC1B,OAAO,EAAE,SAAS,EAAE,MAAM,iBAAiB,CAAC;AAC5C,OAAO,EAAE,OAAO,EAAE,SAAS,EAAE,YAAY,EAAE,MAAM,eAAe,CAAC;AAEjE,aAAa;AACb,OAAO,EAAE,SAAS,EAAE,MAAM,uBAAuB,CAAC;AAElD,UAAU;AACV,MAAM,CAAC,MAAM,OAAO,GAAG,OAAO,CAAC"}

package/dist/playwright/browser-type.d.ts

@@ -0,0 +1,101 @@
+/**
+ * Playwright-compatible BrowserType class for Owl Browser.
+ *
+ * Provides chromium.connect() and chromium.launch() entry points
+ * that mirror Playwright's top-level API. Owl Browser only supports
+ * remote connections (connect), not local browser launches.
+ *
+ * Includes connectOverCDP(), launchPersistentContext(), and launchServer()
+ * stubs for full Playwright API coverage.
+ */
+import { Browser } from './browser.js';
+import type { BrowserConnectOptions, BrowserContextOptions, LaunchOptions } from './types.js';
+/**
+ * BrowserType provides factory methods for creating Browser instances.
+ *
+ * In Playwright, you access this as `chromium`, `firefox`, or `webkit`.
+ * With Owl Browser, all three aliases point to the same implementation
+ * since Owl Browser uses its own Chromium-based engine.
+ *
+ * @example
+ * ```typescript
+ * import { chromium } from '@olib-ai/owl-browser-sdk/playwright';
+ *
+ * // Connect to a running Owl Browser instance
+ * const browser = await chromium.connect({
+ *   wsEndpoint: 'http://localhost:8080',
+ *   token: 'your-token',
+ * });
+ *
+ * const context = await browser.newContext();
+ * const page = await context.newPage();
+ * await page.goto('https://example.com');
+ * ```
+ */
+export declare class BrowserType {
+    private readonly _name;
+    constructor(name: string);
+    /** Browser engine name. */
+    name(): string;
+    /** Executable path (not applicable for Owl Browser). */
+    executablePath(): string;
+    /**
+     * Connect to a running Owl Browser server.
+     *
+     * This is the primary way to create a Browser instance.
+     * The server must already be running and accessible at the given URL.
+     *
+     * @param options - Connection options (endpoint URL and auth token)
+     * @returns A connected Browser instance
+     *
+     * @example
+     * ```typescript
+     * const browser = await chromium.connect({
+     *   wsEndpoint: 'http://localhost:8080',
+     *   token: 'test-token',
+     * });
+     * ```
+     */
+    connect(options: BrowserConnectOptions): Promise<Browser>;
+    /**
+     * Connect to a browser over CDP.
+     *
+     * Alias for connect() in Owl Browser context.
+     *
+     * @param options - Connection options
+     * @returns A connected Browser instance
+     */
+    connectOverCDP(options: BrowserConnectOptions): Promise<Browser>;
+    /**
+     * Launch a new browser instance.
+     *
+     * Owl Browser does not support local browser launches.
+     * Use connect() to connect to a running Owl Browser server instead.
+     *
+     * @param options - Launch options (ignored)
+     * @throws Error always - use connect() instead
+     */
+    launch(options?: LaunchOptions): Promise<never>;
+    /**
+     * Launch a browser with a persistent context (user data directory).
+     *
+     * Not supported by Owl Browser. Use connect() with profile_path option
+     * via browser_create_context for persistent profiles.
+     *
+     * @param userDataDir - User data directory path
+     * @param options - Launch and context options
+     * @throws Error always - use connect() instead
+     */
+    launchPersistentContext(userDataDir: string, options?: LaunchOptions & BrowserContextOptions): Promise<never>;
+    /**
+     * Launch a browser server.
+     *
+     * Not supported by Owl Browser. The server must be started independently
+     * (via Docker or direct binary execution).
+     *
+     * @param options - Launch options
+     * @throws Error always - start server externally
+     */
+    launchServer(options?: LaunchOptions): Promise<never>;
+}
+//# sourceMappingURL=browser-type.d.ts.map

package/dist/playwright/browser-type.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"browser-type.d.ts","sourceRoot":"","sources":["../../src/playwright/browser-type.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAGH,OAAO,EAAE,OAAO,EAAE,MAAM,cAAc,CAAC;AACvC,OAAO,KAAK,EACV,qBAAqB,EACrB,qBAAqB,EACrB,aAAa,EACd,MAAM,YAAY,CAAC;AAEpB;;;;;;;;;;;;;;;;;;;;;GAqBG;AACH,qBAAa,WAAW;IACtB,OAAO,CAAC,QAAQ,CAAC,KAAK,CAAS;gBAEnB,IAAI,EAAE,MAAM;IAIxB,2BAA2B;IAC3B,IAAI,IAAI,MAAM;IAId,wDAAwD;IACxD,cAAc,IAAI,MAAM;IAIxB;;;;;;;;;;;;;;;;OAgBG;IACG,OAAO,CAAC,OAAO,EAAE,qBAAqB,GAAG,OAAO,CAAC,OAAO,CAAC;IAY/D;;;;;;;OAOG;IACG,cAAc,CAAC,OAAO,EAAE,qBAAqB,GAAG,OAAO,CAAC,OAAO,CAAC;IAItE;;;;;;;;OAQG;IACG,MAAM,CAAC,OAAO,CAAC,EAAE,aAAa,GAAG,OAAO,CAAC,KAAK,CAAC;IASrD;;;;;;;;;OASG;IACG,uBAAuB,CAC3B,WAAW,EAAE,MAAM,EACnB,OAAO,CAAC,EAAE,aAAa,GAAG,qBAAqB,GAC9C,OAAO,CAAC,KAAK,CAAC;IAUjB;;;;;;;;OAQG;IACG,YAAY,CAAC,OAAO,CAAC,EAAE,aAAa,GAAG,OAAO,CAAC,KAAK,CAAC;CAQ5D"}