hunter-mate 0.1.2 → 0.1.4

package/dist/assets/server/extraction/linkedin-extractor.js

@@ -0,0 +1,244 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.extractLinkedInProfile = extractLinkedInProfile;
+const cheerio_1 = require("cheerio");
+const SECTION_KEYWORDS = new Set([
+    'About', 'Experience', 'Education', 'Skills', 'Projects', 'Featured', 'Activity',
+    'People you may know', 'You might like', 'Ad Options', 'Don’t want to see this',
+    'Explore Premium profiles', 'People who follow',
+]);
+function isSectionLike(text) {
+    const t = text.trim();
+    if (!t)
+        return true;
+    for (const kw of SECTION_KEYWORDS) {
+        if (t.includes(kw))
+            return true;
+    }
+    return false;
+}
+const NON_NAME_KEYWORDS = /notifications|messages|jobs|network|home|search/i;
+function looksLikeName(text) {
+    const t = text.trim();
+    return (t.length > 0 &&
+        t.length <= 80 &&
+        !isSectionLike(t) &&
+        !NON_NAME_KEYWORDS.test(t));
+}
+function isPunctuationOnly(text) {
+    return /^[^\p{L}\p{N}]*$/u.test(text.trim());
+}
+function looksLikeLocation(text) {
+    const t = text.trim();
+    const parts = t.split(',').map((s) => s.trim()).filter(Boolean);
+    if (parts.length === 0 || parts.length > 3)
+        return false;
+    return parts.every((p) => /^[A-Za-zÀ-ſ\s'-]{2,40}$/.test(p) && !/\d/.test(p));
+}
+function looksLikeUrl(text) {
+    return /^https?:\/\//.test(text.trim());
+}
+function looksLikeFollowers(text) {
+    const t = text.trim();
+    return /followers|connections/i.test(t) || /^[\d,]+\+?\s*$/.test(t);
+}
+function looksLikeExperienceLocation(text) {
+    const t = text.trim();
+    if (t.length > 60)
+        return false;
+    if (/\b(Area|Region|City|Metropolitan)\b/i.test(t))
+        return true;
+    if (/^[A-Z][a-z]+(?:\s+[A-Z][a-z]+){0,5}(?:\s+(Area|Region|City|Metropolitan))?$/i.test(t))
+        return true;
+    return false;
+}
+function looksLikeDateRange(text) {
+    const t = text.trim();
+    const hasYear = /\b(19|20)\d{2}\b/.test(t);
+    const hasMonthOrPresent = /\b(Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec|Present|present|至今)\b/.test(t);
+    const yearRange = /\b(19|20)\d{2}\b\s*[-–—]\s*\b(19|20)\d{2}\b/.test(t);
+    return hasYear && (hasMonthOrPresent || yearRange);
+}
+function firstText($, container) {
+    return $(container).text().replace(/\s+/g, ' ').trim();
+}
+function extractLinkedInProfile(html) {
+    const $ = (0, cheerio_1.load)(html);
+    const out = {};
+    const h2s = $('h2');
+    let name = null;
+    let nameEl = null;
+    h2s.each((_i, el) => {
+        if (name)
+            return;
+        const $el = $(el);
+        const cls = $el.attr('class') || '';
+        const parentCls = $el.parent().attr('class') || '';
+        const t = $el.text().replace(/\s+/g, ' ').trim().replace(/\s*\|\s*$/, '');
+        if (looksLikeName(t) &&
+            (cls.includes('_1f667e81') || parentCls.includes('a8841267'))) {
+            name = t;
+            nameEl = $el;
+        }
+    });
+    if (!name) {
+        h2s.each((_i, el) => {
+            if (name)
+                return;
+            const t = $(el).text().replace(/\s+/g, ' ').trim().replace(/\s*\|\s*$/, '');
+            if (looksLikeName(t)) {
+                name = t;
+                nameEl = $(el);
+            }
+        });
+    }
+    if (name)
+        out.name = name;
+    const topCard = nameEl
+        ? nameEl.closest('section').first()
+        : h2s.first().closest('section').first();
+    const topTexts = [];
+    topCard.find('p').each((_i, el) => {
+        const t = $(el).text().replace(/\s+/g, ' ').trim();
+        if (t && !topTexts.includes(t))
+            topTexts.push(t);
+    });
+    const meaningfulTop = topTexts.filter((t) => !/^·\s*\d+(?:rd|th|nd|st)?$/.test(t));
+    if (meaningfulTop[0] && !isSectionLike(meaningfulTop[0])) {
+        out.headline = meaningfulTop[0];
+    }
+    if (meaningfulTop[1] && !isSectionLike(meaningfulTop[1]) && !isPunctuationOnly(meaningfulTop[1])) {
+        out.currentCompany = meaningfulTop[1];
+    }
+    for (let i = 0; i < meaningfulTop.length; i++) {
+        const t = meaningfulTop[i];
+        if (t === out.headline || t === out.currentCompany)
+            continue;
+        if (!out.location && looksLikeLocation(t))
+            out.location = t;
+        if (looksLikeUrl(t))
+            out.website = t;
+        if (looksLikeFollowers(t)) {
+            const next = meaningfulTop[i + 1];
+            out.followers = next && /connections|followers/i.test(next) ? `${t} ${next}` : t;
+            if (next && /connections|followers/i.test(next))
+                i++;
+        }
+    }
+    const aboutH2 = h2s.filter((_i, el) => $(el).text().includes('About')).first();
+    if (aboutH2.length) {
+        let cursor = aboutH2.parent().next();
+        const summaryParts = [];
+        for (let i = 0; i < 6 && cursor.length; i++) {
+            const tag = cursor.prop('tagName')?.toLowerCase();
+            if (tag === 'h2')
+                break;
+            const t = firstText($, cursor);
+            if (t.length > 20 && !isSectionLike(t))
+                summaryParts.push(t);
+            cursor = cursor.next();
+        }
+        if (summaryParts.length)
+            out.summary = summaryParts.join(' ');
+    }
+    const expH2 = h2s.filter((_i, el) => $(el).text().includes('Experience')).first();
+    if (expH2.length) {
+        const section = expH2.closest('section').first();
+        const pTexts = [];
+        section.find('p').each((_i, el) => {
+            const t = firstText($, el);
+            if (t && !isSectionLike(t))
+                pTexts.push(t);
+        });
+        const experiences = [];
+        let lastDateIdx = -1;
+        for (let i = 0; i < pTexts.length; i++) {
+            if (looksLikeDateRange(pTexts[i])) {
+                const item = { duration: pTexts[i] };
+                if (pTexts[i - 1])
+                    item.company = pTexts[i - 1];
+                if (pTexts[i - 2])
+                    item.title = pTexts[i - 2];
+                if (experiences.length > 0 && lastDateIdx >= 0) {
+                    const desc = pTexts.slice(lastDateIdx + 1, i - 2);
+                    if (desc.length) {
+                        const first = desc[0];
+                        if (looksLikeExperienceLocation(first)) {
+                            experiences[experiences.length - 1].location = first;
+                            if (desc.length > 1) {
+                                experiences[experiences.length - 1].description = desc.slice(1).join(' ');
+                            }
+                        }
+                        else {
+                            experiences[experiences.length - 1].description = desc.join(' ');
+                        }
+                    }
+                }
+                if (Object.keys(item).length > 1)
+                    experiences.push(item);
+                lastDateIdx = i;
+            }
+        }
+        if (experiences.length)
+            out.experiences = experiences;
+    }
+    const eduH2 = h2s.filter((_i, el) => $(el).text().includes('Education')).first();
+    if (eduH2.length) {
+        const section = eduH2.closest('section').first();
+        const pTexts = [];
+        section.find('p').each((_i, el) => {
+            const t = firstText($, el);
+            if (t && !isSectionLike(t))
+                pTexts.push(t);
+        });
+        const education = [];
+        let lastDateIdx = -1;
+        for (let i = 0; i < pTexts.length; i++) {
+            if (looksLikeDateRange(pTexts[i])) {
+                const item = { years: pTexts[i] };
+                if (pTexts[i - 2]) {
+                    item.school = pTexts[i - 2];
+                    item.degree = pTexts[i - 1];
+                }
+                else if (pTexts[i - 1]) {
+                    item.school = pTexts[i - 1];
+                }
+                if (Object.keys(item).length > 1)
+                    education.push(item);
+                lastDateIdx = i;
+            }
+        }
+        if (lastDateIdx >= 0 && lastDateIdx < pTexts.length - 1) {
+            const tail = pTexts.slice(lastDateIdx + 1);
+            if (tail.length >= 2) {
+                education.push({ school: tail[0], degree: tail[1] });
+            }
+            else if (tail.length === 1) {
+                education.push({ school: tail[0] });
+            }
+        }
+        if (education.length)
+            out.education = education;
+    }
+    const skillH2 = h2s.filter((_i, el) => $(el).text().includes('Skills')).first();
+    if (skillH2.length) {
+        const section = skillH2.closest('section, div').first();
+        const skills = new Set();
+        section.find('span, button, a, li, div[role="listitem"]').each((_i, el) => {
+            const t = firstText($, el);
+            if (t &&
+                t.length <= 50 &&
+                t.length >= 2 &&
+                !isSectionLike(t) &&
+                !looksLikeDateRange(t) &&
+                !looksLikeFollowers(t) &&
+                !looksLikeUrl(t)) {
+                skills.add(t);
+            }
+        });
+        if (skills.size)
+            out.skills = Array.from(skills).slice(0, 50);
+    }
+    return out;
+}
+//# sourceMappingURL=linkedin-extractor.js.map

package/dist/assets/server/extraction/linkedin-extractor.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"linkedin-extractor.js","sourceRoot":"","sources":["../../src/extraction/linkedin-extractor.ts"],"names":[],"mappings":";;AA2EA,wDAsLC;AAjQD,qCAA2C;AAO3C,MAAM,gBAAgB,GAAG,IAAI,GAAG,CAAC;IAC/B,OAAO,EAAE,YAAY,EAAE,WAAW,EAAE,QAAQ,EAAE,UAAU,EAAE,UAAU,EAAE,UAAU;IAChF,qBAAqB,EAAE,gBAAgB,EAAE,YAAY,EAAE,wBAAwB;IAC/E,0BAA0B,EAAE,mBAAmB;CAChD,CAAC,CAAC;AAEH,SAAS,aAAa,CAAC,IAAY;IACjC,MAAM,CAAC,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;IACtB,IAAI,CAAC,CAAC;QAAE,OAAO,IAAI,CAAC;IACpB,KAAK,MAAM,EAAE,IAAI,gBAAgB,EAAE,CAAC;QAClC,IAAI,CAAC,CAAC,QAAQ,CAAC,EAAE,CAAC;YAAE,OAAO,IAAI,CAAC;IAClC,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,MAAM,iBAAiB,GAAG,kDAAkD,CAAC;AAE7E,SAAS,aAAa,CAAC,IAAY;IACjC,MAAM,CAAC,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;IACtB,OAAO,CACL,CAAC,CAAC,MAAM,GAAG,CAAC;QACZ,CAAC,CAAC,MAAM,IAAI,EAAE;QACd,CAAC,aAAa,CAAC,CAAC,CAAC;QACjB,CAAC,iBAAiB,CAAC,IAAI,CAAC,CAAC,CAAC,CAC3B,CAAC;AACJ,CAAC;AAED,SAAS,iBAAiB,CAAC,IAAY;IACrC,OAAO,mBAAmB,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC;AAC/C,CAAC;AAED,SAAS,iBAAiB,CAAC,IAAY;IAErC,MAAM,CAAC,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;IACtB,MAAM,KAAK,GAAG,CAAC,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;IAChE,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC;QAAE,OAAO,KAAK,CAAC;IACzD,OAAO,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,yBAAyB,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC;AAChF,CAAC;AAED,SAAS,YAAY,CAAC,IAAY;IAChC,OAAO,cAAc,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC;AAC1C,CAAC;AAED,SAAS,kBAAkB,CAAC,IAAY;IACtC,MAAM,CAAC,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;IACtB,OAAO,wBAAwB,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,gBAAgB,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;AACtE,CAAC;AAED,SAAS,2BAA2B,CAAC,IAAY;IAC/C,MAAM,CAAC,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;IACtB,IAAI,CAAC,CAAC,MAAM,GAAG,EAAE;QAAE,OAAO,KAAK,CAAC;IAChC,IAAI,sCAAsC,CAAC,IAAI,CAAC,CAAC,CAAC;QAAE,OAAO,IAAI,CAAC;IAChE,IAAI,8EAA8E,CAAC,IAAI,CAAC,CAAC,CAAC;QAAE,OAAO,IAAI,CAAC;IACxG,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,kBAAkB,CAAC,IAAY;IACtC,MAAM,CAAC,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;IACtB,MAAM,OAAO,GAAG,kBAAkB,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAC3C,MAAM,iBAAiB,GAAG,0EAA0E,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAC7G,MAAM,SAAS,GAAG,6CAA6C,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IACxE,OAAO,OAAO,IAAI,CAAC,iBAAiB,IAAI,SAAS,CAAC,CAAC;AACrD,CAAC;AAED,SAAS,SAAS,CAAC,CAAa,EAAE,SAAc;IAC9C,OAAO,CAAC,CAAC,SAAS,CAAC,CAAC,IAAI,EAAE,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;AACzD,CAAC;AAED,SAAgB,sBAAsB,CAAC,IAAY;IACjD,MAAM,CAAC,GAAG,IAAA,cAAI,EAAC,IAAI,CAAC,CAAC;IACrB,MAAM,GAAG,GAAwB,EAAE,CAAC;IAGpC,MAAM,GAAG,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC;IACpB,IAAI,IAAI,GAAkB,IAAI,CAAC;IAC/B,IAAI,MAAM,GAAQ,IAAI,CAAC;IACvB,GAAG,CAAC,IAAI,CAAC,CAAC,EAAU,EAAE,EAAO,EAAE,EAAE;QAC/B,IAAI,IAAI;YAAE,OAAO;QACjB,MAAM,GAAG,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC;QAClB,MAAM,GAAG,GAAG,GAAG,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC;QACpC,MAAM,SAAS,GAAG,GAAG,CAAC,MAAM,EAAE,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC;QACnD,MAAM,CAAC,GAAG,GAAG,CAAC,IAAI,EAAE,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,OAAO,CAAC,WAAW,EAAE,EAAE,CAAC,CAAC;QAC1E,IACE,aAAa,CAAC,CAAC,CAAC;YAChB,CAAC,GAAG,CAAC,QAAQ,CAAC,WAAW,CAAC,IAAI,SAAS,CAAC,QAAQ,CAAC,UAAU,CAAC,CAAC,EAC7D,CAAC;YACD,IAAI,GAAG,CAAC,CAAC;YACT,MAAM,GAAG,GAAG,CAAC;QACf,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,IAAI,CAAC,IAAI,EAAE,CAAC;QACV,GAAG,CAAC,IAAI,CAAC,CAAC,EAAU,EAAE,EAAO,EAAE,EAAE;YAC/B,IAAI,IAAI;gBAAE,OAAO;YACjB,MAAM,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,OAAO,CAAC,WAAW,EAAE,EAAE,CAAC,CAAC;YAC5E,IAAI,aAAa,CAAC,CAAC,CAAC,EAAE,CAAC;gBACrB,IAAI,GAAG,CAAC,CAAC;gBACT,MAAM,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC;YACjB,CAAC;QACH,CAAC,CAAC,CAAC;IACL,CAAC;IACD,IAAI,IAAI;QAAE,GAAG,CAAC,IAAI,GAAG,IAAI,CAAC;IAK1B,MAAM,OAAO,GAAG,MAAM;QACpB,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,KAAK,EAAE;QACnC,CAAC,CAAC,GAAG,CAAC,KAAK,EAAE,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,KAAK,EAAE,CAAC;IAC3C,MAAM,QAAQ,GAAa,EAAE,CAAC;IAC9B,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,EAAU,EAAE,EAAO,EAAE,EAAE;QAC7C,MAAM,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;QACnD,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAAC;YAAE,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IACnD,CAAC,CAAC,CAAC;IAIH,MAAM,aAAa,GAAG,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,2BAA2B,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC;IACnF,IAAI,aAAa,CAAC,CAAC,CAAC,IAAI,CAAC,aAAa,CAAC,aAAa,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;QACzD,GAAG,CAAC,QAAQ,GAAG,aAAa,CAAC,CAAC,CAAC,CAAC;IAClC,CAAC;IACD,IAAI,aAAa,CAAC,CAAC,CAAC,IAAI,CAAC,aAAa,CAAC,aAAa,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,iBAAiB,CAAC,aAAa,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;QACjG,GAAG,CAAC,cAAc,GAAG,aAAa,CAAC,CAAC,CAAC,CAAC;IACxC,CAAC;IACD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,aAAa,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAC9C,MAAM,CAAC,GAAG,aAAa,CAAC,CAAC,CAAC,CAAC;QAC3B,IAAI,CAAC,KAAK,GAAG,CAAC,QAAQ,IAAI,CAAC,KAAK,GAAG,CAAC,cAAc;YAAE,SAAS;QAC7D,IAAI,CAAC,GAAG,CAAC,QAAQ,IAAI,iBAAiB,CAAC,CAAC,CAAC;YAAE,GAAG,CAAC,QAAQ,GAAG,CAAC,CAAC;QAC5D,IAAI,YAAY,CAAC,CAAC,CAAC;YAAE,GAAG,CAAC,OAAO,GAAG,CAAC,CAAC;QACrC,IAAI,kBAAkB,CAAC,CAAC,CAAC,EAAE,CAAC;YAE1B,MAAM,IAAI,GAAG,aAAa,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;YAClC,GAAG,CAAC,SAAS,GAAG,IAAI,IAAI,wBAAwB,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,IAAI,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;YACjF,IAAI,IAAI,IAAI,wBAAwB,CAAC,IAAI,CAAC,IAAI,CAAC;gBAAE,CAAC,EAAE,CAAC;QACvD,CAAC;IACH,CAAC;IAGD,MAAM,OAAO,GAAG,GAAG,CAAC,MAAM,CAAC,CAAC,EAAE,EAAE,EAAE,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC,CAAC,KAAK,EAAE,CAAC;IAC/E,IAAI,OAAO,CAAC,MAAM,EAAE,CAAC;QACnB,IAAI,MAAM,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,IAAI,EAAE,CAAC;QACrC,MAAM,YAAY,GAAa,EAAE,CAAC;QAClC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,IAAI,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YAC5C,MAAM,GAAG,GAAG,MAAM,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,WAAW,EAAE,CAAC;YAClD,IAAI,GAAG,KAAK,IAAI;gBAAE,MAAM;YACxB,MAAM,CAAC,GAAG,SAAS,CAAC,CAAC,EAAE,MAAM,CAAC,CAAC;YAC/B,IAAI,CAAC,CAAC,MAAM,GAAG,EAAE,IAAI,CAAC,aAAa,CAAC,CAAC,CAAC;gBAAE,YAAY,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;YAC7D,MAAM,GAAG,MAAM,CAAC,IAAI,EAAE,CAAC;QACzB,CAAC;QACD,IAAI,YAAY,CAAC,MAAM;YAAE,GAAG,CAAC,OAAO,GAAG,YAAY,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IAChE,CAAC;IAID,MAAM,KAAK,GAAG,GAAG,CAAC,MAAM,CAAC,CAAC,EAAE,EAAE,EAAE,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,QAAQ,CAAC,YAAY,CAAC,CAAC,CAAC,KAAK,EAAE,CAAC;IAClF,IAAI,KAAK,CAAC,MAAM,EAAE,CAAC;QACjB,MAAM,OAAO,GAAG,KAAK,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,KAAK,EAAE,CAAC;QACjD,MAAM,MAAM,GAAa,EAAE,CAAC;QAC5B,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,EAAU,EAAE,EAAO,EAAE,EAAE;YAC7C,MAAM,CAAC,GAAG,SAAS,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;YAC3B,IAAI,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC,CAAC;gBAAE,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QAC7C,CAAC,CAAC,CAAC;QACH,MAAM,WAAW,GAA0B,EAAE,CAAC;QAC9C,IAAI,WAAW,GAAG,CAAC,CAAC,CAAC;QACrB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACvC,IAAI,kBAAkB,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;gBAClC,MAAM,IAAI,GAAwB,EAAE,QAAQ,EAAE,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC;gBAC1D,IAAI,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC;oBAAE,IAAI,CAAC,OAAO,GAAG,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;gBAChD,IAAI,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC;oBAAE,IAAI,CAAC,KAAK,GAAG,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;gBAE9C,IAAI,WAAW,CAAC,MAAM,GAAG,CAAC,IAAI,WAAW,IAAI,CAAC,EAAE,CAAC;oBAC/C,MAAM,IAAI,GAAG,MAAM,CAAC,KAAK,CAAC,WAAW,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;oBAClD,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC;wBAChB,MAAM,KAAK,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC;wBACtB,IAAI,2BAA2B,CAAC,KAAK,CAAC,EAAE,CAAC;4BACvC,WAAW,CAAC,WAAW,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,QAAQ,GAAG,KAAK,CAAC;4BACrD,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gCACpB,WAAW,CAAC,WAAW,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,WAAW,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;4BAC5E,CAAC;wBACH,CAAC;6BAAM,CAAC;4BACN,WAAW,CAAC,WAAW,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,WAAW,GAAG,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;wBACnE,CAAC;oBACH,CAAC;gBACH,CAAC;gBACD,IAAI,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,MAAM,GAAG,CAAC;oBAAE,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;gBACzD,WAAW,GAAG,CAAC,CAAC;YAClB,CAAC;QACH,CAAC;QACD,IAAI,WAAW,CAAC,MAAM;YAAE,GAAG,CAAC,WAAW,GAAG,WAAW,CAAC;IACxD,CAAC;IAGD,MAAM,KAAK,GAAG,GAAG,CAAC,MAAM,CAAC,CAAC,EAAE,EAAE,EAAE,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,QAAQ,CAAC,WAAW,CAAC,CAAC,CAAC,KAAK,EAAE,CAAC;IACjF,IAAI,KAAK,CAAC,MAAM,EAAE,CAAC;QACjB,MAAM,OAAO,GAAG,KAAK,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,KAAK,EAAE,CAAC;QACjD,MAAM,MAAM,GAAa,EAAE,CAAC;QAC5B,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,EAAU,EAAE,EAAO,EAAE,EAAE;YAC7C,MAAM,CAAC,GAAG,SAAS,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;YAC3B,IAAI,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC,CAAC;gBAAE,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QAC7C,CAAC,CAAC,CAAC;QACH,MAAM,SAAS,GAA0B,EAAE,CAAC;QAC5C,IAAI,WAAW,GAAG,CAAC,CAAC,CAAC;QACrB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACvC,IAAI,kBAAkB,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;gBAClC,MAAM,IAAI,GAAwB,EAAE,KAAK,EAAE,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC;gBACvD,IAAI,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC;oBAClB,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;oBAC5B,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;gBAC9B,CAAC;qBAAM,IAAI,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC;oBACzB,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;gBAC9B,CAAC;gBACD,IAAI,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,MAAM,GAAG,CAAC;oBAAE,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;gBACvD,WAAW,GAAG,CAAC,CAAC;YAClB,CAAC;QACH,CAAC;QAED,IAAI,WAAW,IAAI,CAAC,IAAI,WAAW,GAAG,MAAM,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACxD,MAAM,IAAI,GAAG,MAAM,CAAC,KAAK,CAAC,WAAW,GAAG,CAAC,CAAC,CAAC;YAC3C,IAAI,IAAI,CAAC,MAAM,IAAI,CAAC,EAAE,CAAC;gBACrB,SAAS,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,IAAI,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;YACvD,CAAC;iBAAM,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;gBAC7B,SAAS,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;YACtC,CAAC;QACH,CAAC;QACD,IAAI,SAAS,CAAC,MAAM;YAAE,GAAG,CAAC,SAAS,GAAG,SAAS,CAAC;IAClD,CAAC;IAGD,MAAM,OAAO,GAAG,GAAG,CAAC,MAAM,CAAC,CAAC,EAAE,EAAE,EAAE,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAAC,KAAK,EAAE,CAAC;IAChF,IAAI,OAAO,CAAC,MAAM,EAAE,CAAC;QACnB,MAAM,OAAO,GAAG,OAAO,CAAC,OAAO,CAAC,cAAc,CAAC,CAAC,KAAK,EAAE,CAAC;QACxD,MAAM,MAAM,GAAG,IAAI,GAAG,EAAU,CAAC;QACjC,OAAO,CAAC,IAAI,CAAC,2CAA2C,CAAC,CAAC,IAAI,CAAC,CAAC,EAAU,EAAE,EAAO,EAAE,EAAE;YACrF,MAAM,CAAC,GAAG,SAAS,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;YAC3B,IACE,CAAC;gBACD,CAAC,CAAC,MAAM,IAAI,EAAE;gBACd,CAAC,CAAC,MAAM,IAAI,CAAC;gBACb,CAAC,aAAa,CAAC,CAAC,CAAC;gBACjB,CAAC,kBAAkB,CAAC,CAAC,CAAC;gBACtB,CAAC,kBAAkB,CAAC,CAAC,CAAC;gBACtB,CAAC,YAAY,CAAC,CAAC,CAAC,EAChB,CAAC;gBACD,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;YAChB,CAAC;QACH,CAAC,CAAC,CAAC;QACH,IAAI,MAAM,CAAC,IAAI;YAAE,GAAG,CAAC,MAAM,GAAG,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;IAChE,CAAC;IAED,OAAO,GAAG,CAAC;AACb,CAAC"}

package/dist/assets/server/extraction/rrweb-snapshot.utils.d.ts

@@ -0,0 +1,15 @@
+export interface SnapshotNode {
+    type: number;
+    tagName?: string;
+    name?: string;
+    publicId?: string;
+    systemId?: string;
+    textContent?: string;
+    data?: string;
+    attributes?: Record<string, string>;
+    childNodes?: SnapshotNode[];
+}
+export declare function nodeToHtml(node: SnapshotNode): string;
+export declare function isFullSnapshot(payload: Record<string, any> | undefined): boolean;
+export declare function isCheckoutEvent(payload: Record<string, any> | undefined): boolean;
+export declare function snapshotTimestamp(payload: Record<string, any> | undefined): Date | undefined;

package/dist/assets/server/extraction/rrweb-snapshot.utils.js

@@ -0,0 +1,48 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.nodeToHtml = nodeToHtml;
+exports.isFullSnapshot = isFullSnapshot;
+exports.isCheckoutEvent = isCheckoutEvent;
+exports.snapshotTimestamp = snapshotTimestamp;
+function nodeToHtml(node) {
+    if (!node)
+        return '';
+    switch (node.type) {
+        case 0:
+            return (node.childNodes || []).map(nodeToHtml).join('');
+        case 1:
+            return `<!DOCTYPE ${node.name || 'html'}>`;
+        case 2: {
+            if (!node.tagName)
+                return '';
+            const attrs = Object.entries(node.attributes || {})
+                .map(([k, v]) => `${k}="${String(v).replace(/"/g, '&quot;')}"`)
+                .join(' ');
+            const open = attrs ? `<${node.tagName} ${attrs}>` : `<${node.tagName}>`;
+            const voidTags = new Set(['area', 'base', 'br', 'col', 'embed', 'hr', 'img', 'input', 'link', 'meta', 'param', 'source', 'track', 'wbr']);
+            if (voidTags.has(node.tagName.toLowerCase()))
+                return open;
+            const children = (node.childNodes || []).map(nodeToHtml).join('');
+            return `${open}${children}</${node.tagName}>`;
+        }
+        case 3:
+            return node.textContent ?? node.data ?? '';
+        case 5:
+            return `<!--${node.data || ''}-->`;
+        default:
+            return '';
+    }
+}
+function isFullSnapshot(payload) {
+    return payload?.type === 2 && !!payload.data?.node;
+}
+function isCheckoutEvent(payload) {
+    return !!payload && isFullSnapshot(payload) && payload.isCheckout === true;
+}
+function snapshotTimestamp(payload) {
+    if (payload && payload.timestamp && typeof payload.timestamp === 'number') {
+        return new Date(payload.timestamp);
+    }
+    return undefined;
+}
+//# sourceMappingURL=rrweb-snapshot.utils.js.map

package/dist/assets/server/extraction/rrweb-snapshot.utils.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"rrweb-snapshot.utils.js","sourceRoot":"","sources":["../../src/extraction/rrweb-snapshot.utils.ts"],"names":[],"mappings":";;AAkBA,gCAyBC;AAED,wCAEC;AAED,0CAEC;AAED,8CAKC;AAxCD,SAAgB,UAAU,CAAC,IAAkB;IAC3C,IAAI,CAAC,IAAI;QAAE,OAAO,EAAE,CAAC;IACrB,QAAQ,IAAI,CAAC,IAAI,EAAE,CAAC;QAClB,KAAK,CAAC;YACJ,OAAO,CAAC,IAAI,CAAC,UAAU,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,UAAU,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QAC1D,KAAK,CAAC;YACJ,OAAO,aAAa,IAAI,CAAC,IAAI,IAAI,MAAM,GAAG,CAAC;QAC7C,KAAK,CAAC,CAAC,CAAC,CAAC;YACP,IAAI,CAAC,IAAI,CAAC,OAAO;gBAAE,OAAO,EAAE,CAAC;YAC7B,MAAM,KAAK,GAAG,MAAM,CAAC,OAAO,CAAC,IAAI,CAAC,UAAU,IAAI,EAAE,CAAC;iBAChD,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,GAAG,CAAC,KAAK,MAAM,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,IAAI,EAAE,QAAQ,CAAC,GAAG,CAAC;iBAC9D,IAAI,CAAC,GAAG,CAAC,CAAC;YACb,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,IAAI,IAAI,CAAC,OAAO,IAAI,KAAK,GAAG,CAAC,CAAC,CAAC,IAAI,IAAI,CAAC,OAAO,GAAG,CAAC;YACxE,MAAM,QAAQ,GAAG,IAAI,GAAG,CAAC,CAAC,MAAM,EAAE,MAAM,EAAE,IAAI,EAAE,KAAK,EAAE,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,OAAO,EAAE,MAAM,EAAE,MAAM,EAAE,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE,KAAK,CAAC,CAAC,CAAC;YAC1I,IAAI,QAAQ,CAAC,GAAG,CAAC,IAAI,CAAC,OAAO,CAAC,WAAW,EAAE,CAAC;gBAAE,OAAO,IAAI,CAAC;YAC1D,MAAM,QAAQ,GAAG,CAAC,IAAI,CAAC,UAAU,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,UAAU,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;YAClE,OAAO,GAAG,IAAI,GAAG,QAAQ,KAAK,IAAI,CAAC,OAAO,GAAG,CAAC;QAChD,CAAC;QACD,KAAK,CAAC;YACJ,OAAO,IAAI,CAAC,WAAW,IAAI,IAAI,CAAC,IAAI,IAAI,EAAE,CAAC;QAC7C,KAAK,CAAC;YACJ,OAAO,OAAO,IAAI,CAAC,IAAI,IAAI,EAAE,KAAK,CAAC;QACrC;YACE,OAAO,EAAE,CAAC;IACd,CAAC;AACH,CAAC;AAED,SAAgB,cAAc,CAAC,OAAwC;IACrE,OAAO,OAAO,EAAE,IAAI,KAAK,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,IAAI,EAAE,IAAI,CAAC;AACrD,CAAC;AAED,SAAgB,eAAe,CAAC,OAAwC;IACtE,OAAO,CAAC,CAAC,OAAO,IAAI,cAAc,CAAC,OAAO,CAAC,IAAI,OAAO,CAAC,UAAU,KAAK,IAAI,CAAC;AAC7E,CAAC;AAED,SAAgB,iBAAiB,CAAC,OAAwC;IACxE,IAAI,OAAO,IAAI,OAAO,CAAC,SAAS,IAAI,OAAO,OAAO,CAAC,SAAS,KAAK,QAAQ,EAAE,CAAC;QAC1E,OAAO,IAAI,IAAI,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC;IACrC,CAAC;IACD,OAAO,SAAS,CAAC;AACnB,CAAC"}

package/dist/assets/server/extraction/rrweb-snapshot.utils.spec.d.ts

@@ -0,0 +1 @@
+export {};

package/dist/assets/server/extraction/rrweb-snapshot.utils.spec.js

@@ -0,0 +1,60 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+const rrweb_snapshot_utils_1 = require("./rrweb-snapshot.utils");
+describe('rrweb-snapshot.utils', () => {
+    describe('nodeToHtml', () => {
+        it('should serialize element with attributes and children', () => {
+            const node = {
+                type: 2,
+                tagName: 'div',
+                attributes: { class: 'profile', id: 'p1' },
+                childNodes: [
+                    { type: 3, textContent: 'Jensen Huang' },
+                ],
+            };
+            expect((0, rrweb_snapshot_utils_1.nodeToHtml)(node)).toContain('<div class="profile" id="p1"');
+            expect((0, rrweb_snapshot_utils_1.nodeToHtml)(node)).toContain('Jensen Huang');
+            expect((0, rrweb_snapshot_utils_1.nodeToHtml)(node)).toContain('</div>');
+        });
+        it('should handle void tags without closing tag', () => {
+            const node = { type: 2, tagName: 'img', attributes: { src: 'a.jpg' } };
+            expect((0, rrweb_snapshot_utils_1.nodeToHtml)(node)).toBe('<img src="a.jpg">');
+        });
+        it('should return empty string for null node', () => {
+            expect((0, rrweb_snapshot_utils_1.nodeToHtml)(null)).toBe('');
+        });
+        it('should escape attribute double quotes', () => {
+            const node = { type: 2, tagName: 'span', attributes: { title: 'say "hi"' } };
+            expect((0, rrweb_snapshot_utils_1.nodeToHtml)(node)).toBe('<span title="say &quot;hi&quot;"></span>');
+        });
+    });
+    describe('isFullSnapshot', () => {
+        it('should recognize rrweb full snapshot payload', () => {
+            expect((0, rrweb_snapshot_utils_1.isFullSnapshot)({ type: 2, data: { node: { type: 2, tagName: 'html' } } })).toBe(true);
+        });
+        it('should reject incremental snapshot', () => {
+            expect((0, rrweb_snapshot_utils_1.isFullSnapshot)({ type: 3, data: { source: 0 } })).toBe(false);
+        });
+        it('should handle undefined payload', () => {
+            expect((0, rrweb_snapshot_utils_1.isFullSnapshot)(undefined)).toBe(false);
+        });
+    });
+    describe('isCheckoutEvent', () => {
+        it('should return true when full snapshot is checkout', () => {
+            expect((0, rrweb_snapshot_utils_1.isCheckoutEvent)({ type: 2, isCheckout: true, data: { node: {} } })).toBe(true);
+        });
+        it('should return false for non-checkout full snapshot', () => {
+            expect((0, rrweb_snapshot_utils_1.isCheckoutEvent)({ type: 2, data: { node: {} } })).toBe(false);
+        });
+    });
+    describe('snapshotTimestamp', () => {
+        it('should convert timestamp to Date', () => {
+            const ts = Date.now();
+            expect((0, rrweb_snapshot_utils_1.snapshotTimestamp)({ timestamp: ts })).toEqual(new Date(ts));
+        });
+        it('should return undefined when no timestamp', () => {
+            expect((0, rrweb_snapshot_utils_1.snapshotTimestamp)({})).toBeUndefined();
+        });
+    });
+});
+//# sourceMappingURL=rrweb-snapshot.utils.spec.js.map

package/dist/assets/server/extraction/rrweb-snapshot.utils.spec.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"rrweb-snapshot.utils.spec.js","sourceRoot":"","sources":["../../src/extraction/rrweb-snapshot.utils.spec.ts"],"names":[],"mappings":";;AAAA,iEAAwG;AAExG,QAAQ,CAAC,sBAAsB,EAAE,GAAG,EAAE;IACpC,QAAQ,CAAC,YAAY,EAAE,GAAG,EAAE;QAC1B,EAAE,CAAC,uDAAuD,EAAE,GAAG,EAAE;YAC/D,MAAM,IAAI,GAAG;gBACX,IAAI,EAAE,CAAC;gBACP,OAAO,EAAE,KAAK;gBACd,UAAU,EAAE,EAAE,KAAK,EAAE,SAAS,EAAE,EAAE,EAAE,IAAI,EAAE;gBAC1C,UAAU,EAAE;oBACV,EAAE,IAAI,EAAE,CAAC,EAAE,WAAW,EAAE,cAAc,EAAE;iBACzC;aACK,CAAC;YACT,MAAM,CAAC,IAAA,iCAAU,EAAC,IAAI,CAAC,CAAC,CAAC,SAAS,CAAC,8BAA8B,CAAC,CAAC;YACnE,MAAM,CAAC,IAAA,iCAAU,EAAC,IAAI,CAAC,CAAC,CAAC,SAAS,CAAC,cAAc,CAAC,CAAC;YACnD,MAAM,CAAC,IAAA,iCAAU,EAAC,IAAI,CAAC,CAAC,CAAC,SAAS,CAAC,QAAQ,CAAC,CAAC;QAC/C,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,6CAA6C,EAAE,GAAG,EAAE;YACrD,MAAM,IAAI,GAAG,EAAE,IAAI,EAAE,CAAC,EAAE,OAAO,EAAE,KAAK,EAAE,UAAU,EAAE,EAAE,GAAG,EAAE,OAAO,EAAE,EAAS,CAAC;YAC9E,MAAM,CAAC,IAAA,iCAAU,EAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,mBAAmB,CAAC,CAAC;QACrD,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,0CAA0C,EAAE,GAAG,EAAE;YAClD,MAAM,CAAC,IAAA,iCAAU,EAAC,IAAW,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QAC3C,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,uCAAuC,EAAE,GAAG,EAAE;YAC/C,MAAM,IAAI,GAAG,EAAE,IAAI,EAAE,CAAC,EAAE,OAAO,EAAE,MAAM,EAAE,UAAU,EAAE,EAAE,KAAK,EAAE,UAAU,EAAE,EAAS,CAAC;YACpF,MAAM,CAAC,IAAA,iCAAU,EAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,0CAA0C,CAAC,CAAC;QAC5E,CAAC,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;IAEH,QAAQ,CAAC,gBAAgB,EAAE,GAAG,EAAE;QAC9B,EAAE,CAAC,8CAA8C,EAAE,GAAG,EAAE;YACtD,MAAM,CAAC,IAAA,qCAAc,EAAC,EAAE,IAAI,EAAE,CAAC,EAAE,IAAI,EAAE,EAAE,IAAI,EAAE,EAAE,IAAI,EAAE,CAAC,EAAE,OAAO,EAAE,MAAM,EAAE,EAAE,EAAS,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACtG,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,oCAAoC,EAAE,GAAG,EAAE;YAC5C,MAAM,CAAC,IAAA,qCAAc,EAAC,EAAE,IAAI,EAAE,CAAC,EAAE,IAAI,EAAE,EAAE,MAAM,EAAE,CAAC,EAAE,EAAS,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QAC9E,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,iCAAiC,EAAE,GAAG,EAAE;YACzC,MAAM,CAAC,IAAA,qCAAc,EAAC,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QAChD,CAAC,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;IAEH,QAAQ,CAAC,iBAAiB,EAAE,GAAG,EAAE;QAC/B,EAAE,CAAC,mDAAmD,EAAE,GAAG,EAAE;YAC3D,MAAM,CAAC,IAAA,sCAAe,EAAC,EAAE,IAAI,EAAE,CAAC,EAAE,UAAU,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,IAAI,EAAE,EAAE,EAAE,EAAS,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAC/F,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,oDAAoD,EAAE,GAAG,EAAE;YAC5D,MAAM,CAAC,IAAA,sCAAe,EAAC,EAAE,IAAI,EAAE,CAAC,EAAE,IAAI,EAAE,EAAE,IAAI,EAAE,EAAE,EAAE,EAAS,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QAC9E,CAAC,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;IAEH,QAAQ,CAAC,mBAAmB,EAAE,GAAG,EAAE;QACjC,EAAE,CAAC,kCAAkC,EAAE,GAAG,EAAE;YAC1C,MAAM,EAAE,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;YACtB,MAAM,CAAC,IAAA,wCAAiB,EAAC,EAAE,SAAS,EAAE,EAAE,EAAS,CAAC,CAAC,CAAC,OAAO,CAAC,IAAI,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC;QAC5E,CAAC,CAAC,CAAC;QAEH,EAAE,CAAC,2CAA2C,EAAE,GAAG,EAAE;YACnD,MAAM,CAAC,IAAA,wCAAiB,EAAC,EAAS,CAAC,CAAC,CAAC,aAAa,EAAE,CAAC;QACvD,CAAC,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;AACL,CAAC,CAAC,CAAC"}

package/dist/assets/server/extraction/script-rule.d.ts

@@ -0,0 +1,14 @@
+export interface ScriptFieldRule {
+    selector: string;
+    attr?: string;
+    regex?: string;
+}
+export interface ScriptRulePayload {
+    engine?: string;
+    fields?: Record<string, ScriptFieldRule>;
+    lists?: Record<string, {
+        selector: string;
+        fields: Record<string, ScriptFieldRule>;
+    }>;
+}
+export declare function applyScriptRule(html: string, payload: ScriptRulePayload): Record<string, any>;

package/dist/assets/server/extraction/script-rule.js

@@ -0,0 +1,49 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.applyScriptRule = applyScriptRule;
+const cheerio_1 = require("cheerio");
+const linkedin_extractor_1 = require("./linkedin-extractor");
+function readField($, root, rule) {
+    const el = root.find(rule.selector).first();
+    if (!el.length)
+        return null;
+    let value;
+    if (!rule.attr || rule.attr === 'text')
+        value = el.text().trim();
+    else if (rule.attr === 'html')
+        value = el.html();
+    else
+        value = el.attr(rule.attr) ?? null;
+    if (value == null)
+        return null;
+    if (rule.regex) {
+        const m = value.match(new RegExp(rule.regex, 's'));
+        value = m?.[1]?.trim() ?? null;
+    }
+    return value || null;
+}
+function applyScriptRule(html, payload) {
+    if (payload.engine === 'linkedin-profile') {
+        return (0, linkedin_extractor_1.extractLinkedInProfile)(html);
+    }
+    const $ = (0, cheerio_1.load)(html);
+    const out = {};
+    for (const [field, rule] of Object.entries(payload.fields || {})) {
+        out[field] = readField($, $('body').length ? $('body') : $.root(), rule);
+    }
+    for (const [listName, listRule] of Object.entries(payload.lists || {})) {
+        const items = [];
+        $(listRule.selector).each((_i, el) => {
+            const item = {};
+            for (const [field, rule] of Object.entries(listRule.fields)) {
+                item[field] = readField($, $(el), rule);
+            }
+            if (Object.values(item).some((v) => v != null))
+                items.push(item);
+        });
+        if (items.length)
+            out[listName] = items;
+    }
+    return out;
+}
+//# sourceMappingURL=script-rule.js.map

package/dist/assets/server/extraction/script-rule.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"script-rule.js","sourceRoot":"","sources":["../../src/extraction/script-rule.ts"],"names":[],"mappings":";;AA0CA,0CAuBC;AAjED,qCAA2C;AAC3C,6DAA8D;AA0B9D,SAAS,SAAS,CAAC,CAAa,EAAE,IAAS,EAAE,IAAqB;IAChE,MAAM,EAAE,GAAG,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,KAAK,EAAE,CAAC;IAC5C,IAAI,CAAC,EAAE,CAAC,MAAM;QAAE,OAAO,IAAI,CAAC;IAC5B,IAAI,KAAoB,CAAC;IACzB,IAAI,CAAC,IAAI,CAAC,IAAI,IAAI,IAAI,CAAC,IAAI,KAAK,MAAM;QAAE,KAAK,GAAG,EAAE,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,CAAC;SAC5D,IAAI,IAAI,CAAC,IAAI,KAAK,MAAM;QAAE,KAAK,GAAG,EAAE,CAAC,IAAI,EAAE,CAAC;;QAC5C,KAAK,GAAG,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC;IACxC,IAAI,KAAK,IAAI,IAAI;QAAE,OAAO,IAAI,CAAC;IAC/B,IAAI,IAAI,CAAC,KAAK,EAAE,CAAC;QACf,MAAM,CAAC,GAAG,KAAK,CAAC,KAAK,CAAC,IAAI,MAAM,CAAC,IAAI,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC,CAAC;QACnD,KAAK,GAAG,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,IAAI,IAAI,CAAC;IACjC,CAAC;IACD,OAAO,KAAK,IAAI,IAAI,CAAC;AACvB,CAAC;AAED,SAAgB,eAAe,CAAC,IAAY,EAAE,OAA0B;IAEtE,IAAI,OAAO,CAAC,MAAM,KAAK,kBAAkB,EAAE,CAAC;QAC1C,OAAO,IAAA,2CAAsB,EAAC,IAAI,CAAC,CAAC;IACtC,CAAC;IAED,MAAM,CAAC,GAAG,IAAA,cAAI,EAAC,IAAI,CAAC,CAAC;IACrB,MAAM,GAAG,GAAwB,EAAE,CAAC;IACpC,KAAK,MAAM,CAAC,KAAK,EAAE,IAAI,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,MAAM,IAAI,EAAE,CAAC,EAAE,CAAC;QACjE,GAAG,CAAC,KAAK,CAAC,GAAG,SAAS,CAAC,CAAC,EAAE,CAAC,CAAC,MAAM,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,EAAE,IAAI,CAAC,CAAC;IAC3E,CAAC;IACD,KAAK,MAAM,CAAC,QAAQ,EAAE,QAAQ,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,KAAK,IAAI,EAAE,CAAC,EAAE,CAAC;QACvE,MAAM,KAAK,GAA0B,EAAE,CAAC;QACxC,CAAC,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC,IAAI,CAAC,CAAC,EAAE,EAAE,EAAE,EAAE,EAAE;YACnC,MAAM,IAAI,GAAwB,EAAE,CAAC;YACrC,KAAK,MAAM,CAAC,KAAK,EAAE,IAAI,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,QAAQ,CAAC,MAAM,CAAC,EAAE,CAAC;gBAC5D,IAAI,CAAC,KAAK,CAAC,GAAG,SAAS,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,CAAC,EAAE,IAAI,CAAC,CAAC;YAC1C,CAAC;YACD,IAAI,MAAM,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,IAAI,IAAI,CAAC;gBAAE,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACnE,CAAC,CAAC,CAAC;QACH,IAAI,KAAK,CAAC,MAAM;YAAE,GAAG,CAAC,QAAQ,CAAC,GAAG,KAAK,CAAC;IAC1C,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC"}

package/dist/assets/server/extraction/snapshot-analysis.service.d.ts

@@ -0,0 +1,23 @@
+import { OnModuleInit } from '@nestjs/common';
+import { Repository } from 'typeorm';
+import { RecordingEventBus } from '../recording/recording-event-bus.service';
+import { RecordingSnapshotService } from '../recording/recording-snapshot.service';
+import { ParseLog } from './entities/parse-log.entity';
+import { AdapterService } from './adapter.service';
+import { ExtractionService } from './extraction.service';
+import { AnomalyReportService } from './anomaly-report.service';
+export declare class SnapshotAnalysisService implements OnModuleInit {
+    private readonly eventBus;
+    private readonly snapshotService;
+    private readonly adapterService;
+    private readonly extractionService;
+    private readonly anomalyReport;
+    private readonly parseLogRepo;
+    private debounceTimers;
+    constructor(eventBus: RecordingEventBus, snapshotService: RecordingSnapshotService, adapterService: AdapterService, extractionService: ExtractionService, anomalyReport: AnomalyReportService, parseLogRepo: Repository<ParseLog>);
+    onModuleInit(): void;
+    private debouncedAnalyze;
+    private analyzeSnapshot;
+    private detectAnomaly;
+    private extractDomain;
+}

package/dist/assets/server/extraction/snapshot-analysis.service.js

@@ -0,0 +1,147 @@
+"use strict";
+var __decorate = (this && this.__decorate) || function (decorators, target, key, desc) {
+    var c = arguments.length, r = c < 3 ? target : desc === null ? desc = Object.getOwnPropertyDescriptor(target, key) : desc, d;
+    if (typeof Reflect === "object" && typeof Reflect.decorate === "function") r = Reflect.decorate(decorators, target, key, desc);
+    else for (var i = decorators.length - 1; i >= 0; i--) if (d = decorators[i]) r = (c < 3 ? d(r) : c > 3 ? d(target, key, r) : d(target, key)) || r;
+    return c > 3 && r && Object.defineProperty(target, key, r), r;
+};
+var __metadata = (this && this.__metadata) || function (k, v) {
+    if (typeof Reflect === "object" && typeof Reflect.metadata === "function") return Reflect.metadata(k, v);
+};
+var __param = (this && this.__param) || function (paramIndex, decorator) {
+    return function (target, key) { decorator(target, key, paramIndex); }
+};
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.SnapshotAnalysisService = void 0;
+const common_1 = require("@nestjs/common");
+const typeorm_1 = require("@nestjs/typeorm");
+const typeorm_2 = require("typeorm");
+const recording_event_bus_service_1 = require("../recording/recording-event-bus.service");
+const recording_snapshot_service_1 = require("../recording/recording-snapshot.service");
+const parse_log_entity_1 = require("./entities/parse-log.entity");
+const adapter_service_1 = require("./adapter.service");
+const extraction_service_1 = require("./extraction.service");
+const anomaly_report_service_1 = require("./anomaly-report.service");
+const CRITICAL_FIELDS = {
+    talent: ['name'],
+    job: ['title'],
+    company: ['name'],
+};
+let SnapshotAnalysisService = class SnapshotAnalysisService {
+    eventBus;
+    snapshotService;
+    adapterService;
+    extractionService;
+    anomalyReport;
+    parseLogRepo;
+    debounceTimers = new Map();
+    constructor(eventBus, snapshotService, adapterService, extractionService, anomalyReport, parseLogRepo) {
+        this.eventBus = eventBus;
+        this.snapshotService = snapshotService;
+        this.adapterService = adapterService;
+        this.extractionService = extractionService;
+        this.anomalyReport = anomalyReport;
+        this.parseLogRepo = parseLogRepo;
+    }
+    onModuleInit() {
+        this.eventBus.on('recording.snapshot.created', (evt) => {
+            this.debouncedAnalyze(evt.sessionId, evt.snapshotId);
+        });
+    }
+    debouncedAnalyze(sessionId, snapshotId) {
+        const existing = this.debounceTimers.get(sessionId);
+        if (existing)
+            clearTimeout(existing);
+        this.debounceTimers.set(sessionId, setTimeout(() => {
+            this.debounceTimers.delete(sessionId);
+            this.analyzeSnapshot(snapshotId).catch((err) => {
+                console.error('[SnapshotAnalysis] analyze failed', err);
+            });
+        }, 1000));
+    }
+    async analyzeSnapshot(snapshotId) {
+        const snapshot = await this.snapshotService.getById(snapshotId);
+        if (!snapshot)
+            return;
+        const recent = await this.parseLogRepo.findOne({
+            where: { sessionId: snapshot.sessionId },
+            order: { createdAt: 'DESC' },
+        });
+        if (recent &&
+            recent.status === 'ok' &&
+            recent.createdAt &&
+            snapshot.snapshotTimestamp.getTime() - recent.createdAt.getTime() < 5000) {
+            return;
+        }
+        const analysis = await this.adapterService.analyzeSnapshot(snapshot);
+        const domain = snapshot.domain ? (0, extraction_service_1.etldPlus1)(snapshot.domain) : this.extractDomain(snapshot.sourceUrl);
+        const entity = await this.extractionService.incrementalUpsertEntity(analysis.pageClass, domain, snapshot.sourceUrl || '', analysis.extracted, [], snapshot.sessionId, analysis.adapterVersion);
+        const anomaly = this.detectAnomaly(analysis.pageClass, analysis.extracted, analysis.matchedRules, domain);
+        const status = anomaly.isAnomaly ? 'anomaly' : entity ? 'ok' : analysis.matchedRules.length ? 'empty' : 'no-rule';
+        await this.snapshotService.updateParseStatus(snapshot.id, status, anomaly.score);
+        const parseLog = await this.parseLogRepo.save(this.parseLogRepo.create({
+            sessionId: snapshot.sessionId,
+            sourceUrl: snapshot.sourceUrl || '',
+            domain,
+            pageClass: analysis.pageClass,
+            status,
+            hasSnapshot: true,
+            snapshotId: snapshot.id,
+            adapterVersion: analysis.adapterVersion,
+            matchedRuleIds: analysis.matchedRules.map((r) => `${r.platform}:${r.version}:${r.id}`),
+            extractedFields: Object.fromEntries(Object.entries(analysis.extracted).filter(([, v]) => v != null)),
+            missingFields: Object.entries(analysis.extracted).filter(([, v]) => v == null).map(([k]) => k),
+            entityId: entity?.id ?? null,
+            anomalyScore: anomaly.score,
+        }));
+        if (status === 'anomaly') {
+            await this.anomalyReport.report(snapshot.sessionId, snapshot.id, parseLog.id, anomaly.reason);
+        }
+    }
+    detectAnomaly(pageClass, extracted, matchedRules, domain) {
+        const reasons = [];
+        if (pageClass === 'unknown' && matchedRules.length > 0) {
+            reasons.push('enabled site but unknown page class');
+        }
+        if (matchedRules.length === 0 && domain) {
+            reasons.push('no adapter rules matched for enabled domain');
+        }
+        const critical = CRITICAL_FIELDS[pageClass] || [];
+        for (const field of critical) {
+            if (!extracted[field])
+                reasons.push(`missing critical field: ${field}`);
+        }
+        const nonEmptyFields = Object.values(extracted).filter((v) => v != null && String(v).trim() !== '').length;
+        if (nonEmptyFields === 0 && matchedRules.length > 0) {
+            reasons.push('all extracted fields are empty after adapter rules');
+        }
+        const score = Math.min(1, reasons.length * 0.25 + 0.1);
+        return {
+            isAnomaly: reasons.length > 0,
+            score,
+            reason: reasons.join('; ') || 'ok',
+        };
+    }
+    extractDomain(sourceUrl) {
+        if (!sourceUrl)
+            return '';
+        try {
+            return (0, extraction_service_1.etldPlus1)(new URL(sourceUrl).hostname);
+        }
+        catch {
+            return '';
+        }
+    }
+};
+exports.SnapshotAnalysisService = SnapshotAnalysisService;
+exports.SnapshotAnalysisService = SnapshotAnalysisService = __decorate([
+    (0, common_1.Injectable)(),
+    __param(5, (0, typeorm_1.InjectRepository)(parse_log_entity_1.ParseLog)),
+    __metadata("design:paramtypes", [recording_event_bus_service_1.RecordingEventBus,
+        recording_snapshot_service_1.RecordingSnapshotService,
+        adapter_service_1.AdapterService,
+        extraction_service_1.ExtractionService,
+        anomaly_report_service_1.AnomalyReportService,
+        typeorm_2.Repository])
+], SnapshotAnalysisService);
+//# sourceMappingURL=snapshot-analysis.service.js.map