hunter-mate 0.1.1 → 0.1.3

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

@@ -0,0 +1,81 @@
+"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.ExtractionRuleService = void 0;
+const common_1 = require("@nestjs/common");
+const typeorm_1 = require("@nestjs/typeorm");
+const typeorm_2 = require("typeorm");
+const extraction_rule_entity_1 = require("./extraction-rule.entity");
+let ExtractionRuleService = class ExtractionRuleService {
+    repo;
+    constructor(repo) {
+        this.repo = repo;
+    }
+    async onModuleInit() {
+        await this.seedLinkedInRule();
+        await this.disableStaleLinkedInRules();
+    }
+    async seedLinkedInRule() {
+        const exists = await this.repo.findOne({ where: { ruleId: 'linkedin-profile-default' } });
+        if (exists)
+            return;
+        await this.repo.save(this.repo.create({
+            ruleId: 'linkedin-profile-default',
+            version: '0.1.0',
+            targetHost: '^www\\.linkedin\\.com$',
+            targetPathPattern: '^/in/',
+            ruleType: 'regex',
+            entityType: 'talent',
+            rulePayload: {
+                mapping: {
+                    name: '<a[^>]*href="[^"]*linkedin\\.com/in/[^"]+"[^>]*>.*?<div[^>]*aria-label="([^"]+)"',
+                    headline: '<a[^>]*href="[^"]*linkedin\\.com/in/[^"]+"[^>]*>.*?<p[^>]*><span>([^<]+)</span>',
+                },
+            },
+            extractedSchema: {
+                name: 'string',
+                headline: 'string',
+            },
+            priority: 1000,
+            enabled: true,
+        }));
+    }
+    async disableStaleLinkedInRules() {
+        const staleIds = [
+            'linkedin-profile',
+            'linkedin-profile-regex',
+            'linkedin-profile-regex-v2',
+            'linkedin-profile-regex-v3',
+        ];
+        await this.repo.update({ ruleId: (0, typeorm_2.In)(staleIds) }, { enabled: false });
+    }
+    create(dto) {
+        const entity = this.repo.create({
+            ...dto,
+            priority: dto.priority ?? 0,
+            enabled: dto.enabled ?? true,
+        });
+        return this.repo.save(entity);
+    }
+    findEnabled() {
+        return this.repo.find({ where: { enabled: true }, order: { priority: 'DESC' } });
+    }
+};
+exports.ExtractionRuleService = ExtractionRuleService;
+exports.ExtractionRuleService = ExtractionRuleService = __decorate([
+    (0, common_1.Injectable)(),
+    __param(0, (0, typeorm_1.InjectRepository)(extraction_rule_entity_1.ExtractionRule)),
+    __metadata("design:paramtypes", [typeorm_2.Repository])
+], ExtractionRuleService);
+//# sourceMappingURL=extraction-rule.service.js.map

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

@@ -0,0 +1 @@
+{"version":3,"file":"extraction-rule.service.js","sourceRoot":"","sources":["../../src/extraction/extraction-rule.service.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;AAAA,2CAA0D;AAC1D,6CAAmD;AACnD,qCAAyC;AACzC,qEAA0D;AAgBnD,IAAM,qBAAqB,GAA3B,MAAM,qBAAqB;IAGb;IAFnB,YAEmB,IAAgC;QAAhC,SAAI,GAAJ,IAAI,CAA4B;IAChD,CAAC;IAEJ,KAAK,CAAC,YAAY;QAChB,MAAM,IAAI,CAAC,gBAAgB,EAAE,CAAC;QAC9B,MAAM,IAAI,CAAC,yBAAyB,EAAE,CAAC;IACzC,CAAC;IAEO,KAAK,CAAC,gBAAgB;QAC5B,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,KAAK,EAAE,EAAE,MAAM,EAAE,0BAA0B,EAAE,EAAE,CAAC,CAAC;QAC1F,IAAI,MAAM;YAAE,OAAO;QACnB,MAAM,IAAI,CAAC,IAAI,CAAC,IAAI,CAClB,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC;YACf,MAAM,EAAE,0BAA0B;YAClC,OAAO,EAAE,OAAO;YAChB,UAAU,EAAE,wBAAwB;YACpC,iBAAiB,EAAE,OAAO;YAC1B,QAAQ,EAAE,OAAO;YACjB,UAAU,EAAE,QAAQ;YACpB,WAAW,EAAE;gBACX,OAAO,EAAE;oBACP,IAAI,EAAE,kFAAkF;oBACxF,QAAQ,EAAE,iFAAiF;iBAC5F;aACF;YACD,eAAe,EAAE;gBACf,IAAI,EAAE,QAAQ;gBACd,QAAQ,EAAE,QAAQ;aACnB;YACD,QAAQ,EAAE,IAAI;YACd,OAAO,EAAE,IAAI;SACd,CAAC,CACH,CAAC;IACJ,CAAC;IAEO,KAAK,CAAC,yBAAyB;QACrC,MAAM,QAAQ,GAAG;YACf,kBAAkB;YAClB,wBAAwB;YACxB,2BAA2B;YAC3B,2BAA2B;SAC5B,CAAC;QAEF,MAAM,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,MAAM,EAAE,IAAA,YAAE,EAAC,QAAQ,CAAC,EAAE,EAAE,EAAE,OAAO,EAAE,KAAK,EAAE,CAAC,CAAC;IACvE,CAAC;IAED,MAAM,CAAC,GAAkB;QACvB,MAAM,MAAM,GAAG,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC;YAC9B,GAAG,GAAG;YACN,QAAQ,EAAE,GAAG,CAAC,QAAQ,IAAI,CAAC;YAC3B,OAAO,EAAE,GAAG,CAAC,OAAO,IAAI,IAAI;SAC7B,CAAC,CAAC;QACH,OAAO,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;IAChC,CAAC;IAED,WAAW;QACT,OAAO,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,EAAE,OAAO,EAAE,IAAI,EAAE,EAAE,KAAK,EAAE,EAAE,QAAQ,EAAE,MAAM,EAAE,EAAE,CAAC,CAAC;IACnF,CAAC;CACF,CAAA;AA7DY,sDAAqB;gCAArB,qBAAqB;IADjC,IAAA,mBAAU,GAAE;IAGR,WAAA,IAAA,0BAAgB,EAAC,uCAAc,CAAC,CAAA;qCACV,oBAAU;GAHxB,qBAAqB,CA6DjC"}

package/dist/assets/server/extraction/extraction.module.d.ts

@@ -0,0 +1,2 @@
+export declare class ExtractionModule {
+}

package/dist/assets/server/extraction/extraction.module.js

@@ -0,0 +1,43 @@
+"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;
+};
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.ExtractionModule = void 0;
+const common_1 = require("@nestjs/common");
+const typeorm_1 = require("@nestjs/typeorm");
+const extension_core_module_1 = require("../extension/extension-core.module");
+const extraction_rule_entity_1 = require("./extraction-rule.entity");
+const extraction_service_1 = require("./extraction.service");
+const extraction_rule_service_1 = require("./extraction-rule.service");
+const extraction_rule_controller_1 = require("./extraction-rule.controller");
+const extraction_entity_controller_1 = require("./extraction-entity.controller");
+const adapter_service_1 = require("./adapter.service");
+const parse_log_entity_1 = require("./entities/parse-log.entity");
+const talent_profile_entity_1 = require("./entities/talent-profile.entity");
+const job_posting_entity_1 = require("./entities/job-posting.entity");
+const company_profile_entity_1 = require("./entities/company-profile.entity");
+let ExtractionModule = class ExtractionModule {
+};
+exports.ExtractionModule = ExtractionModule;
+exports.ExtractionModule = ExtractionModule = __decorate([
+    (0, common_1.Module)({
+        imports: [
+            extension_core_module_1.ExtensionCoreModule,
+            typeorm_1.TypeOrmModule.forFeature([
+                extraction_rule_entity_1.ExtractionRule,
+                parse_log_entity_1.ParseLog,
+                talent_profile_entity_1.TalentProfile,
+                job_posting_entity_1.JobPosting,
+                company_profile_entity_1.CompanyProfile,
+            ]),
+        ],
+        controllers: [extraction_rule_controller_1.ExtractionRuleController, extraction_entity_controller_1.ExtractionEntityController],
+        providers: [extraction_service_1.ExtractionService, extraction_rule_service_1.ExtractionRuleService, adapter_service_1.AdapterService],
+        exports: [extraction_service_1.ExtractionService, extraction_rule_service_1.ExtractionRuleService, adapter_service_1.AdapterService],
+    })
+], ExtractionModule);
+//# sourceMappingURL=extraction.module.js.map

package/dist/assets/server/extraction/extraction.module.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"extraction.module.js","sourceRoot":"","sources":["../../src/extraction/extraction.module.ts"],"names":[],"mappings":";;;;;;;;;AAAA,2CAAwC;AACxC,6CAAgD;AAChD,8EAAyE;AACzE,qEAA0D;AAC1D,6DAAyD;AACzD,uEAAkE;AAClE,6EAAwE;AACxE,iFAA4E;AAC5E,uDAAmD;AACnD,kEAAuD;AACvD,4EAAiE;AACjE,sEAA2D;AAC3D,8EAAmE;AAiB5D,IAAM,gBAAgB,GAAtB,MAAM,gBAAgB;CAAG,CAAA;AAAnB,4CAAgB;2BAAhB,gBAAgB;IAf5B,IAAA,eAAM,EAAC;QACN,OAAO,EAAE;YACP,2CAAmB;YACnB,uBAAa,CAAC,UAAU,CAAC;gBACvB,uCAAc;gBACd,2BAAQ;gBACR,qCAAa;gBACb,+BAAU;gBACV,uCAAc;aACf,CAAC;SACH;QACD,WAAW,EAAE,CAAC,qDAAwB,EAAE,yDAA0B,CAAC;QACnE,SAAS,EAAE,CAAC,sCAAiB,EAAE,+CAAqB,EAAE,gCAAc,CAAC;QACrE,OAAO,EAAE,CAAC,sCAAiB,EAAE,+CAAqB,EAAE,gCAAc,CAAC;KACpE,CAAC;GACW,gBAAgB,CAAG"}

package/dist/assets/server/extraction/extraction.service.d.ts

@@ -0,0 +1,62 @@
+import { Repository } from 'typeorm';
+import { RecordingEvent } from '../recording/entities/recording-event.entity';
+import { ExtractionRule } from './extraction-rule.entity';
+import { ExtractionRuleService } from './extraction-rule.service';
+import { ParseLog } from './entities/parse-log.entity';
+import { TalentProfile } from './entities/talent-profile.entity';
+import { JobPosting } from './entities/job-posting.entity';
+import { CompanyProfile } from './entities/company-profile.entity';
+export declare class ExtractionService {
+    private readonly parseLogRepo;
+    private readonly talentRepo;
+    private readonly jobRepo;
+    private readonly companyRepo;
+    private readonly ruleService;
+    constructor(parseLogRepo: Repository<ParseLog>, talentRepo: Repository<TalentProfile>, jobRepo: Repository<JobPosting>, companyRepo: Repository<CompanyProfile>, ruleService: ExtractionRuleService);
+    applyRules(events: RecordingEvent[], rules: ExtractionRule[]): {
+        sourceUrl: string;
+        matchedRules: {
+            ruleId: string;
+            version: string;
+        }[];
+        extracted: Record<string, any>;
+    } | {
+        sourceUrl: string | null;
+        hasSnapshot: boolean;
+        extracted: null;
+        message: string;
+    };
+    applyRulesToHtml(sourceUrl: string, html: string, rules: ExtractionRule[]): {
+        sourceUrl: string;
+        matchedRules: {
+            ruleId: string;
+            version: string;
+        }[];
+        extracted: Record<string, any>;
+    };
+    parseSession(sessionId: string): Promise<{
+        status: string;
+        message: string;
+    } | {
+        status: string;
+        pageClass: string;
+        matchedRules: {
+            ruleId: string;
+            version: string;
+        }[];
+        jsonLdTypes: string[];
+        extracted: Record<string, any>;
+        entity: {
+            id: number;
+            type: string;
+            isNew: boolean;
+        } | null;
+    }>;
+    private upsertEntity;
+    saveAdapterEntity(sessionId: string, sourceUrl: string, entity: Record<string, any>): Promise<{
+        id: number;
+        type: string;
+    } | null>;
+    private writeParseLog;
+    suggestRule(sessionId: string, events: RecordingEvent[]): Record<string, any>;
+}

package/dist/assets/server/extraction/extraction.service.js

@@ -0,0 +1,347 @@
+"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.ExtractionService = void 0;
+const common_1 = require("@nestjs/common");
+const typeorm_1 = require("@nestjs/typeorm");
+const typeorm_2 = require("typeorm");
+const cheerio_1 = require("cheerio");
+const recording_event_entity_1 = require("../recording/entities/recording-event.entity");
+const extraction_rule_service_1 = require("./extraction-rule.service");
+const parse_log_entity_1 = require("./entities/parse-log.entity");
+const talent_profile_entity_1 = require("./entities/talent-profile.entity");
+const job_posting_entity_1 = require("./entities/job-posting.entity");
+const company_profile_entity_1 = require("./entities/company-profile.entity");
+const structured_data_1 = require("./structured-data");
+const script_rule_1 = require("./script-rule");
+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 extractSnapshotHtml(events) {
+    for (const event of events) {
+        if (event.eventType !== 'rrweb')
+            continue;
+        const payload = event.payload;
+        if (payload?.type === 2 && payload.data?.node) {
+            return nodeToHtml(payload.data.node);
+        }
+    }
+    return null;
+}
+function matchRule(rule, url) {
+    try {
+        const hostPattern = new RegExp(rule.targetHost, 'i');
+        const pathPattern = new RegExp(rule.targetPathPattern, 'i');
+        const parsed = new URL(url);
+        return hostPattern.test(parsed.hostname) && pathPattern.test(parsed.pathname + parsed.search);
+    }
+    catch {
+        return false;
+    }
+}
+function applySelectorRule(payload, html) {
+    const $ = (0, cheerio_1.load)(html);
+    const result = {};
+    const mapping = payload.mapping || {};
+    for (const [field, selector] of Object.entries(mapping)) {
+        const el = $(selector).first();
+        result[field] = el.text().trim() || el.attr('content') || el.attr('value') || null;
+    }
+    return result;
+}
+function applyRegexRule(payload, html) {
+    const result = {};
+    const mapping = payload.mapping || {};
+    for (const [field, pattern] of Object.entries(mapping)) {
+        const match = html.match(new RegExp(pattern, 's'));
+        result[field] = match?.[1] || null;
+    }
+    return result;
+}
+function etldPlus1(hostname) {
+    const parts = hostname.toLowerCase().split('.').filter(Boolean);
+    if (parts.length <= 2)
+        return parts.join('.');
+    const twoLevelTld = new Set(['com.cn', 'co.uk', 'com.au', 'co.jp', 'com.hk', 'com.sg']);
+    const last2 = parts.slice(-2).join('.');
+    if (twoLevelTld.has(last2))
+        return parts.slice(-3).join('.');
+    return last2;
+}
+function normalizeUrlForDedup(url) {
+    try {
+        const u = new URL(url);
+        return `${u.hostname.toLowerCase()}${u.pathname.replace(/\/+$/, '')}${u.hash}`;
+    }
+    catch {
+        return url;
+    }
+}
+let ExtractionService = class ExtractionService {
+    parseLogRepo;
+    talentRepo;
+    jobRepo;
+    companyRepo;
+    ruleService;
+    constructor(parseLogRepo, talentRepo, jobRepo, companyRepo, ruleService) {
+        this.parseLogRepo = parseLogRepo;
+        this.talentRepo = talentRepo;
+        this.jobRepo = jobRepo;
+        this.companyRepo = companyRepo;
+        this.ruleService = ruleService;
+    }
+    applyRules(events, rules) {
+        const sourceUrl = events.find((e) => e.sourceUrl)?.sourceUrl || null;
+        const html = extractSnapshotHtml(events);
+        if (!sourceUrl || !html) {
+            return { sourceUrl, hasSnapshot: !!html, extracted: null, message: 'Missing sourceUrl or rrweb full snapshot.' };
+        }
+        return this.applyRulesToHtml(sourceUrl, html, rules);
+    }
+    applyRulesToHtml(sourceUrl, html, rules) {
+        const matched = rules
+            .filter((r) => r.enabled && matchRule(r, sourceUrl))
+            .sort((a, b) => b.priority - a.priority);
+        const extracted = {};
+        for (const rule of matched) {
+            if (rule.ruleType === 'selector')
+                Object.assign(extracted, applySelectorRule(rule.rulePayload, html));
+            else if (rule.ruleType === 'regex')
+                Object.assign(extracted, applyRegexRule(rule.rulePayload, html));
+            else if (rule.ruleType === 'script')
+                Object.assign(extracted, (0, script_rule_1.applyScriptRule)(html, rule.rulePayload));
+        }
+        return {
+            sourceUrl,
+            matchedRules: matched.map((r) => ({ ruleId: r.ruleId, version: r.version })),
+            extracted,
+        };
+    }
+    async parseSession(sessionId) {
+        const eventRepo = this.parseLogRepo.manager.getRepository(recording_event_entity_1.RecordingEvent);
+        const events = await eventRepo.find({ where: { sessionId }, order: { createdAt: 'ASC' } });
+        const sourceUrl = events.find((e) => e.sourceUrl)?.sourceUrl || null;
+        const html = extractSnapshotHtml(events);
+        const domain = sourceUrl ? etldPlus1(new URL(sourceUrl).hostname) : '';
+        const result = { sessionId, sourceUrl, domain };
+        if (!sourceUrl || !html) {
+            await this.writeParseLog({ sessionId, sourceUrl: sourceUrl || '', domain, pageClass: 'unknown', status: 'no-snapshot', hasSnapshot: false });
+            return { ...result, status: 'no-snapshot', message: 'Missing sourceUrl or rrweb full snapshot.' };
+        }
+        const jsonld = (0, structured_data_1.extractJsonLd)(html);
+        const og = (0, structured_data_1.extractOpenGraph)(html);
+        const rules = await this.ruleService.findEnabled();
+        const matched = rules.filter((r) => r.enabled && matchRule(r, sourceUrl)).sort((a, b) => b.priority - a.priority);
+        const ruleExtracted = {};
+        for (const rule of matched) {
+            if (rule.ruleType === 'selector')
+                Object.assign(ruleExtracted, applySelectorRule(rule.rulePayload, html));
+            else if (rule.ruleType === 'regex')
+                Object.assign(ruleExtracted, applyRegexRule(rule.rulePayload, html));
+            else if (rule.ruleType === 'script')
+                Object.assign(ruleExtracted, (0, script_rule_1.applyScriptRule)(html, rule.rulePayload));
+        }
+        const merged = { ...og, ...(jsonld?.data || {}), ...ruleExtracted };
+        for (const [k, v] of Object.entries({ ...(jsonld?.data || {}), ...(og || {}) })) {
+            if (ruleExtracted[k] == null && v != null)
+                merged[k] = v;
+        }
+        const pageClass = matched.find((r) => r.entityType)?.entityType ||
+            (jsonld ? (0, structured_data_1.pageClassFromJsonLdTypes)(jsonld.types) : null) ||
+            'unknown';
+        const entityId = await this.upsertEntity(pageClass, domain, sourceUrl, merged, matched, sessionId);
+        const status = entityId ? 'ok' : matched.length || jsonld ? 'empty' : 'no-rule';
+        await this.writeParseLog({
+            sessionId, sourceUrl, domain, pageClass,
+            status,
+            hasSnapshot: true,
+            matchedRuleIds: matched.map((r) => r.ruleId),
+            extractedFields: Object.fromEntries(Object.entries(merged).filter(([, v]) => v != null)),
+            missingFields: Object.entries(merged).filter(([, v]) => v == null).map(([k]) => k),
+            entityId: entityId?.id ?? null,
+        });
+        return {
+            ...result,
+            status,
+            pageClass,
+            matchedRules: matched.map((r) => ({ ruleId: r.ruleId, version: r.version })),
+            jsonLdTypes: jsonld?.types || [],
+            extracted: merged,
+            entity: entityId,
+        };
+    }
+    async upsertEntity(pageClass, domain, sourceUrl, extracted, matched, sessionId, ruleVersionFallback = 'jsonld') {
+        const dedupKey = `${domain}:${normalizeUrlForDedup(sourceUrl)}`;
+        const ruleVersion = matched[0]?.version || ruleVersionFallback;
+        const raw = { extracted, ruleVersion };
+        if (pageClass === 'talent') {
+            const name = extracted.name || extracted['og:title'] || null;
+            if (!name)
+                return null;
+            const existing = await this.talentRepo.findOne({ where: { dedupKey } });
+            const entity = existing || this.talentRepo.create({ source: domain, sourceUrl, dedupKey });
+            entity.name = name;
+            entity.headline = extracted.headline || extracted.jobTitle || extracted.description || null;
+            entity.raw = raw;
+            entity.extracted = extracted;
+            entity.ruleVersion = ruleVersion;
+            entity.sessionId = sessionId;
+            const saved = await this.talentRepo.save(entity);
+            return { id: saved.id, type: 'talent', isNew: !existing };
+        }
+        if (pageClass === 'job') {
+            const title = extracted.title || extracted.jobTitle || extracted.name || null;
+            if (!title)
+                return null;
+            const existing = await this.jobRepo.findOne({ where: { dedupKey } });
+            const entity = existing || this.jobRepo.create({ source: domain, sourceUrl, dedupKey });
+            entity.title = title;
+            entity.company = extracted.company || extracted.hiringOrganization || extracted['hiringOrganization.name'] || null;
+            entity.raw = raw;
+            entity.extracted = extracted;
+            entity.ruleVersion = ruleVersion;
+            entity.sessionId = sessionId;
+            const saved = await this.jobRepo.save(entity);
+            return { id: saved.id, type: 'job', isNew: !existing };
+        }
+        if (pageClass === 'company') {
+            const name = extracted.name || extracted.legalName || extracted['og:title'] || null;
+            if (!name)
+                return null;
+            const existing = await this.companyRepo.findOne({ where: { dedupKey } });
+            const entity = existing || this.companyRepo.create({ source: domain, sourceUrl, dedupKey });
+            entity.name = name;
+            entity.raw = raw;
+            entity.extracted = extracted;
+            entity.ruleVersion = ruleVersion;
+            entity.sessionId = sessionId;
+            const saved = await this.companyRepo.save(entity);
+            return { id: saved.id, type: 'company', isNew: !existing };
+        }
+        return null;
+    }
+    async saveAdapterEntity(sessionId, sourceUrl, entity) {
+        const pageClass = String(entity.entityType || entity.type || 'unknown');
+        if (!['talent', 'job', 'company'].includes(pageClass))
+            return null;
+        const entityUrl = typeof entity.url === 'string' && entity.url ? entity.url : sourceUrl;
+        const domain = etldPlus1(new URL(entityUrl).hostname);
+        const extracted = { ...entity };
+        const saved = await this.upsertEntity(pageClass, domain, entityUrl, extracted, [], sessionId, 'adapter');
+        await this.writeParseLog({
+            sessionId,
+            sourceUrl: entityUrl,
+            domain,
+            pageClass,
+            status: saved ? 'ok' : 'empty',
+            hasSnapshot: true,
+            matchedRuleIds: [],
+            extractedFields: Object.fromEntries(Object.entries(extracted).filter(([, v]) => v != null)),
+            entityId: saved?.id ?? null,
+        });
+        return saved;
+    }
+    async writeParseLog(entry) {
+        await this.parseLogRepo.save(this.parseLogRepo.create({
+            sourceUrl: entry.sourceUrl || '',
+            domain: entry.domain || '',
+            pageClass: entry.pageClass || 'unknown',
+            matchedRuleIds: entry.matchedRuleIds || [],
+            extractedFields: entry.extractedFields || {},
+            missingFields: entry.missingFields || [],
+            hasSnapshot: entry.hasSnapshot ?? false,
+            entityId: entry.entityId ?? null,
+            ...entry,
+        }));
+    }
+    suggestRule(sessionId, events) {
+        const sourceUrl = events.find((e) => e.sourceUrl)?.sourceUrl || null;
+        const html = extractSnapshotHtml(events);
+        if (!sourceUrl || !html) {
+            return { sessionId, error: 'missing snapshot or sourceUrl' };
+        }
+        const parsed = new URL(sourceUrl);
+        const jsonld = (0, structured_data_1.extractJsonLd)(html);
+        const og = (0, structured_data_1.extractOpenGraph)(html);
+        const pageClass = jsonld ? (0, structured_data_1.pageClassFromJsonLdTypes)(jsonld.types) : null;
+        const draft = {
+            ruleId: `${etldPlus1(parsed.hostname).replace(/\./g, '-')}-${pageClass || 'page'}-auto`,
+            version: '0.1.0-draft',
+            targetHost: `^${parsed.hostname.replace(/\./g, '\\.')}$`,
+            targetPathPattern: `^${parsed.pathname.replace(/\//g, '\\/')}`,
+            entityType: pageClass,
+            priority: 50,
+        };
+        if (jsonld && pageClass) {
+            draft.ruleType = 'selector';
+            draft.rulePayload = { mapping: {} };
+            draft.note = 'JSON-LD 通道可完整解析,无需 selector 规则;入库字段来自 JSON-LD 扁平化结果';
+            draft.jsonLdSample = Object.fromEntries(Object.entries(jsonld.data).slice(0, 15));
+        }
+        else {
+            draft.ruleType = 'script';
+            draft.rulePayload = {
+                fields: {
+                    name: { selector: 'h1' },
+                    headline: { selector: 'h2' },
+                    description: { selector: 'meta[name="description"]', attr: 'content' },
+                },
+            };
+            draft.note = '无 JSON-LD;以下为通用猜测 selector,请基于快照 HTML 校对(见 htmlPreview)';
+        }
+        draft.ogSample = og;
+        draft.htmlPreview = html.slice(0, 4000);
+        return draft;
+    }
+};
+exports.ExtractionService = ExtractionService;
+exports.ExtractionService = ExtractionService = __decorate([
+    (0, common_1.Injectable)(),
+    __param(0, (0, typeorm_1.InjectRepository)(parse_log_entity_1.ParseLog)),
+    __param(1, (0, typeorm_1.InjectRepository)(talent_profile_entity_1.TalentProfile)),
+    __param(2, (0, typeorm_1.InjectRepository)(job_posting_entity_1.JobPosting)),
+    __param(3, (0, typeorm_1.InjectRepository)(company_profile_entity_1.CompanyProfile)),
+    __metadata("design:paramtypes", [typeorm_2.Repository,
+        typeorm_2.Repository,
+        typeorm_2.Repository,
+        typeorm_2.Repository,
+        extraction_rule_service_1.ExtractionRuleService])
+], ExtractionService);
+//# sourceMappingURL=extraction.service.js.map

package/dist/assets/server/extraction/extraction.service.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"extraction.service.js","sourceRoot":"","sources":["../../src/extraction/extraction.service.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;AAAA,2CAA4C;AAC5C,6CAAmD;AACnD,qCAAqC;AACrC,qCAA+B;AAC/B,yFAA8E;AAE9E,uEAAkE;AAClE,kEAAuD;AACvD,4EAAiE;AACjE,sEAA2D;AAC3D,8EAAmE;AACnE,uDAA8F;AAC9F,+CAAmE;AAenE,SAAS,UAAU,CAAC,IAAkB;IACpC,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,SAAS,mBAAmB,CAAC,MAAwB;IACnD,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;QAC3B,IAAI,KAAK,CAAC,SAAS,KAAK,OAAO;YAAE,SAAS;QAC1C,MAAM,OAAO,GAAG,KAAK,CAAC,OAAc,CAAC;QAErC,IAAI,OAAO,EAAE,IAAI,KAAK,CAAC,IAAI,OAAO,CAAC,IAAI,EAAE,IAAI,EAAE,CAAC;YAC9C,OAAO,UAAU,CAAC,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACvC,CAAC;IACH,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED,SAAS,SAAS,CAAC,IAAoB,EAAE,GAAW;IAClD,IAAI,CAAC;QACH,MAAM,WAAW,GAAG,IAAI,MAAM,CAAC,IAAI,CAAC,UAAU,EAAE,GAAG,CAAC,CAAC;QACrD,MAAM,WAAW,GAAG,IAAI,MAAM,CAAC,IAAI,CAAC,iBAAiB,EAAE,GAAG,CAAC,CAAC;QAC5D,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,GAAG,CAAC,CAAC;QAC5B,OAAO,WAAW,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,IAAI,WAAW,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC;IAChG,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,KAAK,CAAC;IACf,CAAC;AACH,CAAC;AAED,SAAS,iBAAiB,CAAC,OAAY,EAAE,IAAY;IACnD,MAAM,CAAC,GAAG,IAAA,cAAI,EAAC,IAAI,CAAC,CAAC;IACrB,MAAM,MAAM,GAAwB,EAAE,CAAC;IACvC,MAAM,OAAO,GAAG,OAAO,CAAC,OAAO,IAAI,EAAE,CAAC;IACtC,KAAK,MAAM,CAAC,KAAK,EAAE,QAAQ,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;QACxD,MAAM,EAAE,GAAG,CAAC,CAAC,QAAkB,CAAC,CAAC,KAAK,EAAE,CAAC;QACzC,MAAM,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,IAAI,EAAE,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,IAAI,CAAC;IACrF,CAAC;IACD,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,SAAS,cAAc,CAAC,OAAY,EAAE,IAAY;IAChD,MAAM,MAAM,GAAwB,EAAE,CAAC;IACvC,MAAM,OAAO,GAAG,OAAO,CAAC,OAAO,IAAI,EAAE,CAAC;IACtC,KAAK,MAAM,CAAC,KAAK,EAAE,OAAO,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;QACvD,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,MAAM,CAAC,OAAiB,EAAE,GAAG,CAAC,CAAC,CAAC;QAC7D,MAAM,CAAC,KAAK,CAAC,GAAG,KAAK,EAAE,CAAC,CAAC,CAAC,IAAI,IAAI,CAAC;IACrC,CAAC;IACD,OAAO,MAAM,CAAC;AAChB,CAAC;AAGD,SAAS,SAAS,CAAC,QAAgB;IACjC,MAAM,KAAK,GAAG,QAAQ,CAAC,WAAW,EAAE,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;IAChE,IAAI,KAAK,CAAC,MAAM,IAAI,CAAC;QAAE,OAAO,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IAE9C,MAAM,WAAW,GAAG,IAAI,GAAG,CAAC,CAAC,QAAQ,EAAE,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE,QAAQ,EAAE,QAAQ,CAAC,CAAC,CAAC;IACxF,MAAM,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IACxC,IAAI,WAAW,CAAC,GAAG,CAAC,KAAK,CAAC;QAAE,OAAO,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IAC7D,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,oBAAoB,CAAC,GAAW;IACvC,IAAI,CAAC;QACH,MAAM,CAAC,GAAG,IAAI,GAAG,CAAC,GAAG,CAAC,CAAC;QAEvB,OAAO,GAAG,CAAC,CAAC,QAAQ,CAAC,WAAW,EAAE,GAAG,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;IACjF,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,GAAG,CAAC;IACb,CAAC;AACH,CAAC;AAGM,IAAM,iBAAiB,GAAvB,MAAM,iBAAiB;IAGT;IAEA;IAEA;IAEA;IACA;IATnB,YAEmB,YAAkC,EAElC,UAAqC,EAErC,OAA+B,EAE/B,WAAuC,EACvC,WAAkC;QAPlC,iBAAY,GAAZ,YAAY,CAAsB;QAElC,eAAU,GAAV,UAAU,CAA2B;QAErC,YAAO,GAAP,OAAO,CAAwB;QAE/B,gBAAW,GAAX,WAAW,CAA4B;QACvC,gBAAW,GAAX,WAAW,CAAuB;IAClD,CAAC;IAGJ,UAAU,CAAC,MAAwB,EAAE,KAAuB;QAC1D,MAAM,SAAS,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC,EAAE,SAAS,IAAI,IAAI,CAAC;QACrE,MAAM,IAAI,GAAG,mBAAmB,CAAC,MAAM,CAAC,CAAC;QACzC,IAAI,CAAC,SAAS,IAAI,CAAC,IAAI,EAAE,CAAC;YACxB,OAAO,EAAE,SAAS,EAAE,WAAW,EAAE,CAAC,CAAC,IAAI,EAAE,SAAS,EAAE,IAAI,EAAE,OAAO,EAAE,2CAA2C,EAAE,CAAC;QACnH,CAAC;QACD,OAAO,IAAI,CAAC,gBAAgB,CAAC,SAAS,EAAE,IAAI,EAAE,KAAK,CAAC,CAAC;IACvD,CAAC;IAED,gBAAgB,CAAC,SAAiB,EAAE,IAAY,EAAE,KAAuB;QACvE,MAAM,OAAO,GAAG,KAAK;aAClB,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,IAAI,SAAS,CAAC,CAAC,EAAE,SAAS,CAAC,CAAC;aACnD,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,GAAG,CAAC,CAAC,QAAQ,CAAC,CAAC;QAE3C,MAAM,SAAS,GAAwB,EAAE,CAAC;QAC1C,KAAK,MAAM,IAAI,IAAI,OAAO,EAAE,CAAC;YAC3B,IAAI,IAAI,CAAC,QAAQ,KAAK,UAAU;gBAAE,MAAM,CAAC,MAAM,CAAC,SAAS,EAAE,iBAAiB,CAAC,IAAI,CAAC,WAAW,EAAE,IAAI,CAAC,CAAC,CAAC;iBACjG,IAAI,IAAI,CAAC,QAAQ,KAAK,OAAO;gBAAE,MAAM,CAAC,MAAM,CAAC,SAAS,EAAE,cAAc,CAAC,IAAI,CAAC,WAAW,EAAE,IAAI,CAAC,CAAC,CAAC;iBAChG,IAAI,IAAI,CAAC,QAAQ,KAAK,QAAQ;gBAAE,MAAM,CAAC,MAAM,CAAC,SAAS,EAAE,IAAA,6BAAe,EAAC,IAAI,EAAE,IAAI,CAAC,WAA2C,CAAC,CAAC,CAAC;QACzI,CAAC;QACD,OAAO;YACL,SAAS;YACT,YAAY,EAAE,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,MAAM,EAAE,CAAC,CAAC,MAAM,EAAE,OAAO,EAAE,CAAC,CAAC,OAAO,EAAE,CAAC,CAAC;YAC5E,SAAS;SACV,CAAC;IACJ,CAAC;IAOD,KAAK,CAAC,YAAY,CAAC,SAAiB;QAClC,MAAM,SAAS,GAAI,IAAI,CAAC,YAAY,CAAC,OAAe,CAAC,aAAa,CAAC,uCAAc,CAAC,CAAC;QACnF,MAAM,MAAM,GAAqB,MAAM,SAAS,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,EAAE,SAAS,EAAE,EAAE,KAAK,EAAE,EAAE,SAAS,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC;QAC7G,MAAM,SAAS,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC,EAAE,SAAS,IAAI,IAAI,CAAC;QACrE,MAAM,IAAI,GAAG,mBAAmB,CAAC,MAAM,CAAC,CAAC;QACzC,MAAM,MAAM,GAAG,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,IAAI,GAAG,CAAC,SAAS,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;QAEvE,MAAM,MAAM,GAAwB,EAAE,SAAS,EAAE,SAAS,EAAE,MAAM,EAAE,CAAC;QACrE,IAAI,CAAC,SAAS,IAAI,CAAC,IAAI,EAAE,CAAC;YACxB,MAAM,IAAI,CAAC,aAAa,CAAC,EAAE,SAAS,EAAE,SAAS,EAAE,SAAS,IAAI,EAAE,EAAE,MAAM,EAAE,SAAS,EAAE,SAAS,EAAE,MAAM,EAAE,aAAa,EAAE,WAAW,EAAE,KAAK,EAAE,CAAC,CAAC;YAC7I,OAAO,EAAE,GAAG,MAAM,EAAE,MAAM,EAAE,aAAa,EAAE,OAAO,EAAE,2CAA2C,EAAE,CAAC;QACpG,CAAC;QAED,MAAM,MAAM,GAAG,IAAA,+BAAa,EAAC,IAAI,CAAC,CAAC;QACnC,MAAM,EAAE,GAAG,IAAA,kCAAgB,EAAC,IAAI,CAAC,CAAC;QAClC,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,WAAW,CAAC,WAAW,EAAE,CAAC;QACnD,MAAM,OAAO,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,IAAI,SAAS,CAAC,CAAC,EAAE,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,GAAG,CAAC,CAAC,QAAQ,CAAC,CAAC;QAGlH,MAAM,aAAa,GAAwB,EAAE,CAAC;QAC9C,KAAK,MAAM,IAAI,IAAI,OAAO,EAAE,CAAC;YAC3B,IAAI,IAAI,CAAC,QAAQ,KAAK,UAAU;gBAAE,MAAM,CAAC,MAAM,CAAC,aAAa,EAAE,iBAAiB,CAAC,IAAI,CAAC,WAAW,EAAE,IAAI,CAAC,CAAC,CAAC;iBACrG,IAAI,IAAI,CAAC,QAAQ,KAAK,OAAO;gBAAE,MAAM,CAAC,MAAM,CAAC,aAAa,EAAE,cAAc,CAAC,IAAI,CAAC,WAAW,EAAE,IAAI,CAAC,CAAC,CAAC;iBACpG,IAAI,IAAI,CAAC,QAAQ,KAAK,QAAQ;gBAAE,MAAM,CAAC,MAAM,CAAC,aAAa,EAAE,IAAA,6BAAe,EAAC,IAAI,EAAE,IAAI,CAAC,WAA2C,CAAC,CAAC,CAAC;QAC7I,CAAC;QAGD,MAAM,MAAM,GAAwB,EAAE,GAAG,EAAE,EAAE,GAAG,CAAC,MAAM,EAAE,IAAI,IAAI,EAAE,CAAC,EAAE,GAAG,aAAa,EAAE,CAAC;QACzF,KAAK,MAAM,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,EAAE,GAAG,CAAC,MAAM,EAAE,IAAI,IAAI,EAAE,CAAC,EAAE,GAAG,CAAC,EAAE,IAAI,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC;YAChF,IAAI,aAAa,CAAC,CAAC,CAAC,IAAI,IAAI,IAAI,CAAC,IAAI,IAAI;gBAAE,MAAM,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC;QAC3D,CAAC;QAGD,MAAM,SAAS,GACb,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,UAAU,CAAC,EAAE,UAAU;YAC7C,CAAC,MAAM,CAAC,CAAC,CAAC,IAAA,0CAAwB,EAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;YACxD,SAAS,CAAC;QAGZ,MAAM,QAAQ,GAAG,MAAM,IAAI,CAAC,YAAY,CAAC,SAAS,EAAE,MAAM,EAAE,SAAS,EAAE,MAAM,EAAE,OAAO,EAAE,SAAS,CAAC,CAAC;QACnG,MAAM,MAAM,GAAG,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,OAAO,CAAC,MAAM,IAAI,MAAM,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,SAAS,CAAC;QAChF,MAAM,IAAI,CAAC,aAAa,CAAC;YACvB,SAAS,EAAE,SAAS,EAAE,MAAM,EAAE,SAAS;YACvC,MAAM;YACN,WAAW,EAAE,IAAI;YACjB,cAAc,EAAE,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,CAAC;YAC5C,eAAe,EAAE,MAAM,CAAC,WAAW,CAAC,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,IAAI,IAAI,CAAC,CAAC;YACxF,aAAa,EAAE,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,IAAI,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC;YAClF,QAAQ,EAAE,QAAQ,EAAE,EAAE,IAAI,IAAI;SAC/B,CAAC,CAAC;QAEH,OAAO;YACL,GAAG,MAAM;YACT,MAAM;YACN,SAAS;YACT,YAAY,EAAE,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,MAAM,EAAE,CAAC,CAAC,MAAM,EAAE,OAAO,EAAE,CAAC,CAAC,OAAO,EAAE,CAAC,CAAC;YAC5E,WAAW,EAAE,MAAM,EAAE,KAAK,IAAI,EAAE;YAChC,SAAS,EAAE,MAAM;YACjB,MAAM,EAAE,QAAQ;SACjB,CAAC;IACJ,CAAC;IAGO,KAAK,CAAC,YAAY,CACxB,SAAiB,EACjB,MAAc,EACd,SAAiB,EACjB,SAA8B,EAC9B,OAAyB,EACzB,SAAiB,EACjB,mBAAmB,GAAG,QAAQ;QAE9B,MAAM,QAAQ,GAAG,GAAG,MAAM,IAAI,oBAAoB,CAAC,SAAS,CAAC,EAAE,CAAC;QAChE,MAAM,WAAW,GAAG,OAAO,CAAC,CAAC,CAAC,EAAE,OAAO,IAAI,mBAAmB,CAAC;QAC/D,MAAM,GAAG,GAAG,EAAE,SAAS,EAAE,WAAW,EAAE,CAAC;QAEvC,IAAI,SAAS,KAAK,QAAQ,EAAE,CAAC;YAC3B,MAAM,IAAI,GAAG,SAAS,CAAC,IAAI,IAAI,SAAS,CAAC,UAAU,CAAC,IAAI,IAAI,CAAC;YAC7D,IAAI,CAAC,IAAI;gBAAE,OAAO,IAAI,CAAC;YACvB,MAAM,QAAQ,GAAG,MAAM,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC,EAAE,KAAK,EAAE,EAAE,QAAQ,EAAE,EAAE,CAAC,CAAC;YACxE,MAAM,MAAM,GAAG,QAAQ,IAAI,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC,EAAE,MAAM,EAAE,MAAM,EAAE,SAAS,EAAE,QAAQ,EAAE,CAAC,CAAC;YAC3F,MAAM,CAAC,IAAI,GAAG,IAAI,CAAC;YACnB,MAAM,CAAC,QAAQ,GAAG,SAAS,CAAC,QAAQ,IAAI,SAAS,CAAC,QAAQ,IAAI,SAAS,CAAC,WAAW,IAAI,IAAI,CAAC;YAC5F,MAAM,CAAC,GAAG,GAAG,GAAG,CAAC;YACjB,MAAM,CAAC,SAAS,GAAG,SAAS,CAAC;YAC7B,MAAM,CAAC,WAAW,GAAG,WAAW,CAAC;YACjC,MAAM,CAAC,SAAS,GAAG,SAAS,CAAC;YAC7B,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;YACjD,OAAO,EAAE,EAAE,EAAE,KAAK,CAAC,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,KAAK,EAAE,CAAC,QAAQ,EAAE,CAAC;QAC5D,CAAC;QACD,IAAI,SAAS,KAAK,KAAK,EAAE,CAAC;YACxB,MAAM,KAAK,GAAG,SAAS,CAAC,KAAK,IAAI,SAAS,CAAC,QAAQ,IAAI,SAAS,CAAC,IAAI,IAAI,IAAI,CAAC;YAC9E,IAAI,CAAC,KAAK;gBAAE,OAAO,IAAI,CAAC;YACxB,MAAM,QAAQ,GAAG,MAAM,IAAI,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,KAAK,EAAE,EAAE,QAAQ,EAAE,EAAE,CAAC,CAAC;YACrE,MAAM,MAAM,GAAG,QAAQ,IAAI,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,MAAM,EAAE,MAAM,EAAE,SAAS,EAAE,QAAQ,EAAE,CAAC,CAAC;YACxF,MAAM,CAAC,KAAK,GAAG,KAAK,CAAC;YACrB,MAAM,CAAC,OAAO,GAAG,SAAS,CAAC,OAAO,IAAI,SAAS,CAAC,kBAAkB,IAAI,SAAS,CAAC,yBAAyB,CAAC,IAAI,IAAI,CAAC;YACnH,MAAM,CAAC,GAAG,GAAG,GAAG,CAAC;YACjB,MAAM,CAAC,SAAS,GAAG,SAAS,CAAC;YAC7B,MAAM,CAAC,WAAW,GAAG,WAAW,CAAC;YACjC,MAAM,CAAC,SAAS,GAAG,SAAS,CAAC;YAC7B,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;YAC9C,OAAO,EAAE,EAAE,EAAE,KAAK,CAAC,EAAE,EAAE,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,CAAC,QAAQ,EAAE,CAAC;QACzD,CAAC;QACD,IAAI,SAAS,KAAK,SAAS,EAAE,CAAC;YAC5B,MAAM,IAAI,GAAG,SAAS,CAAC,IAAI,IAAI,SAAS,CAAC,SAAS,IAAI,SAAS,CAAC,UAAU,CAAC,IAAI,IAAI,CAAC;YACpF,IAAI,CAAC,IAAI;gBAAE,OAAO,IAAI,CAAC;YACvB,MAAM,QAAQ,GAAG,MAAM,IAAI,CAAC,WAAW,CAAC,OAAO,CAAC,EAAE,KAAK,EAAE,EAAE,QAAQ,EAAE,EAAE,CAAC,CAAC;YACzE,MAAM,MAAM,GAAG,QAAQ,IAAI,IAAI,CAAC,WAAW,CAAC,MAAM,CAAC,EAAE,MAAM,EAAE,MAAM,EAAE,SAAS,EAAE,QAAQ,EAAE,CAAC,CAAC;YAC5F,MAAM,CAAC,IAAI,GAAG,IAAI,CAAC;YACnB,MAAM,CAAC,GAAG,GAAG,GAAG,CAAC;YACjB,MAAM,CAAC,SAAS,GAAG,SAAS,CAAC;YAC7B,MAAM,CAAC,WAAW,GAAG,WAAW,CAAC;YACjC,MAAM,CAAC,SAAS,GAAG,SAAS,CAAC;YAC7B,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,WAAW,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;YAClD,OAAO,EAAE,EAAE,EAAE,KAAK,CAAC,EAAE,EAAE,IAAI,EAAE,SAAS,EAAE,KAAK,EAAE,CAAC,QAAQ,EAAE,CAAC;QAC7D,CAAC;QACD,OAAO,IAAI,CAAC;IACd,CAAC;IAMD,KAAK,CAAC,iBAAiB,CACrB,SAAiB,EACjB,SAAiB,EACjB,MAA2B;QAE3B,MAAM,SAAS,GAAG,MAAM,CAAC,MAAM,CAAC,UAAU,IAAI,MAAM,CAAC,IAAI,IAAI,SAAS,CAAC,CAAC;QACxE,IAAI,CAAC,CAAC,QAAQ,EAAE,KAAK,EAAE,SAAS,CAAC,CAAC,QAAQ,CAAC,SAAS,CAAC;YAAE,OAAO,IAAI,CAAC;QACnE,MAAM,SAAS,GACb,OAAO,MAAM,CAAC,GAAG,KAAK,QAAQ,IAAI,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,SAAS,CAAC;QACxE,MAAM,MAAM,GAAG,SAAS,CAAC,IAAI,GAAG,CAAC,SAAS,CAAC,CAAC,QAAQ,CAAC,CAAC;QACtD,MAAM,SAAS,GAAwB,EAAE,GAAG,MAAM,EAAE,CAAC;QACrD,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,YAAY,CACnC,SAAS,EACT,MAAM,EACN,SAAS,EACT,SAAS,EACT,EAAE,EACF,SAAS,EACT,SAAS,CACV,CAAC;QACF,MAAM,IAAI,CAAC,aAAa,CAAC;YACvB,SAAS;YACT,SAAS,EAAE,SAAS;YACpB,MAAM;YACN,SAAS;YACT,MAAM,EAAE,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,OAAO;YAC9B,WAAW,EAAE,IAAI;YACjB,cAAc,EAAE,EAAE;YAClB,eAAe,EAAE,MAAM,CAAC,WAAW,CACjC,MAAM,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,IAAI,IAAI,CAAC,CACvD;YACD,QAAQ,EAAE,KAAK,EAAE,EAAE,IAAI,IAAI;SAC5B,CAAC,CAAC;QACH,OAAO,KAAK,CAAC;IACf,CAAC;IAEO,KAAK,CAAC,aAAa,CAAC,KAA4E;QACtG,MAAM,IAAI,CAAC,YAAY,CAAC,IAAI,CAC1B,IAAI,CAAC,YAAY,CAAC,MAAM,CAAC;YACvB,SAAS,EAAE,KAAK,CAAC,SAAS,IAAI,EAAE;YAChC,MAAM,EAAE,KAAK,CAAC,MAAM,IAAI,EAAE;YAC1B,SAAS,EAAE,KAAK,CAAC,SAAS,IAAI,SAAS;YACvC,cAAc,EAAE,KAAK,CAAC,cAAc,IAAI,EAAE;YAC1C,eAAe,EAAE,KAAK,CAAC,eAAe,IAAI,EAAE;YAC5C,aAAa,EAAE,KAAK,CAAC,aAAa,IAAI,EAAE;YACxC,WAAW,EAAE,KAAK,CAAC,WAAW,IAAI,KAAK;YACvC,QAAQ,EAAE,KAAK,CAAC,QAAQ,IAAI,IAAI;YAChC,GAAG,KAAK;SACT,CAAC,CACH,CAAC;IACJ,CAAC;IAMD,WAAW,CAAC,SAAiB,EAAE,MAAwB;QACrD,MAAM,SAAS,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC,EAAE,SAAS,IAAI,IAAI,CAAC;QACrE,MAAM,IAAI,GAAG,mBAAmB,CAAC,MAAM,CAAC,CAAC;QACzC,IAAI,CAAC,SAAS,IAAI,CAAC,IAAI,EAAE,CAAC;YACxB,OAAO,EAAE,SAAS,EAAE,KAAK,EAAE,+BAA+B,EAAE,CAAC;QAC/D,CAAC;QACD,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,SAAS,CAAC,CAAC;QAClC,MAAM,MAAM,GAAG,IAAA,+BAAa,EAAC,IAAI,CAAC,CAAC;QACnC,MAAM,EAAE,GAAG,IAAA,kCAAgB,EAAC,IAAI,CAAC,CAAC;QAClC,MAAM,SAAS,GAAG,MAAM,CAAC,CAAC,CAAC,IAAA,0CAAwB,EAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;QAEzE,MAAM,KAAK,GAAwB;YACjC,MAAM,EAAE,GAAG,SAAS,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC,OAAO,CAAC,KAAK,EAAE,GAAG,CAAC,IAAI,SAAS,IAAI,MAAM,OAAO;YACvF,OAAO,EAAE,aAAa;YACtB,UAAU,EAAE,IAAI,MAAM,CAAC,QAAQ,CAAC,OAAO,CAAC,KAAK,EAAE,KAAK,CAAC,GAAG;YACxD,iBAAiB,EAAE,IAAI,MAAM,CAAC,QAAQ,CAAC,OAAO,CAAC,KAAK,EAAE,KAAK,CAAC,EAAE;YAC9D,UAAU,EAAE,SAAS;YACrB,QAAQ,EAAE,EAAE;SACb,CAAC;QAEF,IAAI,MAAM,IAAI,SAAS,EAAE,CAAC;YAExB,KAAK,CAAC,QAAQ,GAAG,UAAU,CAAC;YAC5B,KAAK,CAAC,WAAW,GAAG,EAAE,OAAO,EAAE,EAAE,EAAE,CAAC;YACpC,KAAK,CAAC,IAAI,GAAG,qDAAqD,CAAC;YACnE,KAAK,CAAC,YAAY,GAAG,MAAM,CAAC,WAAW,CAAC,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC;QACpF,CAAC;aAAM,CAAC;YAEN,KAAK,CAAC,QAAQ,GAAG,QAAQ,CAAC;YAC1B,KAAK,CAAC,WAAW,GAAG;gBAClB,MAAM,EAAE;oBACN,IAAI,EAAE,EAAE,QAAQ,EAAE,IAAI,EAAE;oBACxB,QAAQ,EAAE,EAAE,QAAQ,EAAE,IAAI,EAAE;oBAC5B,WAAW,EAAE,EAAE,QAAQ,EAAE,0BAA0B,EAAE,IAAI,EAAE,SAAS,EAAE;iBACvE;aACF,CAAC;YACF,KAAK,CAAC,IAAI,GAAG,yDAAyD,CAAC;QACzE,CAAC;QACD,KAAK,CAAC,QAAQ,GAAG,EAAE,CAAC;QACpB,KAAK,CAAC,WAAW,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC;QACxC,OAAO,KAAK,CAAC;IACf,CAAC;CACF,CAAA;AA5QY,8CAAiB;4BAAjB,iBAAiB;IAD7B,IAAA,mBAAU,GAAE;IAGR,WAAA,IAAA,0BAAgB,EAAC,2BAAQ,CAAC,CAAA;IAE1B,WAAA,IAAA,0BAAgB,EAAC,qCAAa,CAAC,CAAA;IAE/B,WAAA,IAAA,0BAAgB,EAAC,+BAAU,CAAC,CAAA;IAE5B,WAAA,IAAA,0BAAgB,EAAC,uCAAc,CAAC,CAAA;qCALF,oBAAU;QAEZ,oBAAU;QAEb,oBAAU;QAEN,oBAAU;QACV,+CAAqB;GAV1C,iBAAiB,CA4Q7B"}

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

@@ -0,0 +1,13 @@
+export interface ScriptFieldRule {
+    selector: string;
+    attr?: string;
+    regex?: string;
+}
+export interface ScriptRulePayload {
+    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 +1,45 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.applyScriptRule = applyScriptRule;
+const cheerio_1 = require("cheerio");
+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) {
+    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":";;AAuCA,0CAkBC;AAzDD,qCAA2C;AAwB3C,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;IACtE,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/structured-data.d.ts

@@ -0,0 +1,7 @@
+export interface JsonLdResult {
+    types: string[];
+    data: Record<string, any>;
+}
+export declare function pageClassFromJsonLdTypes(types: string[]): 'talent' | 'job' | 'company' | null;
+export declare function extractJsonLd(html: string): JsonLdResult | null;
+export declare function extractOpenGraph(html: string): Record<string, string>;