hunter-mate 0.1.2 → 0.1.4

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

@@ -0,0 +1,275 @@
+"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.AdapterService = void 0;
+const common_1 = require("@nestjs/common");
+const crypto_1 = require("crypto");
+const typeorm_1 = require("@nestjs/typeorm");
+const typeorm_2 = require("typeorm");
+const extension_control_service_1 = require("../extension/extension-control.service");
+const extraction_service_1 = require("./extraction.service");
+const adapter_rule_entity_1 = require("./entities/adapter-rule.entity");
+const structured_data_1 = require("./structured-data");
+const sleep = (ms) => new Promise((r) => setTimeout(r, ms));
+let AdapterService = class AdapterService {
+    ruleRepo;
+    ext;
+    extraction;
+    constructor(ruleRepo, ext, extraction) {
+        this.ruleRepo = ruleRepo;
+        this.ext = ext;
+        this.extraction = extraction;
+    }
+    async onModuleInit() {
+        await this.seedLinkedInRule();
+    }
+    async seedLinkedInRule() {
+        const exists = await this.ruleRepo.findOne({
+            where: { platform: 'linkedin', version: '0.2.0', hostPattern: 'linkedin\\.com' },
+        });
+        if (exists)
+            return;
+        await this.ruleRepo.save(this.ruleRepo.create({
+            platform: 'linkedin',
+            version: '0.2.0',
+            hostPattern: 'linkedin\\.com',
+            pathPattern: '/in/',
+            ruleType: 'script',
+            rulePayload: { engine: 'linkedin-profile', entityType: 'talent' },
+            priority: 1000,
+            enabled: true,
+        }));
+    }
+    async findRulesForUrl(url) {
+        let parsed;
+        try {
+            parsed = new URL(url);
+        }
+        catch {
+            return [];
+        }
+        const all = await this.ruleRepo.find({ where: { enabled: true } });
+        return all
+            .filter((r) => {
+            try {
+                const hostMatch = new RegExp(r.hostPattern, 'i').test(parsed.hostname);
+                const pathMatch = new RegExp(r.pathPattern, 'i').test(parsed.pathname + parsed.search);
+                return hostMatch && pathMatch;
+            }
+            catch {
+                return false;
+            }
+        })
+            .sort((a, b) => b.priority - a.priority);
+    }
+    async analyzeSnapshot(snapshot) {
+        const sourceUrl = snapshot.sourceUrl || '';
+        const html = snapshot.html;
+        const matched = sourceUrl ? await this.findRulesForUrl(sourceUrl) : [];
+        const ruleExtracted = {};
+        for (const rule of matched) {
+            if (rule.ruleType === 'selector') {
+                const { load } = await Promise.resolve().then(() => require('cheerio'));
+                const $ = load(html);
+                const mapping = rule.rulePayload.mapping || {};
+                for (const [field, selector] of Object.entries(mapping)) {
+                    const el = $(selector).first();
+                    ruleExtracted[field] = el.text().trim() || el.attr('content') || el.attr('value') || null;
+                }
+            }
+            else if (rule.ruleType === 'regex') {
+                const mapping = rule.rulePayload.mapping || {};
+                for (const [field, pattern] of Object.entries(mapping)) {
+                    const hit = html.match(new RegExp(pattern, 's'));
+                    ruleExtracted[field] = hit?.[1] || null;
+                }
+            }
+            else if (rule.ruleType === 'script') {
+                const { applyScriptRule } = await Promise.resolve().then(() => require('./script-rule'));
+                Object.assign(ruleExtracted, applyScriptRule(html, rule.rulePayload));
+            }
+            else if (rule.ruleType === 'jsonld') {
+            }
+        }
+        const jsonld = (0, structured_data_1.extractJsonLd)(html);
+        const og = (0, structured_data_1.extractOpenGraph)(html);
+        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.rulePayload.entityType)?.rulePayload.entityType ||
+            (jsonld ? (0, structured_data_1.pageClassFromJsonLdTypes)(jsonld.types) : null) ||
+            snapshot.pageClass ||
+            'unknown';
+        return {
+            pageClass,
+            extracted: merged,
+            matchedRules: matched,
+            adapterVersion: matched[0]?.version || (jsonld ? 'jsonld' : 'unknown'),
+        };
+    }
+    list() {
+        return {
+            adapters: [
+                { platform: 'linkedin', host: 'linkedin.com', command: 'hm.linkedin.extract', mode: 'cdp-extract' },
+                { platform: 'hiredchina', host: 'hiredchina.com', command: 'hm.hiredchina.extract', mode: 'cdp-extract' },
+            ],
+        };
+    }
+    async run(body) {
+        if (!body?.url)
+            throw new common_1.BadRequestException('url required');
+        let parsed;
+        try {
+            parsed = new URL(body.url);
+        }
+        catch {
+            throw new common_1.BadRequestException(`invalid url: ${body.url}`);
+        }
+        const rules = await this.findRulesForUrl(body.url);
+        const adapter = rules[0];
+        if (!adapter) {
+            const fallback = [
+                { platform: 'linkedin', hostSuffix: 'linkedin.com', command: 'hm.linkedin.extract' },
+                { platform: 'hiredchina', hostSuffix: 'hiredchina.com', command: 'hm.hiredchina.extract' },
+            ];
+            const hardcoded = fallback.find((a) => parsed.hostname === a.hostSuffix || parsed.hostname.endsWith(`.${a.hostSuffix}`));
+            if (!hardcoded) {
+                throw new common_1.BadRequestException(`unsupported_site: ${parsed.hostname}`);
+            }
+            return this.runCdpExtract(hardcoded, body);
+        }
+        if (adapter.rulePayload.strategy === 'cdp') {
+            const cdpAdapter = {
+                platform: adapter.platform,
+                hostSuffix: parsed.hostname,
+                command: adapter.rulePayload.command || `hm.${adapter.platform}.extract`,
+            };
+            return this.runCdpExtract(cdpAdapter, body);
+        }
+        throw new common_1.BadRequestException('adapter rule without cdp strategy requires recording snapshot');
+    }
+    async runCdpExtract(adapter, body) {
+        let tabId = body.tab != null && Number.isFinite(Number(body.tab)) ? Number(body.tab) : null;
+        let createdTab = false;
+        if (tabId != null && tabId > 0) {
+            await this.expectOk(this.ext.sendCommand({ command: 'cp.goto', params: { url: body.url, __tab: tabId } }), 'cp.goto');
+        }
+        else {
+            tabId = await this.findTabOnHost(new URL(body.url).hostname);
+            if (tabId != null) {
+                await this.expectOk(this.ext.sendCommand({ command: 'cp.goto', params: { url: body.url, __tab: tabId } }), 'cp.goto');
+            }
+            else {
+                const resp = await this.expectOk(this.ext.sendCommand({ command: 'cp.tab.new', params: { url: body.url, active: false } }), 'cp.tab.new');
+                tabId = Number(resp.result?.page);
+                createdTab = true;
+            }
+        }
+        if (!Number.isFinite(tabId) || !(tabId > 0)) {
+            throw new common_1.HttpException('failed to resolve target tab', 500);
+        }
+        let extract = null;
+        let lastError = null;
+        for (let i = 0; i < 3; i++) {
+            await sleep(i === 0 ? (body.settleMs ?? 1200) : 1500);
+            try {
+                const resp = await this.expectOk(this.ext.sendCommand({
+                    command: adapter.command,
+                    params: { __tab: tabId },
+                    timeoutMs: 30000,
+                }), adapter.command);
+                const result = resp.result ?? null;
+                if (result) {
+                    const hasEntities = Array.isArray(result.entities) && result.entities.length > 0;
+                    const hasLinks = Object.values(result.links ?? {}).some((v) => Array.isArray(v) && v.length > 0);
+                    if (hasEntities || hasLinks) {
+                        extract = result;
+                        break;
+                    }
+                    if (i === 2)
+                        extract = result;
+                }
+            }
+            catch (e) {
+                lastError = e.message;
+            }
+        }
+        if (!extract) {
+            throw new common_1.HttpException(`extract_failed: ${lastError ?? 'no result from extension'}`, 502);
+        }
+        const runId = `adapter-${(0, crypto_1.randomUUID)().slice(0, 8)}`;
+        const saved = [];
+        for (const entity of extract.entities ?? []) {
+            const rec = await this.extraction.saveAdapterEntity(runId, extract.url || body.url, entity);
+            if (rec)
+                saved.push(rec);
+        }
+        const maxLinks = body.maxLinks ?? 10;
+        const links = {};
+        for (const [k, v] of Object.entries(extract.links ?? {})) {
+            links[k] = Array.isArray(v) ? v.slice(0, maxLinks) : [];
+        }
+        return {
+            ok: true,
+            runId,
+            url: extract.url || body.url,
+            canonical: extract.canonical ?? null,
+            pageClass: extract.pageClass ?? 'unknown',
+            platform: adapter.platform,
+            tab: String(tabId),
+            createdTab,
+            savedCount: saved.length,
+            saved,
+            links,
+            title: extract.title ?? null,
+            extractedAt: extract.extractedAt ?? null,
+        };
+    }
+    async expectOk(p, label) {
+        const resp = (await p);
+        if (resp && resp.status === 'error') {
+            throw new Error(`${label}: ${resp.message ?? 'extension error'}`);
+        }
+        return resp;
+    }
+    async findTabOnHost(hostname) {
+        try {
+            const resp = (await this.expectOk(this.ext.sendCommand({ command: 'cp.tab.list', params: { action: 'list' } }), 'cp.tab.list'));
+            const tabs = resp.result ?? [];
+            const hit = tabs.find((t) => {
+                try {
+                    return new URL(t.url).hostname === hostname;
+                }
+                catch {
+                    return false;
+                }
+            });
+            return hit ? Number(hit.page) : null;
+        }
+        catch {
+            return null;
+        }
+    }
+};
+exports.AdapterService = AdapterService;
+exports.AdapterService = AdapterService = __decorate([
+    (0, common_1.Injectable)(),
+    __param(0, (0, typeorm_1.InjectRepository)(adapter_rule_entity_1.AdapterRule)),
+    __metadata("design:paramtypes", [typeorm_2.Repository,
+        extension_control_service_1.ExtensionControlService,
+        extraction_service_1.ExtractionService])
+], AdapterService);
+//# sourceMappingURL=adapter.service.js.map

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

@@ -0,0 +1 @@
+{"version":3,"file":"adapter.service.js","sourceRoot":"","sources":["../../src/extraction/adapter.service.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;AAAA,2CAA8F;AAC9F,mCAAoC;AACpC,6CAAmD;AACnD,qCAAqC;AAErC,sFAAiF;AACjF,6DAAyD;AACzD,wEAA6D;AAE7D,uDAA8F;AAS9F,MAAM,KAAK,GAAG,CAAC,EAAU,EAAE,EAAE,CAAC,IAAI,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,UAAU,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC;AAQ7D,IAAM,cAAc,GAApB,MAAM,cAAc;IAGN;IACA;IACA;IAJnB,YAEmB,QAAiC,EACjC,GAA4B,EAC5B,UAA6B;QAF7B,aAAQ,GAAR,QAAQ,CAAyB;QACjC,QAAG,GAAH,GAAG,CAAyB;QAC5B,eAAU,GAAV,UAAU,CAAmB;IAC7C,CAAC;IAEJ,KAAK,CAAC,YAAY;QAChB,MAAM,IAAI,CAAC,gBAAgB,EAAE,CAAC;IAChC,CAAC;IAEO,KAAK,CAAC,gBAAgB;QAC5B,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC;YACzC,KAAK,EAAE,EAAE,QAAQ,EAAE,UAAU,EAAE,OAAO,EAAE,OAAO,EAAE,WAAW,EAAE,gBAAgB,EAAE;SACjF,CAAC,CAAC;QACH,IAAI,MAAM;YAAE,OAAO;QACnB,MAAM,IAAI,CAAC,QAAQ,CAAC,IAAI,CACtB,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC;YACnB,QAAQ,EAAE,UAAU;YACpB,OAAO,EAAE,OAAO;YAChB,WAAW,EAAE,gBAAgB;YAC7B,WAAW,EAAE,MAAM;YACnB,QAAQ,EAAE,QAAQ;YAClB,WAAW,EAAE,EAAE,MAAM,EAAE,kBAAkB,EAAE,UAAU,EAAE,QAAQ,EAAE;YACjE,QAAQ,EAAE,IAAI;YACd,OAAO,EAAE,IAAI;SACd,CAAC,CACH,CAAC;IACJ,CAAC;IAGD,KAAK,CAAC,eAAe,CAAC,GAAW;QAC/B,IAAI,MAAW,CAAC;QAChB,IAAI,CAAC;YACH,MAAM,GAAG,IAAI,GAAG,CAAC,GAAG,CAAC,CAAC;QACxB,CAAC;QAAC,MAAM,CAAC;YACP,OAAO,EAAE,CAAC;QACZ,CAAC;QACD,MAAM,GAAG,GAAG,MAAM,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,EAAE,OAAO,EAAE,IAAI,EAAE,EAAE,CAAC,CAAC;QACnE,OAAO,GAAG;aACP,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE;YACZ,IAAI,CAAC;gBACH,MAAM,SAAS,GAAG,IAAI,MAAM,CAAC,CAAC,CAAC,WAAW,EAAE,GAAG,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;gBACvE,MAAM,SAAS,GAAG,IAAI,MAAM,CAAC,CAAC,CAAC,WAAW,EAAE,GAAG,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC;gBACvF,OAAO,SAAS,IAAI,SAAS,CAAC;YAChC,CAAC;YAAC,MAAM,CAAC;gBACP,OAAO,KAAK,CAAC;YACf,CAAC;QACH,CAAC,CAAC;aACD,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,GAAG,CAAC,CAAC,QAAQ,CAAC,CAAC;IAC7C,CAAC;IAGD,KAAK,CAAC,eAAe,CAAC,QAA2B;QAM/C,MAAM,SAAS,GAAG,QAAQ,CAAC,SAAS,IAAI,EAAE,CAAC;QAC3C,MAAM,IAAI,GAAG,QAAQ,CAAC,IAAI,CAAC;QAC3B,MAAM,OAAO,GAAG,SAAS,CAAC,CAAC,CAAC,MAAM,IAAI,CAAC,eAAe,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;QAEvE,MAAM,aAAa,GAAwB,EAAE,CAAC;QAC9C,KAAK,MAAM,IAAI,IAAI,OAAO,EAAE,CAAC;YAC3B,IAAI,IAAI,CAAC,QAAQ,KAAK,UAAU,EAAE,CAAC;gBACjC,MAAM,EAAE,IAAI,EAAE,GAAG,2CAAa,SAAS,EAAC,CAAC;gBACzC,MAAM,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC;gBACrB,MAAM,OAAO,GAAG,IAAI,CAAC,WAAW,CAAC,OAAO,IAAI,EAAE,CAAC;gBAC/C,KAAK,MAAM,CAAC,KAAK,EAAE,QAAQ,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;oBACxD,MAAM,EAAE,GAAG,CAAC,CAAC,QAAkB,CAAC,CAAC,KAAK,EAAE,CAAC;oBACzC,aAAa,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;gBAC5F,CAAC;YACH,CAAC;iBAAM,IAAI,IAAI,CAAC,QAAQ,KAAK,OAAO,EAAE,CAAC;gBACrC,MAAM,OAAO,GAAG,IAAI,CAAC,WAAW,CAAC,OAAO,IAAI,EAAE,CAAC;gBAC/C,KAAK,MAAM,CAAC,KAAK,EAAE,OAAO,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;oBACvD,MAAM,GAAG,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,MAAM,CAAC,OAAiB,EAAE,GAAG,CAAC,CAAC,CAAC;oBAC3D,aAAa,CAAC,KAAK,CAAC,GAAG,GAAG,EAAE,CAAC,CAAC,CAAC,IAAI,IAAI,CAAC;gBAC1C,CAAC;YACH,CAAC;iBAAM,IAAI,IAAI,CAAC,QAAQ,KAAK,QAAQ,EAAE,CAAC;gBACtC,MAAM,EAAE,eAAe,EAAE,GAAG,2CAAa,eAAe,EAAC,CAAC;gBAC1D,MAAM,CAAC,MAAM,CAAC,aAAa,EAAE,eAAe,CAAC,IAAI,EAAE,IAAI,CAAC,WAAgC,CAAC,CAAC,CAAC;YAC7F,CAAC;iBAAM,IAAI,IAAI,CAAC,QAAQ,KAAK,QAAQ,EAAE,CAAC;YAExC,CAAC;QACH,CAAC;QAED,MAAM,MAAM,GAAG,IAAA,+BAAa,EAAC,IAAI,CAAC,CAAC;QACnC,MAAM,EAAE,GAAG,IAAA,kCAAgB,EAAC,IAAI,CAAC,CAAC;QAGlC,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;QAED,MAAM,SAAS,GACb,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,WAAW,CAAC,UAAU,CAAC,EAAE,WAAW,CAAC,UAAU;YACrE,CAAC,MAAM,CAAC,CAAC,CAAC,IAAA,0CAAwB,EAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;YACxD,QAAQ,CAAC,SAAS;YAClB,SAAS,CAAC;QAEZ,OAAO;YACL,SAAS;YACT,SAAS,EAAE,MAAM;YACjB,YAAY,EAAE,OAAO;YACrB,cAAc,EAAE,OAAO,CAAC,CAAC,CAAC,EAAE,OAAO,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,SAAS,CAAC;SACvE,CAAC;IACJ,CAAC;IAED,IAAI;QACF,OAAO;YACL,QAAQ,EAAE;gBACR,EAAE,QAAQ,EAAE,UAAU,EAAE,IAAI,EAAE,cAAc,EAAE,OAAO,EAAE,qBAAqB,EAAE,IAAI,EAAE,aAAa,EAAE;gBACnG,EAAE,QAAQ,EAAE,YAAY,EAAE,IAAI,EAAE,gBAAgB,EAAE,OAAO,EAAE,uBAAuB,EAAE,IAAI,EAAE,aAAa,EAAE;aAC1G;SACF,CAAC;IACJ,CAAC;IAED,KAAK,CAAC,GAAG,CAAC,IAAkF;QAC1F,IAAI,CAAC,IAAI,EAAE,GAAG;YAAE,MAAM,IAAI,4BAAmB,CAAC,cAAc,CAAC,CAAC;QAC9D,IAAI,MAAW,CAAC;QAChB,IAAI,CAAC;YACH,MAAM,GAAG,IAAI,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;QAC7B,CAAC;QAAC,MAAM,CAAC;YACP,MAAM,IAAI,4BAAmB,CAAC,gBAAgB,IAAI,CAAC,GAAG,EAAE,CAAC,CAAC;QAC5D,CAAC;QAED,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,eAAe,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;QACnD,MAAM,OAAO,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;QACzB,IAAI,CAAC,OAAO,EAAE,CAAC;YAEb,MAAM,QAAQ,GAAiB;gBAC7B,EAAE,QAAQ,EAAE,UAAU,EAAE,UAAU,EAAE,cAAc,EAAE,OAAO,EAAE,qBAAqB,EAAE;gBACpF,EAAE,QAAQ,EAAE,YAAY,EAAE,UAAU,EAAE,gBAAgB,EAAE,OAAO,EAAE,uBAAuB,EAAE;aAC3F,CAAC;YACF,MAAM,SAAS,GAAG,QAAQ,CAAC,IAAI,CAC7B,CAAC,CAAC,EAAE,EAAE,CAAC,MAAM,CAAC,QAAQ,KAAK,CAAC,CAAC,UAAU,IAAI,MAAM,CAAC,QAAQ,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,UAAU,EAAE,CAAC,CACxF,CAAC;YACF,IAAI,CAAC,SAAS,EAAE,CAAC;gBACf,MAAM,IAAI,4BAAmB,CAAC,qBAAqB,MAAM,CAAC,QAAQ,EAAE,CAAC,CAAC;YACxE,CAAC;YACD,OAAO,IAAI,CAAC,aAAa,CAAC,SAAS,EAAE,IAAI,CAAC,CAAC;QAC7C,CAAC;QAGD,IAAI,OAAO,CAAC,WAAW,CAAC,QAAQ,KAAK,KAAK,EAAE,CAAC;YAC3C,MAAM,UAAU,GAAe;gBAC7B,QAAQ,EAAE,OAAO,CAAC,QAAQ;gBAC1B,UAAU,EAAE,MAAM,CAAC,QAAQ;gBAC3B,OAAO,EAAE,OAAO,CAAC,WAAW,CAAC,OAAO,IAAI,MAAM,OAAO,CAAC,QAAQ,UAAU;aACzE,CAAC;YACF,OAAO,IAAI,CAAC,aAAa,CAAC,UAAU,EAAE,IAAI,CAAC,CAAC;QAC9C,CAAC;QAGD,MAAM,IAAI,4BAAmB,CAAC,+DAA+D,CAAC,CAAC;IACjG,CAAC;IAEO,KAAK,CAAC,aAAa,CAAC,OAAmB,EAAE,IAAkF;QAEjI,IAAI,KAAK,GAAkB,IAAI,CAAC,GAAG,IAAI,IAAI,IAAI,MAAM,CAAC,QAAQ,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;QAC3G,IAAI,UAAU,GAAG,KAAK,CAAC;QACvB,IAAI,KAAK,IAAI,IAAI,IAAI,KAAK,GAAG,CAAC,EAAE,CAAC;YAC/B,MAAM,IAAI,CAAC,QAAQ,CACjB,IAAI,CAAC,GAAG,CAAC,WAAW,CAAC,EAAE,OAAO,EAAE,SAAS,EAAE,MAAM,EAAE,EAAE,GAAG,EAAE,IAAI,CAAC,GAAG,EAAE,KAAK,EAAE,KAAK,EAAE,EAAE,CAAC,EACrF,SAAS,CACV,CAAC;QACJ,CAAC;aAAM,CAAC;YACN,KAAK,GAAG,MAAM,IAAI,CAAC,aAAa,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,QAAQ,CAAC,CAAC;YAC7D,IAAI,KAAK,IAAI,IAAI,EAAE,CAAC;gBAClB,MAAM,IAAI,CAAC,QAAQ,CACjB,IAAI,CAAC,GAAG,CAAC,WAAW,CAAC,EAAE,OAAO,EAAE,SAAS,EAAE,MAAM,EAAE,EAAE,GAAG,EAAE,IAAI,CAAC,GAAG,EAAE,KAAK,EAAE,KAAK,EAAE,EAAE,CAAC,EACrF,SAAS,CACV,CAAC;YACJ,CAAC;iBAAM,CAAC;gBACN,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,QAAQ,CAC9B,IAAI,CAAC,GAAG,CAAC,WAAW,CAAC,EAAE,OAAO,EAAE,YAAY,EAAE,MAAM,EAAE,EAAE,GAAG,EAAE,IAAI,CAAC,GAAG,EAAE,MAAM,EAAE,KAAK,EAAE,EAAE,CAAC,EACzF,YAAY,CACb,CAAC;gBACF,KAAK,GAAG,MAAM,CAAE,IAAgD,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC;gBAC/E,UAAU,GAAG,IAAI,CAAC;YACpB,CAAC;QACH,CAAC;QACD,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,EAAE,CAAC;YAC5C,MAAM,IAAI,sBAAa,CAAC,8BAA8B,EAAE,GAAG,CAAC,CAAC;QAC/D,CAAC;QAGD,IAAI,OAAO,GAA+B,IAAI,CAAC;QAC/C,IAAI,SAAS,GAAkB,IAAI,CAAC;QACpC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;YAC3B,MAAM,KAAK,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,QAAQ,IAAI,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;YACtD,IAAI,CAAC;gBACH,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,QAAQ,CAC9B,IAAI,CAAC,GAAG,CAAC,WAAW,CAAC;oBACnB,OAAO,EAAE,OAAO,CAAC,OAAO;oBACxB,MAAM,EAAE,EAAE,KAAK,EAAE,KAAK,EAAE;oBACxB,SAAS,EAAE,KAAK;iBACjB,CAAC,EACF,OAAO,CAAC,OAAO,CAChB,CAAC;gBACF,MAAM,MAAM,GAAI,IAAyC,CAAC,MAAM,IAAI,IAAI,CAAC;gBACzE,IAAI,MAAM,EAAE,CAAC;oBACX,MAAM,WAAW,GAAG,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,QAAQ,CAAC,IAAI,MAAM,CAAC,QAAQ,CAAC,MAAM,GAAG,CAAC,CAAC;oBACjF,MAAM,QAAQ,GAAG,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,KAAK,IAAI,EAAE,CAAC,CAAC,IAAI,CACrD,CAAC,CAAC,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,MAAM,GAAG,CAAC,CACxC,CAAC;oBACF,IAAI,WAAW,IAAI,QAAQ,EAAE,CAAC;wBAC5B,OAAO,GAAG,MAAM,CAAC;wBACjB,MAAM;oBACR,CAAC;oBACD,IAAI,CAAC,KAAK,CAAC;wBAAE,OAAO,GAAG,MAAM,CAAC;gBAChC,CAAC;YACH,CAAC;YAAC,OAAO,CAAC,EAAE,CAAC;gBACX,SAAS,GAAI,CAAW,CAAC,OAAO,CAAC;YACnC,CAAC;QACH,CAAC;QACD,IAAI,CAAC,OAAO,EAAE,CAAC;YACb,MAAM,IAAI,sBAAa,CAAC,mBAAmB,SAAS,IAAI,0BAA0B,EAAE,EAAE,GAAG,CAAC,CAAC;QAC7F,CAAC;QAGD,MAAM,KAAK,GAAG,WAAW,IAAA,mBAAU,GAAE,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,CAAC;QACpD,MAAM,KAAK,GAAwC,EAAE,CAAC;QACtD,KAAK,MAAM,MAAM,IAAI,OAAO,CAAC,QAAQ,IAAI,EAAE,EAAE,CAAC;YAC5C,MAAM,GAAG,GAAG,MAAM,IAAI,CAAC,UAAU,CAAC,iBAAiB,CAAC,KAAK,EAAE,OAAO,CAAC,GAAG,IAAI,IAAI,CAAC,GAAG,EAAE,MAAM,CAAC,CAAC;YAC5F,IAAI,GAAG;gBAAE,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;QAC3B,CAAC;QACD,MAAM,QAAQ,GAAG,IAAI,CAAC,QAAQ,IAAI,EAAE,CAAC;QACrC,MAAM,KAAK,GAA8B,EAAE,CAAC;QAC5C,KAAK,MAAM,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,KAAK,IAAI,EAAE,CAAC,EAAE,CAAC;YACzD,KAAK,CAAC,CAAC,CAAC,GAAG,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,QAAQ,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;QAC1D,CAAC;QAED,OAAO;YACL,EAAE,EAAE,IAAI;YACR,KAAK;YACL,GAAG,EAAE,OAAO,CAAC,GAAG,IAAI,IAAI,CAAC,GAAG;YAC5B,SAAS,EAAE,OAAO,CAAC,SAAS,IAAI,IAAI;YACpC,SAAS,EAAE,OAAO,CAAC,SAAS,IAAI,SAAS;YACzC,QAAQ,EAAE,OAAO,CAAC,QAAQ;YAC1B,GAAG,EAAE,MAAM,CAAC,KAAK,CAAC;YAClB,UAAU;YACV,UAAU,EAAE,KAAK,CAAC,MAAM;YACxB,KAAK;YACL,KAAK;YACL,KAAK,EAAE,OAAO,CAAC,KAAK,IAAI,IAAI;YAC5B,WAAW,EAAE,OAAO,CAAC,WAAW,IAAI,IAAI;SACzC,CAAC;IACJ,CAAC;IAEO,KAAK,CAAC,QAAQ,CAAC,CAAmB,EAAE,KAAa;QACvD,MAAM,IAAI,GAAG,CAAC,MAAM,CAAC,CAA0C,CAAC;QAChE,IAAI,IAAI,IAAI,IAAI,CAAC,MAAM,KAAK,OAAO,EAAE,CAAC;YACpC,MAAM,IAAI,KAAK,CAAC,GAAG,KAAK,KAAK,IAAI,CAAC,OAAO,IAAI,iBAAiB,EAAE,CAAC,CAAC;QACpE,CAAC;QACD,OAAO,IAAI,CAAC;IACd,CAAC;IAGO,KAAK,CAAC,aAAa,CAAC,QAAgB;QAC1C,IAAI,CAAC;YACH,MAAM,IAAI,GAAG,CAAC,MAAM,IAAI,CAAC,QAAQ,CAC/B,IAAI,CAAC,GAAG,CAAC,WAAW,CAAC,EAAE,OAAO,EAAE,aAAa,EAAE,MAAM,EAAE,EAAE,MAAM,EAAE,MAAM,EAAE,EAAE,CAAC,EAC5E,aAAa,CACd,CAAsD,CAAC;YACxD,MAAM,IAAI,GAAG,IAAI,CAAC,MAAM,IAAI,EAAE,CAAC;YAC/B,MAAM,GAAG,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE;gBAC1B,IAAI,CAAC;oBACH,OAAO,IAAI,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,QAAQ,KAAK,QAAQ,CAAC;gBAC9C,CAAC;gBAAC,MAAM,CAAC;oBACP,OAAO,KAAK,CAAC;gBACf,CAAC;YACH,CAAC,CAAC,CAAC;YACH,OAAO,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;QACvC,CAAC;QAAC,MAAM,CAAC;YACP,OAAO,IAAI,CAAC;QACd,CAAC;IACH,CAAC;CACF,CAAA;AAzRY,wCAAc;yBAAd,cAAc;IAD1B,IAAA,mBAAU,GAAE;IAGR,WAAA,IAAA,0BAAgB,EAAC,iCAAW,CAAC,CAAA;qCACH,oBAAU;QACf,mDAAuB;QAChB,sCAAiB;GALrC,cAAc,CAyR1B"}

package/dist/assets/server/extraction/anomaly-report.service.d.ts

@@ -0,0 +1,12 @@
+import { Repository } from 'typeorm';
+import { RecordingSnapshot } from '../recording/entities/recording-snapshot.entity';
+import { RecordingEvent } from '../recording/entities/recording-event.entity';
+import { DevopsSinkInit } from '../devops/devops-sink-init';
+export declare class AnomalyReportService {
+    private readonly snapshotRepo;
+    private readonly eventRepo;
+    private readonly devops;
+    constructor(snapshotRepo: Repository<RecordingSnapshot>, eventRepo: Repository<RecordingEvent>, devops: DevopsSinkInit);
+    report(sessionId: string, snapshotId: number, parseLogId: number, reason: string): Promise<void>;
+    private sanitizeHtml;
+}

package/dist/assets/server/extraction/anomaly-report.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.AnomalyReportService = void 0;
+const common_1 = require("@nestjs/common");
+const typeorm_1 = require("@nestjs/typeorm");
+const typeorm_2 = require("typeorm");
+const recording_snapshot_entity_1 = require("../recording/entities/recording-snapshot.entity");
+const recording_event_entity_1 = require("../recording/entities/recording-event.entity");
+const devops_sink_init_1 = require("../devops/devops-sink-init");
+let AnomalyReportService = class AnomalyReportService {
+    snapshotRepo;
+    eventRepo;
+    devops;
+    constructor(snapshotRepo, eventRepo, devops) {
+        this.snapshotRepo = snapshotRepo;
+        this.eventRepo = eventRepo;
+        this.devops = devops;
+    }
+    async report(sessionId, snapshotId, parseLogId, reason) {
+        if (process.env.HUNTER_MATE_DEVOPS_ENABLED !== '1')
+            return;
+        if (!this.devops.usage)
+            return;
+        const snapshot = await this.snapshotRepo.findOne({ where: { id: snapshotId } });
+        if (!snapshot)
+            return;
+        const recentEvents = await this.eventRepo.find({
+            where: { sessionId },
+            order: { createdAt: 'DESC' },
+            take: 50,
+        });
+        const sanitizedHtml = this.sanitizeHtml(snapshot.html);
+        this.devops.usage.capture({
+            feature: 'parse-anomaly',
+            event: 'report',
+            context: {
+                sessionId,
+                snapshotId,
+                parseLogId,
+                sourceUrl: snapshot.sourceUrl,
+                domain: snapshot.domain,
+                pageClass: snapshot.pageClass,
+                htmlLength: snapshot.html.length,
+                htmlSample: sanitizedHtml.slice(0, 200_000),
+                eventCount: recentEvents.length,
+                eventSample: recentEvents.slice(0, 10).map((e) => ({ type: e.eventType, payloadType: e.payload?.type })),
+                reason,
+            },
+        });
+        await this.snapshotRepo.update(snapshotId, { reportedAt: new Date() });
+    }
+    sanitizeHtml(html) {
+        return html
+            .replace(/<script\b[^>]*>[\s\S]*?<\/script\b[^>]*>/gi, '')
+            .replace(/<input\b[^>]*>/gi, '<input>')
+            .replace(/<textarea\b[^>]*>[\s\S]*?<\/textarea\b[^>]*>/gi, '<textarea></textarea>')
+            .slice(0, 250_000);
+    }
+};
+exports.AnomalyReportService = AnomalyReportService;
+exports.AnomalyReportService = AnomalyReportService = __decorate([
+    (0, common_1.Injectable)(),
+    __param(0, (0, typeorm_1.InjectRepository)(recording_snapshot_entity_1.RecordingSnapshot)),
+    __param(1, (0, typeorm_1.InjectRepository)(recording_event_entity_1.RecordingEvent)),
+    __metadata("design:paramtypes", [typeorm_2.Repository,
+        typeorm_2.Repository,
+        devops_sink_init_1.DevopsSinkInit])
+], AnomalyReportService);
+//# sourceMappingURL=anomaly-report.service.js.map

package/dist/assets/server/extraction/anomaly-report.service.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"anomaly-report.service.js","sourceRoot":"","sources":["../../src/extraction/anomaly-report.service.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;AAAA,2CAA4C;AAC5C,6CAAmD;AACnD,qCAAqC;AACrC,+FAAoF;AACpF,yFAA8E;AAE9E,iEAA4D;AAQrD,IAAM,oBAAoB,GAA1B,MAAM,oBAAoB;IAGZ;IAEA;IACA;IALnB,YAEmB,YAA2C,EAE3C,SAAqC,EACrC,MAAsB;QAHtB,iBAAY,GAAZ,YAAY,CAA+B;QAE3C,cAAS,GAAT,SAAS,CAA4B;QACrC,WAAM,GAAN,MAAM,CAAgB;IACtC,CAAC;IAEJ,KAAK,CAAC,MAAM,CAAC,SAAiB,EAAE,UAAkB,EAAE,UAAkB,EAAE,MAAc;QACpF,IAAI,OAAO,CAAC,GAAG,CAAC,0BAA0B,KAAK,GAAG;YAAE,OAAO;QAC3D,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK;YAAE,OAAO;QAE/B,MAAM,QAAQ,GAAG,MAAM,IAAI,CAAC,YAAY,CAAC,OAAO,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,EAAE,UAAU,EAAE,EAAE,CAAC,CAAC;QAChF,IAAI,CAAC,QAAQ;YAAE,OAAO;QAEtB,MAAM,YAAY,GAAG,MAAM,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC;YAC7C,KAAK,EAAE,EAAE,SAAS,EAAE;YACpB,KAAK,EAAE,EAAE,SAAS,EAAE,MAAM,EAAE;YAC5B,IAAI,EAAE,EAAE;SACT,CAAC,CAAC;QAEH,MAAM,aAAa,GAAG,IAAI,CAAC,YAAY,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;QAEvD,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,OAAO,CAAC;YACxB,OAAO,EAAE,eAAe;YACxB,KAAK,EAAE,QAAQ;YACf,OAAO,EAAE;gBACP,SAAS;gBACT,UAAU;gBACV,UAAU;gBACV,SAAS,EAAE,QAAQ,CAAC,SAAS;gBAC7B,MAAM,EAAE,QAAQ,CAAC,MAAM;gBACvB,SAAS,EAAE,QAAQ,CAAC,SAAS;gBAC7B,UAAU,EAAE,QAAQ,CAAC,IAAI,CAAC,MAAM;gBAChC,UAAU,EAAE,aAAa,CAAC,KAAK,CAAC,CAAC,EAAE,OAAO,CAAC;gBAC3C,UAAU,EAAE,YAAY,CAAC,MAAM;gBAC/B,WAAW,EAAE,YAAY,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC,CAAC,SAAS,EAAE,WAAW,EAAE,CAAC,CAAC,OAAO,EAAE,IAAI,EAAE,CAAC,CAAC;gBACxG,MAAM;aACP;SACF,CAAC,CAAC;QAEH,MAAM,IAAI,CAAC,YAAY,CAAC,MAAM,CAAC,UAAU,EAAE,EAAE,UAAU,EAAE,IAAI,IAAI,EAAE,EAAE,CAAC,CAAC;IACzE,CAAC;IAEO,YAAY,CAAC,IAAY;QAC/B,OAAO,IAAI;aACR,OAAO,CAAC,4CAA4C,EAAE,EAAE,CAAC;aACzD,OAAO,CAAC,kBAAkB,EAAE,SAAS,CAAC;aACtC,OAAO,CAAC,gDAAgD,EAAE,uBAAuB,CAAC;aAClF,KAAK,CAAC,CAAC,EAAE,OAAO,CAAC,CAAC;IACvB,CAAC;CACF,CAAA;AApDY,oDAAoB;+BAApB,oBAAoB;IADhC,IAAA,mBAAU,GAAE;IAGR,WAAA,IAAA,0BAAgB,EAAC,6CAAiB,CAAC,CAAA;IAEnC,WAAA,IAAA,0BAAgB,EAAC,uCAAc,CAAC,CAAA;qCADF,oBAAU;QAEb,oBAAU;QACb,iCAAc;GAN9B,oBAAoB,CAoDhC"}

package/dist/assets/server/extraction/entities/adapter-rule.entity.d.ts

@@ -0,0 +1,13 @@
+export type RuleType = 'selector' | 'regex' | 'script' | 'jsonld';
+export declare class AdapterRule {
+    id: number;
+    platform: string;
+    version: string;
+    hostPattern: string;
+    pathPattern: string;
+    ruleType: RuleType;
+    rulePayload: Record<string, any>;
+    priority: number;
+    enabled: boolean;
+    createdAt: Date;
+}

package/dist/assets/server/extraction/entities/adapter-rule.entity.js

@@ -0,0 +1,71 @@
+"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);
+};
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.AdapterRule = void 0;
+const typeorm_1 = require("typeorm");
+let AdapterRule = class AdapterRule {
+    id;
+    platform;
+    version;
+    hostPattern;
+    pathPattern;
+    ruleType;
+    rulePayload;
+    priority;
+    enabled;
+    createdAt;
+};
+exports.AdapterRule = AdapterRule;
+__decorate([
+    (0, typeorm_1.PrimaryGeneratedColumn)(),
+    __metadata("design:type", Number)
+], AdapterRule.prototype, "id", void 0);
+__decorate([
+    (0, typeorm_1.Column)({ type: 'varchar', length: 64 }),
+    __metadata("design:type", String)
+], AdapterRule.prototype, "platform", void 0);
+__decorate([
+    (0, typeorm_1.Column)({ type: 'varchar', length: 32 }),
+    __metadata("design:type", String)
+], AdapterRule.prototype, "version", void 0);
+__decorate([
+    (0, typeorm_1.Column)({ type: 'varchar', length: 256 }),
+    __metadata("design:type", String)
+], AdapterRule.prototype, "hostPattern", void 0);
+__decorate([
+    (0, typeorm_1.Column)({ type: 'varchar', length: 256 }),
+    __metadata("design:type", String)
+], AdapterRule.prototype, "pathPattern", void 0);
+__decorate([
+    (0, typeorm_1.Column)({ type: 'simple-enum', enum: ['selector', 'regex', 'script', 'jsonld'] }),
+    __metadata("design:type", String)
+], AdapterRule.prototype, "ruleType", void 0);
+__decorate([
+    (0, typeorm_1.Column)({ type: 'simple-json' }),
+    __metadata("design:type", Object)
+], AdapterRule.prototype, "rulePayload", void 0);
+__decorate([
+    (0, typeorm_1.Column)({ type: 'int', default: 0 }),
+    __metadata("design:type", Number)
+], AdapterRule.prototype, "priority", void 0);
+__decorate([
+    (0, typeorm_1.Column)({ type: 'boolean', default: true }),
+    __metadata("design:type", Boolean)
+], AdapterRule.prototype, "enabled", void 0);
+__decorate([
+    (0, typeorm_1.CreateDateColumn)(),
+    __metadata("design:type", Date)
+], AdapterRule.prototype, "createdAt", void 0);
+exports.AdapterRule = AdapterRule = __decorate([
+    (0, typeorm_1.Entity)('adapter_rules'),
+    (0, typeorm_1.Index)(['platform', 'version', 'hostPattern'])
+], AdapterRule);
+//# sourceMappingURL=adapter-rule.entity.js.map

package/dist/assets/server/extraction/entities/adapter-rule.entity.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"adapter-rule.entity.js","sourceRoot":"","sources":["../../../src/extraction/entities/adapter-rule.entity.ts"],"names":[],"mappings":";;;;;;;;;;;;AAAA,qCAA0F;AAUnF,IAAM,WAAW,GAAjB,MAAM,WAAW;IAEtB,EAAE,CAAS;IAGX,QAAQ,CAAS;IAGjB,OAAO,CAAS;IAGhB,WAAW,CAAS;IAGpB,WAAW,CAAS;IAGpB,QAAQ,CAAW;IAGnB,WAAW,CAAsB;IAGjC,QAAQ,CAAS;IAGjB,OAAO,CAAU;IAGjB,SAAS,CAAO;CACjB,CAAA;AA9BY,kCAAW;AAEtB;IADC,IAAA,gCAAsB,GAAE;;uCACd;AAGX;IADC,IAAA,gBAAM,EAAC,EAAE,IAAI,EAAE,SAAS,EAAE,MAAM,EAAE,EAAE,EAAE,CAAC;;6CACvB;AAGjB;IADC,IAAA,gBAAM,EAAC,EAAE,IAAI,EAAE,SAAS,EAAE,MAAM,EAAE,EAAE,EAAE,CAAC;;4CACxB;AAGhB;IADC,IAAA,gBAAM,EAAC,EAAE,IAAI,EAAE,SAAS,EAAE,MAAM,EAAE,GAAG,EAAE,CAAC;;gDACrB;AAGpB;IADC,IAAA,gBAAM,EAAC,EAAE,IAAI,EAAE,SAAS,EAAE,MAAM,EAAE,GAAG,EAAE,CAAC;;gDACrB;AAGpB;IADC,IAAA,gBAAM,EAAC,EAAE,IAAI,EAAE,aAAa,EAAE,IAAI,EAAE,CAAC,UAAU,EAAE,OAAO,EAAE,QAAQ,EAAE,QAAQ,CAAC,EAAE,CAAC;;6CAC9D;AAGnB;IADC,IAAA,gBAAM,EAAC,EAAE,IAAI,EAAE,aAAa,EAAE,CAAC;;gDACC;AAGjC;IADC,IAAA,gBAAM,EAAC,EAAE,IAAI,EAAE,KAAK,EAAE,OAAO,EAAE,CAAC,EAAE,CAAC;;6CACnB;AAGjB;IADC,IAAA,gBAAM,EAAC,EAAE,IAAI,EAAE,SAAS,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC;;4CAC1B;AAGjB;IADC,IAAA,0BAAgB,GAAE;8BACR,IAAI;8CAAC;sBA7BL,WAAW;IAFvB,IAAA,gBAAM,EAAC,eAAe,CAAC;IACvB,IAAA,eAAK,EAAC,CAAC,UAAU,EAAE,SAAS,EAAE,aAAa,CAAC,CAAC;GACjC,WAAW,CA8BvB"}

package/dist/assets/server/extraction/entities/company-profile.entity.d.ts

@@ -0,0 +1,14 @@
+export declare class CompanyProfile {
+    id: number;
+    source: string;
+    sourceUrl: string;
+    dedupKey: string;
+    name: string;
+    raw: Record<string, any>;
+    extracted: Record<string, any>;
+    ruleVersion: string;
+    sessionId: string;
+    tags: string[];
+    createdAt: Date;
+    updatedAt: Date;
+}

package/dist/assets/server/extraction/entities/company-profile.entity.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);
+};
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.CompanyProfile = void 0;
+const typeorm_1 = require("typeorm");
+let CompanyProfile = class CompanyProfile {
+    id;
+    source;
+    sourceUrl;
+    dedupKey;
+    name;
+    raw;
+    extracted;
+    ruleVersion;
+    sessionId;
+    tags;
+    createdAt;
+    updatedAt;
+};
+exports.CompanyProfile = CompanyProfile;
+__decorate([
+    (0, typeorm_1.PrimaryGeneratedColumn)(),
+    __metadata("design:type", Number)
+], CompanyProfile.prototype, "id", void 0);
+__decorate([
+    (0, typeorm_1.Column)({ type: 'varchar', length: 64 }),
+    __metadata("design:type", String)
+], CompanyProfile.prototype, "source", void 0);
+__decorate([
+    (0, typeorm_1.Column)({ type: 'varchar', length: 512 }),
+    __metadata("design:type", String)
+], CompanyProfile.prototype, "sourceUrl", void 0);
+__decorate([
+    (0, typeorm_1.Column)({ type: 'varchar', length: 256 }),
+    __metadata("design:type", String)
+], CompanyProfile.prototype, "dedupKey", void 0);
+__decorate([
+    (0, typeorm_1.Column)({ type: 'varchar', length: 256, nullable: true }),
+    __metadata("design:type", String)
+], CompanyProfile.prototype, "name", void 0);
+__decorate([
+    (0, typeorm_1.Column)({ type: 'simple-json', nullable: true }),
+    __metadata("design:type", Object)
+], CompanyProfile.prototype, "raw", void 0);
+__decorate([
+    (0, typeorm_1.Column)({ type: 'simple-json', nullable: true }),
+    __metadata("design:type", Object)
+], CompanyProfile.prototype, "extracted", void 0);
+__decorate([
+    (0, typeorm_1.Column)({ type: 'varchar', length: 32, nullable: true }),
+    __metadata("design:type", String)
+], CompanyProfile.prototype, "ruleVersion", void 0);
+__decorate([
+    (0, typeorm_1.Column)({ type: 'varchar', length: 64, nullable: true }),
+    __metadata("design:type", String)
+], CompanyProfile.prototype, "sessionId", void 0);
+__decorate([
+    (0, typeorm_1.Column)({ type: 'simple-json', nullable: true }),
+    __metadata("design:type", Array)
+], CompanyProfile.prototype, "tags", void 0);
+__decorate([
+    (0, typeorm_1.CreateDateColumn)(),
+    __metadata("design:type", Date)
+], CompanyProfile.prototype, "createdAt", void 0);
+__decorate([
+    (0, typeorm_1.UpdateDateColumn)(),
+    __metadata("design:type", Date)
+], CompanyProfile.prototype, "updatedAt", void 0);
+exports.CompanyProfile = CompanyProfile = __decorate([
+    (0, typeorm_1.Entity)('company_profiles'),
+    (0, typeorm_1.Index)(['dedupKey'], { unique: true })
+], CompanyProfile);
+//# sourceMappingURL=company-profile.entity.js.map

package/dist/assets/server/extraction/entities/company-profile.entity.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"company-profile.entity.js","sourceRoot":"","sources":["../../../src/extraction/entities/company-profile.entity.ts"],"names":[],"mappings":";;;;;;;;;;;;AAAA,qCAOiB;AAKV,IAAM,cAAc,GAApB,MAAM,cAAc;IAEzB,EAAE,CAAS;IAGX,MAAM,CAAS;IAGf,SAAS,CAAS;IAGlB,QAAQ,CAAS;IAGjB,IAAI,CAAS;IAGb,GAAG,CAAsB;IAGzB,SAAS,CAAsB;IAG/B,WAAW,CAAS;IAGpB,SAAS,CAAS;IAGlB,IAAI,CAAW;IAGf,SAAS,CAAO;IAGhB,SAAS,CAAO;CACjB,CAAA;AApCY,wCAAc;AAEzB;IADC,IAAA,gCAAsB,GAAE;;0CACd;AAGX;IADC,IAAA,gBAAM,EAAC,EAAE,IAAI,EAAE,SAAS,EAAE,MAAM,EAAE,EAAE,EAAE,CAAC;;8CACzB;AAGf;IADC,IAAA,gBAAM,EAAC,EAAE,IAAI,EAAE,SAAS,EAAE,MAAM,EAAE,GAAG,EAAE,CAAC;;iDACvB;AAGlB;IADC,IAAA,gBAAM,EAAC,EAAE,IAAI,EAAE,SAAS,EAAE,MAAM,EAAE,GAAG,EAAE,CAAC;;gDACxB;AAGjB;IADC,IAAA,gBAAM,EAAC,EAAE,IAAI,EAAE,SAAS,EAAE,MAAM,EAAE,GAAG,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAC;;4CAC5C;AAGb;IADC,IAAA,gBAAM,EAAC,EAAE,IAAI,EAAE,aAAa,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAC;;2CACvB;AAGzB;IADC,IAAA,gBAAM,EAAC,EAAE,IAAI,EAAE,aAAa,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAC;;iDACjB;AAG/B;IADC,IAAA,gBAAM,EAAC,EAAE,IAAI,EAAE,SAAS,EAAE,MAAM,EAAE,EAAE,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAC;;mDACpC;AAGpB;IADC,IAAA,gBAAM,EAAC,EAAE,IAAI,EAAE,SAAS,EAAE,MAAM,EAAE,EAAE,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAC;;iDACtC;AAGlB;IADC,IAAA,gBAAM,EAAC,EAAE,IAAI,EAAE,aAAa,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAC;;4CACjC;AAGf;IADC,IAAA,0BAAgB,GAAE;8BACR,IAAI;iDAAC;AAGhB;IADC,IAAA,0BAAgB,GAAE;8BACR,IAAI;iDAAC;yBAnCL,cAAc;IAF1B,IAAA,gBAAM,EAAC,kBAAkB,CAAC;IAC1B,IAAA,eAAK,EAAC,CAAC,UAAU,CAAC,EAAE,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC;GACzB,cAAc,CAoC1B"}

package/dist/assets/server/extraction/entities/job-posting.entity.d.ts

@@ -0,0 +1,15 @@
+export declare class JobPosting {
+    id: number;
+    source: string;
+    sourceUrl: string;
+    dedupKey: string;
+    title: string;
+    company: string;
+    raw: Record<string, any>;
+    extracted: Record<string, any>;
+    ruleVersion: string;
+    sessionId: string;
+    tags: string[];
+    createdAt: Date;
+    updatedAt: Date;
+}