job-pro 0.7.0 → 0.7.2

package/dist/galaxyuniversal.js

@@ -1,37 +1,389 @@
-// 银河通用 / Galaxy Universal (Galbot — embodied AI robotics) — stub for `job-pro`.
+// 银河通用 / Galaxy Universal (Galbot) recruiting via app.mokahr.com.
 //
-// STATUS: stub-only. Galaxy Universal lists careers via Moka social-recruitment
-// (orgId 165930, slug yinhetongyong), which is auth-gated for anonymous access.
-// Probe results:
-//   www.galbot.com/careers, galaxyuniversal.com/careers → no public API discoverable
-//   app.mokahr.com/social-recruitment/yinhetongyong/165930  → Moka SPA, session auth required
-import { extractResumeSignals, checkResume } from "./tencent.js";
-export { checkResume };
-const SOURCE = "galbot.com";
-const STUB_MESSAGE = "Galaxy Universal / 银河通用: no public job API — Moka social-recruitment portal " +
-    "(yinhetongyong/165930) requires session auth (verified Moka anon path is gated).";
-export async function searchPositions(_opts = {}) {
-    return { ok: false, source: SOURCE, message: STUB_MESSAGE, query: {}, positions: [] };
-}
-export async function fetchAllPositions(_opts = {}) {
-    return { ok: false, source: SOURCE, message: STUB_MESSAGE, total: 0, fetched: 0, positions: [] };
+// ============================================================
+// HOW THIS WORKS (probed 2026-05):
+//
+//   www.galbot.com/careers → SPA shell that links to
+//     https://app.mokahr.com/social-recruitment/yinhetongyong/165929
+//   (corrected from the previously documented 165930, which 404s).
+//
+//   The Moka SSR HTML embeds `<input id="init-data" value="<JSON>">`
+//   with the first 15 jobs and `jobStats.total`. Pagination uses
+//   POST /api/outer/ats-apply/website/jobs/v2?orgId=yinhetongyong
+//   (AES-128-CBC envelope: key=necromancer, iv=aesIv from SSR HTML).
+//
+// CONFIRMED MOKA ORG:
+//   slug=yinhetongyong, siteId=165929, mode=social
+//   Portal: https://app.mokahr.com/social-recruitment/yinhetongyong/165929
+//
+// PositionSummary field mapping:
+//   post_id       ← job.id
+//   title         ← job.title
+//   project       ← job.zhineng?.name (e.g. "算法类")
+//   recruit_label ← job.commitment || hireMode label
+//   bgs           ← job.department?.name (e.g. "算法中心")
+//   work_cities   ← locations[].cityId → label via jobsGroupedByLocation
+//   apply_url     ← portal#/jobs/{id}
+import { extractResumeSignals, scoreOverlap, checkResume } from "./tencent.js";
+import { createDecipheriv } from "node:crypto";
+export { checkResume, extractResumeSignals, scoreOverlap };
+const SOURCE = "app.mokahr.com/yinhetongyong";
+const ORG_SLUG = "yinhetongyong";
+const SITE_ID = 165929;
+const PORTAL_URL = `https://app.mokahr.com/social-recruitment/${ORG_SLUG}/${SITE_ID}`;
+const API_ENDPOINT = "https://app.mokahr.com/api/outer/ats-apply/website/jobs/v2";
+const DEFAULT_HEADERS = {
+    "User-Agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/124.0.0.0 Safari/537.36",
+    Accept: "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8",
+    "Accept-Language": "zh-CN,zh;q=0.9,en;q=0.8",
+};
+function htmlDecode(s) {
+    return s
+        .replace(/&quot;/g, '"')
+        .replace(/&amp;/g, "&")
+        .replace(/&lt;/g, "<")
+        .replace(/&gt;/g, ">")
+        .replace(/&#x27;/g, "'")
+        .replace(/&#39;/g, "'");
+}
+function parseInitData(html) {
+    const m = html.match(/<input[^>]*id="init-data"[^>]*value="([^"]+)"/);
+    if (!m)
+        return null;
+    try {
+        return JSON.parse(htmlDecode(m[1]));
+    }
+    catch {
+        return null;
+    }
+}
+async function fetchPortalHtml() {
+    let r1;
+    try {
+        r1 = await fetch(PORTAL_URL, { method: "GET", headers: DEFAULT_HEADERS, redirect: "manual" });
+    }
+    catch (err) {
+        return { ok: false, message: `network error: ${err instanceof Error ? err.message : err}` };
+    }
+    const cookies = [];
+    const headersAny = r1.headers;
+    if (typeof headersAny.getSetCookie === "function") {
+        for (const v of headersAny.getSetCookie.call(r1.headers) ?? []) {
+            const c = v.split(";")[0];
+            if (c)
+                cookies.push(c);
+        }
+    }
+    if (cookies.length === 0) {
+        const raw = r1.headers.get("set-cookie");
+        if (raw)
+            cookies.push(...raw.split(/,(?=[^;]+=)/).map((c) => c.split(";")[0].trim()));
+    }
+    const cookieHeader = cookies.join("; ");
+    let r2;
+    try {
+        r2 = await fetch(PORTAL_URL, {
+            method: "GET",
+            headers: { ...DEFAULT_HEADERS, Cookie: cookieHeader },
+            redirect: "follow",
+        });
+    }
+    catch (err) {
+        return { ok: false, message: `network error: ${err instanceof Error ? err.message : err}` };
+    }
+    if (!r2.ok)
+        return { ok: false, message: `HTTP ${r2.status}` };
+    return { ok: true, html: await r2.text(), cookieHeader, message: "ok" };
+}
+function decryptMoka(envelope, aesIv) {
+    if (!envelope.data || !envelope.necromancer)
+        return null;
+    try {
+        const decipher = createDecipheriv("aes-128-cbc", Buffer.from(envelope.necromancer, "utf8"), Buffer.from(aesIv, "utf8"));
+        const plain = Buffer.concat([
+            decipher.update(Buffer.from(envelope.data, "base64")),
+            decipher.final(),
+        ]);
+        return JSON.parse(plain.toString("utf8"));
+    }
+    catch {
+        return null;
+    }
+}
+async function fetchEncryptedPage(pageNum, pageSize, aesIv, cookieHeader) {
+    let response;
+    try {
+        response = await fetch(`${API_ENDPOINT}?orgId=${encodeURIComponent(ORG_SLUG)}`, {
+            method: "POST",
+            headers: {
+                ...DEFAULT_HEADERS,
+                Accept: "application/json,*/*",
+                "Content-Type": "application/json",
+                Origin: "https://app.mokahr.com",
+                Referer: PORTAL_URL,
+                Cookie: cookieHeader,
+            },
+            body: JSON.stringify({
+                orgId: ORG_SLUG,
+                siteId: String(SITE_ID),
+                pageNum,
+                pageSize,
+                needStat: true,
+            }),
+        });
+    }
+    catch (err) {
+        return { ok: false, message: `network error: ${err instanceof Error ? err.message : err}` };
+    }
+    if (!response.ok)
+        return { ok: false, message: `HTTP ${response.status}` };
+    let envelope;
+    try {
+        envelope = await response.json();
+    }
+    catch {
+        return { ok: false, message: "bad JSON" };
+    }
+    const decoded = decryptMoka(envelope, aesIv);
+    if (!decoded || decoded.code !== 0 || !decoded.data) {
+        return { ok: false, message: decoded?.msg || envelope.msg || "decrypt error" };
+    }
+    return {
+        ok: true,
+        jobs: decoded.data.jobs ?? [],
+        total: decoded.data.jobStats?.total ?? 0,
+        message: "ok",
+    };
+}
+function buildCityMap(groups) {
+    const out = {};
+    if (!groups)
+        return out;
+    for (const g of groups) {
+        if (typeof g.cityId === "number" && g.label)
+            out[g.cityId] = g.label;
+    }
+    return out;
+}
+function workCities(job, cityMap) {
+    const uniq = [];
+    for (const loc of job.locations ?? []) {
+        const label = (typeof loc.cityId === "number" && cityMap[loc.cityId]) || loc.country || "";
+        if (label && !uniq.includes(label))
+            uniq.push(label);
+    }
+    return uniq.join(" / ");
+}
+function recruitLabel(job) {
+    if (job.commitment)
+        return job.commitment;
+    if (job.hireMode === 1)
+        return "全职";
+    if (job.hireMode === 2)
+        return "实习";
+    return "";
+}
+function summarize(job, cityMap) {
+    return {
+        post_id: String(job.id),
+        title: job.title ?? "",
+        project: job.zhineng?.name ?? "",
+        recruit_label: recruitLabel(job),
+        bgs: job.department?.name ?? "",
+        work_cities: workCities(job, cityMap),
+        apply_url: `${PORTAL_URL}#/jobs/${encodeURIComponent(job.id)}`,
+    };
+}
+function matchesKeyword(job, kw) {
+    if (!kw)
+        return true;
+    const lc = kw.toLowerCase();
+    return ((job.title ?? "").toLowerCase().includes(lc) ||
+        (job.zhineng?.name ?? "").toLowerCase().includes(lc) ||
+        (job.department?.name ?? "").toLowerCase().includes(lc));
+}
+export async function searchPositions(opts = {}) {
+    const pageSize = opts.pageSize ?? 20;
+    const page = opts.page ?? 1;
+    const keyword = opts.keyword ?? "";
+    const portal = await fetchPortalHtml();
+    if (!portal.ok || !portal.html) {
+        return {
+            ok: false,
+            source: SOURCE,
+            message: portal.message,
+            query: { keyword, page, pageSize },
+            positions: [],
+            total: 0,
+        };
+    }
+    const init = parseInitData(portal.html);
+    if (!init || !init.jobs || !init.jobStats) {
+        return {
+            ok: false,
+            source: SOURCE,
+            message: "Moka init-data missing jobs/jobStats",
+            query: { keyword, page, pageSize },
+            positions: [],
+            total: 0,
+        };
+    }
+    const cityMap = buildCityMap(init.jobsGroupedByLocation);
+    let jobs = init.jobs;
+    const total = init.jobStats.total ?? jobs.length;
+    if (page > 1 && init.aesIv && portal.cookieHeader) {
+        const more = await fetchEncryptedPage(page, pageSize, init.aesIv, portal.cookieHeader);
+        if (!more.ok || !more.jobs) {
+            return {
+                ok: false,
+                source: SOURCE,
+                message: `pagination failed: ${more.message}`,
+                query: { keyword, page, pageSize },
+                positions: [],
+                total,
+            };
+        }
+        jobs = more.jobs;
+    }
+    const filtered = jobs.filter((j) => matchesKeyword(j, keyword)).slice(0, pageSize);
+    return {
+        ok: true,
+        source: SOURCE,
+        query: { keyword, page, pageSize },
+        page,
+        page_size: pageSize,
+        total,
+        positions: filtered.map((j) => summarize(j, cityMap)),
+    };
+}
+export async function fetchAllPositions(opts = {}) {
+    const pageSize = opts.pageSize ?? 20;
+    const maxPages = Math.max(1, opts.maxPages ?? 50);
+    const keyword = opts.keyword ?? "";
+    const portal = await fetchPortalHtml();
+    if (!portal.ok || !portal.html) {
+        return {
+            ok: false,
+            source: SOURCE,
+            message: portal.message,
+            total: 0,
+            fetched: 0,
+            positions: [],
+        };
+    }
+    const init = parseInitData(portal.html);
+    if (!init || !init.jobs || !init.jobStats || !init.aesIv) {
+        return {
+            ok: false,
+            source: SOURCE,
+            message: "Moka init-data missing required fields",
+            total: 0,
+            fetched: 0,
+            positions: [],
+        };
+    }
+    const cityMap = buildCityMap(init.jobsGroupedByLocation);
+    const total = init.jobStats.total ?? 0;
+    const collected = [...init.jobs];
+    let page = 2;
+    while (collected.length < total && page <= maxPages) {
+        const more = await fetchEncryptedPage(page, pageSize, init.aesIv, portal.cookieHeader ?? "");
+        if (!more.ok || !more.jobs || more.jobs.length === 0)
+            break;
+        collected.push(...more.jobs);
+        page += 1;
+    }
+    const filtered = collected.filter((j) => matchesKeyword(j, keyword));
+    return {
+        ok: true,
+        source: SOURCE,
+        total,
+        fetched: filtered.length,
+        positions: filtered.map((j) => summarize(j, cityMap)),
+    };
 }
 export async function fetchPositionDetail(postId) {
-    return { ok: false, source: SOURCE, message: STUB_MESSAGE, post_id: postId };
+    return {
+        ok: false,
+        source: SOURCE,
+        message: "Moka detail endpoint is also AES-encrypted and not implemented; " +
+            "use the apply_url deeplink for the full JD.",
+        post_id: postId,
+        apply_url: `${PORTAL_URL}#/jobs/${encodeURIComponent(postId)}`,
+    };
 }
 export async function fetchDictionaries() {
-    return { ok: false, source: SOURCE, message: STUB_MESSAGE };
+    const portal = await fetchPortalHtml();
+    if (!portal.ok || !portal.html) {
+        return { ok: false, source: SOURCE, message: portal.message };
+    }
+    const init = parseInitData(portal.html);
+    if (!init) {
+        return { ok: false, source: SOURCE, message: "Moka init-data missing" };
+    }
+    return {
+        ok: true,
+        source: SOURCE,
+        locations: init.jobsGroupedByLocation ?? [],
+        moka_org: { slug: ORG_SLUG, siteId: SITE_ID, url: PORTAL_URL },
+    };
 }
 export async function listNotices() {
-    return { ok: false, source: SOURCE, message: STUB_MESSAGE, notices: [] };
+    return {
+        ok: false,
+        source: SOURCE,
+        message: "Galaxy Universal / 银河通用: no public notices endpoint",
+        notices: [],
+    };
 }
 export async function getNotice(noticeId) {
-    return { ok: false, source: SOURCE, message: STUB_MESSAGE, notice_id: noticeId };
+    return {
+        ok: false,
+        source: SOURCE,
+        message: "Galaxy Universal / 银河通用: no public notices endpoint",
+        notice_id: noticeId,
+    };
 }
 export async function findNoticesByQuestion(question, _opts = {}) {
-    return { ok: false, source: SOURCE, question, message: STUB_MESSAGE, matches: [] };
+    return {
+        ok: false,
+        source: SOURCE,
+        question,
+        message: "Galaxy Universal / 银河通用: no public notices endpoint",
+        matches: [],
+    };
 }
-export async function matchResume(text, _opts = {}) {
+export async function matchResume(text, opts = {}) {
     const { terms, cities } = extractResumeSignals(text ?? "");
-    return { ok: false, source: SOURCE, extracted_terms: terms, city_preferences: cities, matches: [], message: STUB_MESSAGE };
+    const candidates = Math.max(20, opts.candidates ?? 100);
+    const all = await fetchAllPositions({
+        pageSize: 20,
+        maxPages: Math.ceil(candidates / 15),
+    });
+    if (!all.ok) {
+        return {
+            ok: false,
+            source: SOURCE,
+            extracted_terms: terms,
+            city_preferences: cities,
+            matches: [],
+            message: all.message,
+        };
+    }
+    const topN = Math.max(1, opts.topN ?? 10);
+    const scored = all.positions
+        .map((p) => ({
+        p,
+        score: scoreOverlap(`${p.title} ${p.project} ${p.bgs}`, terms, cities).score,
+    }))
+        .sort((a, b) => b.score - a.score)
+        .slice(0, topN)
+        .map((x) => x.p);
+    return {
+        ok: true,
+        source: SOURCE,
+        extracted_terms: terms,
+        city_preferences: cities,
+        matches: scored,
+    };
 }

package/dist/horizonrobotics.js

@@ -1,66 +1,45 @@
-// 地平线 (Horizon Robotics) — stub adapter for `job-pro`.
-//
-// STATUS: stub-only. Horizon's careers portal is hosted on Moka and gated
-// behind the Moka SPA's login flow; the per-slug JSON endpoint returns the
-// "您访问的页面不存在" Moka error page for anonymous requests.
+// 地平线 (Horizon Robotics) careers adapter for `job-pro`.
 //
 // ============================================================
-// RECONNAISSANCE RESULTS (probed 2026-05):
-//
-//   https://career.horizon.ai             — 000 (no public DNS / unreachable)
-//   https://campus.horizon.ai             — 000 (no public DNS / unreachable)
-//   https://horizon.app.mokahr.com        — Moka SPA shell renders, but
-//     /api/career/website/horizon/jobs returns {"code":-1,"message":"您访问的页面不存在"}
+// API DISCOVERY (probed 2026-05-16 via puppeteer-core network capture)
 //
-//   Feishu ATSX:    horizonrobotics.jobs.feishu.cn — HTTP 405 (DNS but no portal)
-//                   horizon.jobs.feishu.cn         — HTTP 405
-//                   horizon-robotics.jobs.feishu.cn — HTTP 400 (no portal)
-//   Greenhouse:     horizon / horizon-robotics    — HTTP 404
-//   Lever:          horizonrobotics                — HTTP 404
+// Horizon's careers run on `wecruit.hotjob.cn`, the same Beisen Wecruit
+// stack as SenseTime (see cli/src/sensetime.ts). The `/{SU…}/pb/<channel>.html`
+// SPA path returns nginx 405 on any anonymous POST. The real XHR is fired
+// at the sibling `/wecruit/positionInfo/listPosition/{SU…}` route.
 //
-//   The Moka portal exists (the slug 'horizonrobotics' returns 200 page
-//   shell) but the underlying job-list endpoint requires the Moka SPA's
-//   user-session JWT, which is only minted post-login.
+// Channels (probed 2026-05-16):
+//   * school  — `SU6409ef49bef57c635fd390a6` (校园招聘 / 实习生) ~84 positions
+//   * social  — `SU64819a4f2f9d2433ba8b043a` (社会招聘)            ~216 positions
 //
-// Conclusion: no unauthenticated public API. Visit Moka careers shell at
-// https://app.mokahr.com/social-recruitment/horizonrobotics for the portal.
-import { extractResumeSignals, scoreOverlap, checkResume } from "./tencent.js";
-export { checkResume };
-const SOURCE = "app.mokahr.com/horizonrobotics";
-const STUB_MESSAGE = "Horizon Robotics (地平线): Moka careers portal (slug horizonrobotics) is gated — the public " +
-    "/api/career/website/horizon/jobs endpoint returns the Moka 'page not found' error for anonymous " +
-    "requests; positions are visible only after a candidate session is established. No Greenhouse / Lever / " +
-    "Feishu tenant provisioned. No unauthenticated public API available.";
-export async function searchPositions(_opts = {}) {
-    return { ok: false, source: SOURCE, message: STUB_MESSAGE, query: {}, positions: [] };
-}
-export async function fetchAllPositions(_opts = {}) {
-    return { ok: false, source: SOURCE, message: STUB_MESSAGE, total: 0, fetched: 0, positions: [] };
-}
-export async function fetchPositionDetail(postId) {
-    return { ok: false, source: SOURCE, message: STUB_MESSAGE, post_id: postId };
-}
-export async function fetchDictionaries() {
-    return { ok: false, source: SOURCE, message: STUB_MESSAGE };
-}
-export async function listNotices() {
-    return { ok: false, source: SOURCE, message: STUB_MESSAGE, notices: [] };
-}
-export async function getNotice(noticeId) {
-    return { ok: false, source: SOURCE, message: STUB_MESSAGE, notice_id: noticeId };
-}
-export async function findNoticesByQuestion(question, _opts = {}) {
-    return { ok: false, source: SOURCE, question, message: STUB_MESSAGE, matches: [] };
-}
-export async function matchResume(text, _opts = {}) {
-    const { terms, cities } = extractResumeSignals(text ?? "");
-    return {
-        ok: false,
-        source: SOURCE,
-        extracted_terms: terms,
-        city_preferences: cities,
-        matches: [],
-        message: STUB_MESSAGE,
-    };
-}
-export { extractResumeSignals, scoreOverlap };
+// Anonymous, no token, no cookie. See cli/src/wecruit.ts for the shared
+// factory: POST to `/wecruit/positionInfo/listPosition/{channelId}` with
+// `application/x-www-form-urlencoded` body containing
+// `isFrompb=true&recruitType=<1|2>&pageSize=N&currentPage=N`. Response is
+// `{ data:{ pageForm:{ totalPage, pageData[…] } }, state:"200" }`.
+import { createAdapter } from "./wecruit.js";
+const adapter = createAdapter({
+    host: "wecruit.hotjob.cn",
+    label: "Horizon Robotics",
+    channels: [
+        {
+            channelId: "SU6409ef49bef57c635fd390a6",
+            recruitType: "campus",
+            pagePath: "school",
+        },
+        {
+            channelId: "SU64819a4f2f9d2433ba8b043a",
+            recruitType: "social",
+            pagePath: "social",
+        },
+    ],
+});
+export const searchPositions = adapter.searchPositions;
+export const fetchAllPositions = adapter.fetchAllPositions;
+export const fetchPositionDetail = adapter.fetchPositionDetail;
+export const fetchDictionaries = adapter.fetchDictionaries;
+export const listNotices = adapter.listNotices;
+export const getNotice = adapter.getNotice;
+export const findNoticesByQuestion = adapter.findNoticesByQuestion;
+export const matchResume = adapter.matchResume;
+export const checkResume = adapter.checkResume;