boss-agent-cli 1.3.0__py3-none-any.whl

boss_agent_cli/__init__.py

@@ -0,0 +1 @@
+__version__ = "1.3.0"

boss_agent_cli/api/boss.yaml

@@ -0,0 +1,209 @@
+base_url: "https://www.zhipin.com"
+
+web_pages:
+  geek_base: "/web/geek"
+  geek_job: "/web/geek/job"
+  geek_recommend: "/web/geek/recommend"
+  geek_chat: "/web/geek/chat"
+  geek_resume: "/web/geek/resume"
+
+default_headers:
+  User-Agent: "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/145.0.0.0 Safari/537.36"
+  sec-ch-ua: '"Chromium";v="145", "Not(A:Brand";v="99", "Google Chrome";v="145"'
+  sec-ch-ua-mobile: "?0"
+  sec-ch-ua-platform: '"macOS"'
+  Sec-Fetch-Dest: "empty"
+  Sec-Fetch-Mode: "cors"
+  Sec-Fetch-Site: "same-origin"
+  Accept: "application/json, text/plain, */*"
+  Accept-Language: "zh-CN,zh;q=0.9,en;q=0.8"
+  DNT: "1"
+
+response_codes:
+  success: 0
+  stoken_expired: 37
+  rate_limited: 9
+  account_risk: 36
+
+endpoints:
+  search:
+    method: GET
+    path: "/wapi/zpgeek/search/joblist.json"
+    referer: "/web/geek/job"
+  recommend:
+    method: GET
+    path: "/wapi/zpgeek/pc/recommend/job/list.json"
+    referer: "/web/geek/recommend"
+  detail:
+    method: GET
+    path: "/wapi/zpgeek/job/detail.json"
+    referer: "/web/geek/job"
+  greet:
+    method: POST
+    path: "/wapi/zpgeek/friend/add.json"
+    referer: "/web/geek/chat"
+  job_card:
+    method: GET
+    path: "/wapi/zpgeek/job/card.json"
+    referer: "/web/geek/job"
+  user_info:
+    method: GET
+    path: "/wapi/zpuser/wap/getUserInfo.json"
+    referer: "/"
+  resume_baseinfo:
+    method: GET
+    path: "/wapi/zpgeek/resume/baseinfo/query.json"
+    referer: "/web/geek/resume"
+  resume_expect:
+    method: GET
+    path: "/wapi/zpgeek/resume/expect/query.json"
+    referer: "/web/geek/resume"
+  deliver_list:
+    method: GET
+    path: "/wapi/zprelation/resume/geekDeliverList"
+    referer: "/web/geek/chat"
+  friend_list:
+    method: GET
+    path: "/wapi/zprelation/friend/getGeekFriendList.json"
+    referer: "/web/geek/chat"
+  interview_data:
+    method: GET
+    path: "/wapi/zpinterview/geek/interview/data.json"
+    referer: "/web/geek/job"
+  job_history:
+    method: GET
+    path: "/wapi/zpgeek/history/joblist.json"
+    referer: "/web/geek/job"
+  chat_history:
+    method: GET
+    path: "/wapi/zpchat/geek/historyMsg"
+    referer: "/web/geek/chat"
+  friend_label_add:
+    method: GET
+    path: "/wapi/zprelation/friend/label/addMark"
+    referer: "/web/geek/chat"
+  friend_label_delete:
+    method: GET
+    path: "/wapi/zprelation/friend/label/deleteMark"
+    referer: "/web/geek/chat"
+  exchange_request:
+    method: POST
+    path: "/wapi/zpchat/exchange/request"
+    referer: "/web/geek/chat"
+
+lookups:
+  city:
+    北京: "101010100"
+    上海: "101020100"
+    广州: "101280100"
+    深圳: "101280600"
+    杭州: "101210100"
+    成都: "101270100"
+    南京: "101190100"
+    武汉: "101200100"
+    西安: "101110100"
+    苏州: "101190400"
+    长沙: "101250100"
+    郑州: "101180100"
+    重庆: "101040100"
+    天津: "101030100"
+    合肥: "101220100"
+    厦门: "101230200"
+    济南: "101120100"
+    青岛: "101120200"
+    大连: "101070200"
+    宁波: "101210400"
+    福州: "101230100"
+    东莞: "101281600"
+    珠海: "101280700"
+    佛山: "101280800"
+    昆明: "101290100"
+    贵阳: "101260100"
+    太原: "101100100"
+    南昌: "101240100"
+    南宁: "101300100"
+    石家庄: "101090100"
+    哈尔滨: "101050100"
+    长春: "101060100"
+    沈阳: "101070100"
+    海口: "101310100"
+    兰州: "101160100"
+    乌鲁木齐: "101130100"
+    无锡: "101190200"
+    常州: "101191100"
+    温州: "101210700"
+    惠州: "101280300"
+
+  salary:
+    3K以下: "401"
+    3-5K: "402"
+    5-10K: "403"
+    10-15K: "404"
+    10-20K: "405"
+    20-50K: "406"
+    50K以上: "407"
+
+  experience:
+    应届: "108"
+    1年以内: "101"
+    1-3年: "103"
+    3-5年: "104"
+    5-10年: "105"
+    10年以上: "106"
+
+  education:
+    大专: "202"
+    本科: "203"
+    硕士: "204"
+    博士: "205"
+
+  scale:
+    0-20人: "301"
+    20-99人: "302"
+    100-499人: "303"
+    500-999人: "304"
+    1000-9999人: "305"
+    10000人以上: "306"
+
+  industry:
+    不限: "0"
+    互联网: "100020"
+    电子商务: "100021"
+    游戏: "100024"
+    软件/信息服务: "100032"
+    人工智能: "100901"
+    大数据: "100902"
+    云计算: "100903"
+    区块链: "100904"
+    物联网: "100905"
+    金融: "100101"
+    银行: "100102"
+    保险: "100103"
+    证券/基金: "100104"
+    教育培训: "100200"
+    医疗健康: "100300"
+    房地产: "100400"
+    汽车: "100500"
+    物流/运输: "100600"
+    广告/传媒: "100700"
+    消费品: "100800"
+    制造业: "101000"
+    能源/环保: "101100"
+    政府/非营利: "101200"
+    农业: "101300"
+
+  stage:
+    不限: "0"
+    未融资: "801"
+    天使轮: "802"
+    A轮: "803"
+    B轮: "804"
+    C轮: "805"
+    D轮及以上: "806"
+    已上市: "807"
+    不需要融资: "808"
+
+  job_type:
+    全职: "1901"
+    兼职: "1903"
+    实习: "1903"

boss_agent_cli/api/browser_client.py

@@ -0,0 +1,325 @@
+"""Browser-based API client — uses patchright to make requests from a real browser context.
+
+stoken is generated by the browser's JS environment automatically, eliminating expiration issues.
+Used for high-risk operations: search, recommend, greet, batch-greet.
+
+Supports two modes:
+  1. CDP mode (preferred): connects to user's existing Chrome via DevTools Protocol
+  2. Patchright mode (fallback): launches a headless Chromium instance
+"""
+import sys
+from pathlib import Path
+
+from patchright.sync_api import sync_playwright
+
+from boss_agent_cli.api import endpoints
+from boss_agent_cli.api.throttle import RequestThrottle
+
+HOME_URL = "https://www.zhipin.com/"
+CDP_DEFAULT_URL = "http://localhost:9222"
+
+# 超时常量
+_CDP_PROBE_TIMEOUT = 3           # CDP 探测 HTTP 超时（秒）
+_NAV_TIMEOUT_MS = 15000          # 页面导航超时（毫秒）
+
+# macOS / Linux / Windows Chrome user data 默认路径
+_CHROME_USER_DATA_CANDIDATES = [
+	Path.home() / "Library/Application Support/Google/Chrome",              # macOS
+	Path.home() / ".config/google-chrome",                                  # Linux
+	Path.home() / "AppData/Local/Google/Chrome/User Data",                  # Windows
+]
+
+
+class BrowserSession:
+	"""Persistent browser session that makes API calls via page.evaluate(fetch).
+
+	Tries CDP connection to user's Chrome first; falls back to headless patchright.
+	"""
+
+	def __init__(self, cookies: dict, user_agent: str, *, delay: tuple[float, float] = (1.5, 3.0), cdp_url: str | None = None, logger=None):
+		self._throttle = RequestThrottle(delay)
+		self._pw = None
+		self._browser = None
+		self._context = None
+		self._page = None
+		self._cookies = cookies
+		self._user_agent = user_agent
+		self._started = False
+		self._cdp_url = cdp_url
+		self._is_cdp = False
+		self._own_context = False  # 是否由我们创建的 context（需要在 close 时清理）
+		self._logger = logger
+		self._bridge_client = None
+		self._is_bridge = False
+
+	def _log(self, message: str) -> None:
+		"""通过注入的 logger 输出，受 --log-level 控制。"""
+		if self._logger:
+			self._logger.info(message)
+		else:
+			print(message, file=sys.stderr)
+
+	def _ensure_started(self):
+		if self._started:
+			return
+
+		# 优先尝试 Bridge 模式（Chrome 扩展 + daemon，零配置）
+		if self._try_bridge():
+			return
+
+		self._pw = sync_playwright().start()
+
+		# 第二优先：CDP 连接用户 Chrome（真实指纹 + 天然登录态）
+		if self._try_cdp():
+			return
+
+		# 兜底：启动 headless patchright
+		self._start_headless()
+
+	def _try_bridge(self) -> bool:
+		"""尝试通过 Browser Bridge（Chrome 扩展 + daemon）连接。"""
+		try:
+			from boss_agent_cli.bridge.client import BridgeClient
+			client = BridgeClient()
+			if not client.is_running():
+				return False
+			if not client.is_extension_connected():
+				self._log("[boss] Bridge daemon 运行中但扩展未连接")
+				return False
+			self._bridge_client = client
+			self._started = True
+			self._is_cdp = False
+			self._is_bridge = True
+			self._log("[boss] Bridge 模式连接成功（Chrome 扩展 + daemon）")
+			return True
+		except Exception:
+			return False
+
+	def _try_cdp(self) -> bool:
+		"""Try connecting to user's Chrome via CDP.
+
+		Attempts in order:
+		  1. Explicit cdp_url if provided
+		  2. Default http://localhost:9222 (+ auto WS fallback)
+		  3. WebSocket URL from Chrome's DevToolsActivePort file
+		"""
+		urls_to_try = []
+		if self._cdp_url:
+			urls_to_try.append(self._cdp_url)
+		urls_to_try.append(CDP_DEFAULT_URL)
+
+		# 从 DevToolsActivePort 文件读取 WebSocket URL
+		ws_url = self._read_devtools_active_port()
+		if ws_url:
+			urls_to_try.append(ws_url)
+
+		for url in urls_to_try:
+			if self._try_connect(url):
+				return True
+			# HTTP URL 连接失败时，尝试从 /json/version 获取 WS URL
+			if url.startswith("http"):
+				ws = self._fetch_ws_url(url)
+				if ws and self._try_connect(ws):
+					return True
+		return False
+
+	def _try_connect(self, url: str) -> bool:
+		"""Attempt a single CDP connection, reusing user's existing context.
+
+		Reuses the first existing browser context (preserves login state and
+		avoids Chrome's automation detection on new contexts). Only creates
+		a new context when none exists.
+		"""
+		try:
+			self._browser = self._pw.chromium.connect_over_cdp(url)
+			contexts = self._browser.contexts
+
+			if contexts:
+				# 复用用户现有 context（保留登录态 + 规避 automation 检测）
+				self._context = contexts[0]
+				self._own_context = False  # 非我们创建，close 时不关闭
+			else:
+				# 没有已存在 context，创建新的并注入 cookies
+				self._context = self._browser.new_context()
+				if self._cookies:
+					self._context.add_cookies([
+						{"name": name, "value": value, "domain": ".zhipin.com", "path": "/"}
+						for name, value in self._cookies.items()
+					])
+				self._own_context = True
+
+			self._page = self._context.new_page()
+			# CDP 模式下用较长超时 + commit 级等待（避免 networkidle 卡住）
+			try:
+				self._page.goto(HOME_URL, wait_until="commit", timeout=_NAV_TIMEOUT_MS)
+			except Exception:
+				pass  # 即使导航超时，页面 JS 环境已可用
+			self._started = True
+			self._is_cdp = True
+			reuse_label = "复用用户 context" if not self._own_context else "新建 context"
+			self._log(f"[boss] CDP 连接成功 ({url})，{reuse_label}")
+			return True
+		except Exception:
+			if self._browser:
+				try:
+					self._browser.close()
+				except Exception:
+					pass
+				self._browser = None
+			return False
+
+	@staticmethod
+	def _fetch_ws_url(http_url: str) -> str | None:
+		"""Fetch WebSocket debugger URL from Chrome's /json/version endpoint."""
+		import httpx
+		try:
+			resp = httpx.get(f"{http_url}/json/version", timeout=_CDP_PROBE_TIMEOUT)
+			return resp.json().get("webSocketDebuggerUrl")
+		except Exception:
+			return None
+
+	@staticmethod
+	def _read_devtools_active_port() -> str | None:
+		"""Read Chrome's DevToolsActivePort file to get WS endpoint."""
+		for user_data_dir in _CHROME_USER_DATA_CANDIDATES:
+			port_file = user_data_dir / "DevToolsActivePort"
+			if port_file.exists():
+				try:
+					lines = port_file.read_text().strip().splitlines()
+					if len(lines) >= 2:
+						port = lines[0].strip()
+						ws_path = lines[1].strip()
+						return f"ws://127.0.0.1:{port}{ws_path}"
+				except Exception:
+					continue
+		return None
+
+	def _start_headless(self):
+		"""Fallback: launch headless patchright Chromium."""
+		self._browser = self._pw.chromium.launch(headless=True)
+		self._context = self._browser.new_context(
+			user_agent=self._user_agent,
+			viewport={"width": 1280, "height": 800},
+			locale="zh-CN",
+			timezone_id="Asia/Shanghai",
+		)
+		self._own_context = True
+		self._context.add_cookies([
+			{"name": name, "value": value, "domain": ".zhipin.com", "path": "/"}
+			for name, value in self._cookies.items()
+		])
+		self._page = self._context.new_page()
+		self._page.goto(HOME_URL, wait_until="domcontentloaded")
+		self._page.wait_for_load_state("networkidle")
+		self._started = True
+		self._is_cdp = False
+		self._log("[boss] CDP 不可用（提示：需以 --remote-debugging-port=9222 启动 Chrome），降级到 headless patchright")
+
+	# ── Core request via browser fetch() ─────────────────────────────
+
+	def request(self, method: str, url: str, *, params: dict | None = None, data: dict | None = None) -> dict:
+		self._ensure_started()
+		self._throttle.wait()
+
+		referer = endpoints.REFERER_MAP.get(url, f"{endpoints.BASE_URL}/")
+
+		# Bridge 模式：通过扩展 fetch
+		if self._is_bridge and self._bridge_client:
+			import urllib.parse
+			full_url = url
+			if params:
+				full_url = f"{url}?{urllib.parse.urlencode(params)}"
+			result = self._bridge_client.fetch_json(
+				full_url,
+				method=method,
+				data=data,
+				referer=referer,
+			)
+			self._throttle.mark()
+			return result
+
+		# Playwright 模式（CDP 或 headless）
+		result = self._page.evaluate("""
+			async ({method, url, params, data, referer}) => {
+				try {
+					let fetchUrl = url;
+					if (params && Object.keys(params).length > 0) {
+						const sp = new URLSearchParams();
+						for (const [k, v] of Object.entries(params)) {
+							if (v !== null && v !== undefined) sp.append(k, String(v));
+						}
+						fetchUrl = url + '?' + sp.toString();
+					}
+
+					const options = {
+						method: method,
+						credentials: 'include',
+						headers: {
+							'Accept': 'application/json, text/plain, */*',
+							'Referer': referer,
+							'X-Requested-With': 'XMLHttpRequest',
+						},
+					};
+
+					if (method === 'POST' && data) {
+						options.headers['Content-Type'] = 'application/x-www-form-urlencoded';
+						const formData = new URLSearchParams();
+						for (const [k, v] of Object.entries(data)) {
+							formData.append(k, String(v));
+						}
+						options.body = formData.toString();
+					}
+
+					const resp = await fetch(fetchUrl, options);
+					return await resp.json();
+				} catch (e) {
+					return {code: -1, message: e.message, zpData: {}};
+				}
+			}
+		""", {"method": method, "url": url, "params": params or {}, "data": data, "referer": referer})
+
+		self._throttle.mark()
+		return result
+
+	# ── Lifecycle ────────────────────────────────────────────────────
+
+	@property
+	def is_cdp(self) -> bool:
+		"""Whether this session is using CDP connection to user's Chrome."""
+		return self._is_cdp
+
+	def close(self):
+		if self._is_bridge and self._bridge_client:
+			try:
+				self._bridge_client.close_window()
+			except Exception:
+				pass
+			self._bridge_client = None
+			self._started = False
+			return
+
+		if self._is_cdp:
+			# CDP 模式：关闭我们创建的 page 和 context，不关闭用户浏览器
+			if self._page:
+				try:
+					self._page.close()
+				except Exception:
+					pass
+			if self._own_context and self._context:
+				try:
+					self._context.close()
+				except Exception:
+					pass
+		else:
+			# Headless 模式：关闭整个浏览器实例
+			if self._browser:
+				self._browser.close()
+		if self._pw:
+			self._pw.stop()
+		self._started = False
+
+	def __enter__(self):
+		return self
+
+	def __exit__(self, *args):
+		self.close()