first commit

Co-Authored-By: Claude <noreply@anthropic.com>
This commit is contained in:
kunkunxiaomaoandClaude committed 2026-11-16 20:45:22 +08:00
commit 9c5fef1a4d
412 files changed
+94582

No files matched your search

+1
View File
@@ -0,0 +1 @@
"""Local service entry points for Sakura downloader."""
+500
View File
@@ -0,0 +1,500 @@
from __future__ import annotations
import json
import mimetypes
import time
import uuid
from dataclasses import asdict
from http.server import BaseHTTPRequestHandler, ThreadingHTTPServer
from pathlib import Path
from urllib.parse import parse_qs, urlparse
from pixiv_app.core.auth import SessionStore
from pixiv_app.core.cookie_import import cookie_summary, cookies_to_header, cookies_to_playwright, parse_cookie_text
from pixiv_app.core.gallery_service import LocalGalleryService
from pixiv_app.core.library import DEFAULT_LIBRARY_DB
from pixiv_app.core.origin_follow_sync import OriginFollowSyncService
from pixiv_app.core.paths import app_session_file, downloads_root, plugin_roots, webui_root
from pixiv_app.core.plugin.manager import PluginManager
from pixiv_app.core.thumbnails import DEFAULT_THUMBNAIL_DIR
from pixiv_app.core.x_media_service import XMediaService
class GalleryApiServer:
def __init__(
self,
*,
host: str = "127.0.0.1",
port: int = 8765,
db_path: str | Path = DEFAULT_LIBRARY_DB,
thumbnail_dir: str | Path = DEFAULT_THUMBNAIL_DIR,
web_root: str | Path | None = None,
) -> None:
self.host = host
self.port = port
self.service = LocalGalleryService(db_path=db_path, thumbnail_dir=thumbnail_dir)
self.x_media_service = XMediaService()
self.origin_follow_sync_service = OriginFollowSyncService(
db_path=str(db_path),
gallery_service=self.service,
x_media_service=self.x_media_service,
)
self.plugin_manager = PluginManager(plugin_roots())
self.plugin_manager.load_all()
self.plugin_preview_cache: dict[str, tuple[str, list]] = {}
self.web_root = Path(web_root) if web_root else webui_root()
def serve_forever(self) -> None:
service = self.service
x_media_service = self.x_media_service
origin_follow_sync_service = self.origin_follow_sync_service
plugin_manager = self.plugin_manager
plugin_preview_cache = self.plugin_preview_cache
web_root = self.web_root
Handler = type(
"GalleryApiHandlerForServer",
(GalleryApiHandler,),
{
"gallery_service": service,
"x_media_service": x_media_service,
"origin_follow_sync_service": origin_follow_sync_service,
"plugin_manager": plugin_manager,
"plugin_preview_cache": plugin_preview_cache,
"web_root": web_root,
},
)
server = ThreadingHTTPServer((self.host, self.port), Handler)
try:
server.serve_forever()
finally:
server.server_close()
class GalleryApiHandler(BaseHTTPRequestHandler):
gallery_service: LocalGalleryService
x_media_service: XMediaService
origin_follow_sync_service: OriginFollowSyncService
plugin_manager: PluginManager
plugin_preview_cache: dict[str, tuple[str, list]]
web_root: Path
def do_GET(self) -> None:
parsed = urlparse(self.path)
query = parse_qs(parsed.query)
try:
if parsed.path in {"/", "/index.html"}:
self._send_static_file(self.web_root / "index.html")
elif parsed.path.startswith("/assets/"):
self._send_static_file(self.web_root / parsed.path.removeprefix("/assets/"))
elif parsed.path == "/api/health":
self._send_json({"status": "ok"})
elif parsed.path == "/api/session/status":
self._send_json(self._handle_session_status())
elif parsed.path == "/api/plugins/pages":
self._send_json(self._handle_plugin_pages())
elif parsed.path == "/api/artworks":
self._send_json(self._handle_artworks(query))
elif parsed.path == "/api/favorites":
self._send_json(self.gallery_service.list_favorites(limit=_int(query, "limit", 60), offset=_int(query, "offset", 0)))
elif parsed.path.startswith("/api/artworks/"):
self._send_json(self._handle_artwork_detail(parsed.path))
elif parsed.path == "/api/tags":
self._send_json(self.gallery_service.list_tags(query=_first(query, "q"), limit=_int(query, "limit", 100)))
elif parsed.path == "/api/artists":
self._send_json(self.gallery_service.list_artists(query=_first(query, "q"), limit=_int(query, "limit", 100)))
elif parsed.path.startswith("/api/artists/"):
self._send_json(self._handle_artist_detail(parsed.path, query))
elif parsed.path == "/api/followed-artists":
self._send_json(self.gallery_service.list_followed_artists())
elif parsed.path == "/api/thumbnails/build":
self._send_json({"items": self.gallery_service.build_missing_thumbnails(limit=_int(query, "limit", 200))})
elif parsed.path == "/api/import/legacy":
self._send_json(self.gallery_service.scan_legacy_downloads(root=_first(query, "root", "Sakura_Downloads")))
elif parsed.path == "/api/sync/followed":
self._send_json(
self.gallery_service.sync_followed_artists(
max_new_per_artist=_int(query, "max_new", 20),
download=_bool(query, "download", True),
)
)
elif parsed.path == "/api/x/local-media":
self._send_json(
self.x_media_service.list_local_media(
limit=_int(query, "limit", 160),
offset=_int(query, "offset", 0),
)
)
elif parsed.path == "/api/x/follows":
self._send_json(self.x_media_service.list_followed_authors())
elif parsed.path.startswith("/media/thumbnail/"):
self._send_file(self._resolve_media_path(parsed.path, thumbnail=True))
elif parsed.path.startswith("/media/source/"):
self._send_file(self._resolve_media_path(parsed.path, thumbnail=False))
elif parsed.path.startswith("/media/x-local/"):
self._send_file(self.x_media_service.resolve_local_media(parsed.path.rsplit("/", 1)[-1]))
else:
self._send_json({"error": "not found"}, status=404)
except ValueError as exc:
self._send_json({"error": str(exc)}, status=400)
except Exception as exc:
self._send_json({"error": str(exc)}, status=500)
def do_POST(self) -> None:
parsed = urlparse(self.path)
try:
payload = self._read_json_body()
if parsed.path == "/api/favorites":
self._send_json(
self.gallery_service.set_favorite(
int(payload["artwork_id"]),
bool(payload.get("favorite", True)),
str(payload.get("note", "")),
)
)
elif parsed.path == "/api/followed-artists":
self._send_json(
self.gallery_service.set_followed_artist(
payload.get("artist_id", ""),
bool(payload.get("followed", True)),
str(payload.get("name", "")),
)
)
elif parsed.path == "/api/import/legacy":
self._send_json(self.gallery_service.scan_legacy_downloads(root=str(payload.get("root", "Sakura_Downloads"))))
elif parsed.path == "/api/sync/followed":
self._send_json(
self.gallery_service.sync_followed_artists(
cookie=str(payload.get("cookie", "")),
max_new_per_artist=int(payload.get("max_new", 20)),
download=bool(payload.get("download", True)),
)
)
elif parsed.path == "/api/cookies/import":
self._send_json(self._handle_cookie_import(payload))
elif parsed.path == "/api/origin-follows/sync":
self._send_json(
self.origin_follow_sync_service.sync(
str(payload.get("platform", "")),
)
)
elif parsed.path == "/api/x/media/preview":
self._send_json(
self.x_media_service.preview_author_media(
str(payload.get("username", "")),
media_type=str(payload.get("media_type", "all")),
max_items=int(payload.get("max_items", 120)),
)
)
elif parsed.path == "/api/x/media/download":
selected_ids = payload.get("selected_ids")
if selected_ids is not None and not isinstance(selected_ids, list):
raise ValueError("selected_ids 必须是数组。")
self._send_json(
self.x_media_service.download_preview(
str(payload.get("preview_id", "")),
selected_ids=[str(item) for item in selected_ids] if selected_ids is not None else None,
)
)
elif parsed.path == "/api/x/local-media/delete":
media_ids = payload.get("ids", [])
if not isinstance(media_ids, list):
raise ValueError("ids 必须是数组。")
self._send_json(self.x_media_service.delete_local_media([str(item) for item in media_ids]))
elif parsed.path == "/api/x/follows":
self._send_json(
self.x_media_service.set_followed_author(
str(payload.get("username", "")),
followed=bool(payload.get("followed", True)),
creator_name=str(payload.get("name", "")),
)
)
elif parsed.path == "/api/x/follows/sync":
self._send_json(
self.x_media_service.sync_followed_authors(
download=bool(payload.get("download", True)),
max_new_per_author=int(payload.get("max_new", 20)),
)
)
elif parsed.path == "/api/xiaohongshu/preview":
self._send_json(self._handle_plugin_preview("小红书", payload))
elif parsed.path == "/api/xiaohongshu/download":
self._send_json(self._handle_plugin_download(payload))
else:
self._send_json({"error": "not found"}, status=404)
except ValueError as exc:
self._send_json({"error": str(exc)}, status=400)
except Exception as exc:
self._send_json({"error": str(exc)}, status=500)
def do_OPTIONS(self) -> None:
self.send_response(204)
self._send_cors_headers()
self.end_headers()
def log_message(self, format: str, *args) -> None:
return None
def _handle_artworks(self, query: dict[str, list[str]]) -> dict:
artist_id = _first(query, "artist_id")
return self.gallery_service.list_artworks(
query=_first(query, "q"),
search_scope=_first(query, "search_scope") or "all",
tag=_first(query, "tag"),
artist_id=int(artist_id) if artist_id else None,
artwork_type=_first(query, "type"),
include_restricted=_bool(query, "include_restricted", True),
limit=_int(query, "limit", 60),
offset=_int(query, "offset", 0),
ensure_thumbnails=_bool(query, "ensure_thumbnails", True),
)
def _handle_plugin_pages(self) -> dict:
self.plugin_manager.reload_all_from_disk()
items = []
if "X" in self.plugin_manager.plugins:
items.append(
{
"id": "x-media",
"view": "xMedia",
"title": "X 作者媒体",
"description": "按作者名称解析图片或视频并选择下载",
}
)
if "小红书" in self.plugin_manager.plugins:
plugin = self.plugin_manager.plugins["小红书"]
items.append(
{
"id": "xiaohongshu",
"view": "xiaohongshu",
"title": "小红书",
"description": "搜索关键词或解析笔记链接,预览后下载",
"available": bool(plugin.validate()),
}
)
return {"items": items}
def _handle_plugin_preview(self, plugin_name: str, payload: dict) -> dict:
plugin = self.plugin_manager.plugins.get(plugin_name)
if plugin is None:
raise ValueError(f"未安装插件: {plugin_name}")
if not plugin.validate():
raise ValueError(f"{plugin_name} 插件依赖不可用,请先安装 Playwright。")
target = str(payload.get("target", "")).strip()
if not target:
raise ValueError("请输入解析目标。")
limit = max(1, min(int(payload.get("limit", 20)), 50))
resources = plugin.parse(target)[:limit]
preview_id = uuid.uuid4().hex
self.plugin_preview_cache[preview_id] = (plugin_name, resources)
return {
"preview_id": preview_id,
"count": len(resources),
"items": [_resource_to_json(item, i) for i, item in enumerate(resources)],
}
def _handle_plugin_download(self, payload: dict) -> dict:
preview_id = str(payload.get("preview_id", ""))
cached = self.plugin_preview_cache.get(preview_id)
if cached is None:
raise ValueError("预览结果已过期,请重新解析。")
plugin_name, resources = cached
plugin = self.plugin_manager.plugins.get(plugin_name)
if plugin is None:
raise ValueError(f"插件已卸载: {plugin_name}")
selected_ids = payload.get("selected_ids")
selected = set(str(item) for item in selected_ids) if isinstance(selected_ids, list) else set()
if selected:
resources = [item for item in resources if item.id in selected]
save_root = downloads_root()
results = []
downloaded = 0
failed = 0
for resource in resources:
try:
paths = plugin.download(resource, save_root)
downloaded += len(paths)
results.append({
"id": resource.id,
"ok": bool(paths),
"paths": [str(path) for path in paths],
"synced_files": 0,
})
except Exception as exc:
failed += 1
results.append({"id": resource.id, "ok": False, "error": str(exc)})
return {
"requested": len(resources),
"downloaded_files": downloaded,
"synced_files": 0,
"failed_items": failed,
"items": results,
}
def _handle_artwork_detail(self, path: str) -> dict:
artwork_id = int(path.rsplit("/", 1)[-1])
item = self.gallery_service.get_artwork(artwork_id)
if item is None:
raise ValueError(f"Artwork not found: {artwork_id}")
return item
def _handle_artist_detail(self, path: str, query: dict[str, list[str]]) -> dict:
artist_id = int(path.rsplit("/", 1)[-1])
item = self.gallery_service.get_artist(
artist_id,
limit=_int(query, "limit", 60),
offset=_int(query, "offset", 0),
)
if item is None:
raise ValueError(f"Artist not found: {artist_id}")
return item
def _resolve_media_path(self, path: str, *, thumbnail: bool) -> Path:
parts = path.strip("/").split("/")
if len(parts) != 4:
raise ValueError("Expected /media/{thumbnail|source}/{artwork_id}/{page_index}")
artwork_id = int(parts[2])
page_index = int(parts[3])
media_path = (
self.gallery_service.get_thumbnail_path(artwork_id, page_index=page_index)
if thumbnail
else self.gallery_service.get_source_file_path(artwork_id, page_index=page_index)
)
if media_path is None:
raise ValueError("Media file not found")
return media_path
def _handle_cookie_import(self, payload: dict) -> dict:
platform = str(payload.get("platform", "")).strip().lower()
text = str(payload.get("text", ""))
if platform in {"pixiv", "px"}:
cookies = parse_cookie_text(text)
pixiv_cookies = cookies_to_playwright(cookies, ("pixiv.net",))
header = cookies_to_header(pixiv_cookies, ("pixiv.net",))
store = SessionStore(app_session_file())
session_payload = store.load()
session_payload.update(
{
"login_mode": "cookie",
"cookie": header,
"cookie_json": pixiv_cookies,
"login_id": "",
"saved_at": time.strftime("%Y-%m-%d %H:%M:%S"),
}
)
store.save(session_payload)
return {"platform": "pixiv", "count": len(pixiv_cookies), "summary": cookie_summary(pixiv_cookies)}
if platform in {"x", "twitter"}:
result = self.x_media_service.import_cookie_text(text)
result["platform"] = "x"
return result
raise ValueError("platform 必须是 pixiv 或 x。")
def _handle_session_status(self) -> dict:
payload = SessionStore(app_session_file()).load()
pixiv_cookie = str(payload.get("cookie") or "").strip()
cookie_json = payload.get("cookie_json")
cloud = payload.get("cloud_sync") if isinstance(payload.get("cloud_sync"), dict) else {}
return {
"pixiv": {
"has_cookie": bool(pixiv_cookie or cookie_json),
"saved_at": str(payload.get("saved_at") or ""),
"login_mode": str(payload.get("login_mode") or ""),
},
"cloud": {
"enabled": bool(cloud.get("enabled")),
"has_token": bool(str(cloud.get("access_token") or "").strip()),
"base_url": str(cloud.get("base_url") or "").strip(),
},
}
def _send_json(self, payload: dict | list, *, status: int = 200) -> None:
body = json.dumps(payload, ensure_ascii=False).encode("utf-8")
self.send_response(status)
self.send_header("Content-Type", "application/json; charset=utf-8")
self._send_cors_headers()
self.send_header("Content-Length", str(len(body)))
self.end_headers()
self.wfile.write(body)
def _send_file(self, path: Path) -> None:
content_type = mimetypes.guess_type(path.name)[0] or "application/octet-stream"
body = path.read_bytes()
self.send_response(200)
self.send_header("Content-Type", content_type)
self.send_header("Content-Length", str(len(body)))
self.end_headers()
self.wfile.write(body)
def _send_static_file(self, path: Path) -> None:
if not path.exists() or not path.is_file():
self._send_json({"error": "not found"}, status=404)
return
self._send_file(path)
def _send_cors_headers(self) -> None:
self.send_header("Access-Control-Allow-Origin", "*")
self.send_header("Access-Control-Allow-Methods", "GET, POST, OPTIONS")
self.send_header("Access-Control-Allow-Headers", "Content-Type")
def _read_json_body(self) -> dict:
length = int(self.headers.get("Content-Length", "0") or "0")
if length == 0:
return {}
body = self.rfile.read(length)
payload = json.loads(body.decode("utf-8"))
if not isinstance(payload, dict):
raise ValueError("JSON body must be an object")
return payload
def _first(query: dict[str, list[str]], key: str, default: str = "") -> str:
values = query.get(key)
if not values:
return default
return values[0]
def _int(query: dict[str, list[str]], key: str, default: int) -> int:
value = _first(query, key)
if not value:
return default
return int(value)
def _bool(query: dict[str, list[str]], key: str, default: bool) -> bool:
value = _first(query, key)
if not value:
return default
return value.lower() in {"1", "true", "yes", "on"}
def _resource_to_json(resource, index: int) -> dict:
payload = asdict(resource)
payload["index"] = index
payload["kind"] = "video" if any(file.get("type") == "video" for file in resource.files) else "image"
payload["file_count"] = len(resource.files)
return payload
def _plugin_platform_slug(plugin_name: str) -> str:
lowered = plugin_name.lower()
if "douyin" in lowered or "抖音" in plugin_name:
return "douyin"
if "xiaohongshu" in lowered or "小红书" in plugin_name:
return "xiaohongshu"
if lowered == "x":
return "x"
return lowered.replace(" ", "_")[:20] or "plugin"
def _resource_file_type(resource, index: int, path: Path) -> str:
if index < len(resource.files):
file_type = str(resource.files[index].get("type") or "").lower()
if file_type == "video":
return "video"
mime_type = mimetypes.guess_type(str(path))[0] or ""
return "video" if mime_type.startswith("video/") else "illust"
+30
View File
@@ -0,0 +1,30 @@
from __future__ import annotations
import argparse
import logging
from pixiv_app.services.gallery_api import GalleryApiServer
def main() -> None:
parser = argparse.ArgumentParser(description="Run local Sakura gallery API")
parser.add_argument("--host", default="127.0.0.1")
parser.add_argument("--port", type=int, default=8765)
parser.add_argument("--db", default="pixiv_app_library.db")
parser.add_argument("--thumbnail-dir", default="runtime/thumbnails")
parser.add_argument("--web-root", default="")
args = parser.parse_args()
logging.basicConfig(level=logging.INFO, format="%(asctime)s %(levelname)s %(message)s")
logging.info("Local gallery API listening on http://%s:%s", args.host, args.port)
GalleryApiServer(
host=args.host,
port=args.port,
db_path=args.db,
thumbnail_dir=args.thumbnail_dir,
web_root=args.web_root or None,
).serve_forever()
if __name__ == "__main__":
main()
+356
View File
@@ -0,0 +1,356 @@
"""
X (Twitter) authentication helper — Playwright **persistent context** only.
- First-time login is always manual (including email verification); no credential stuffing.
- Session persistence is Chromium's user_data_dir (cookies + localStorage live inside it).
- Optional JSON cookie export into ``profiles/x/<profile>/cookies/`` for backup only.
Directory layout (created automatically)::
profiles/
x/
<profile_key>/
chromium_profile/ # user_data_dir → persistent session (real storage)
cookies/ # optional Playwright cookie exports (JSON)
localStorage/ # reserved / notes (Chromium keeps LS inside chromium_profile)
Usage::
auth = XAuthManager(profile_key="default")
await auth.start_session()
page = await auth.get_page()
await page.goto("https://x.com/i/flow/login")
# ... user completes login manually ...
if await auth.is_logged_in():
...
await auth.close()
"""
from __future__ import annotations
import asyncio
import json
import logging
from pathlib import Path
from typing import Any
from playwright.async_api import BrowserContext, Page, Playwright, async_playwright
from pixiv_app.core.paths import profiles_root
from pixiv_app.core.playwright_runtime import configure_playwright_runtime_env, preferred_chromium_executable
logger = logging.getLogger(__name__)
def _default_profiles_root() -> Path:
return profiles_root()
class XAuthManager:
"""
Manage a long-lived Playwright Chromium **persistent** profile for x.com.
Not thread-safe: use one instance per logical browser session, or guard externally.
"""
HOME_URL = "https://x.com/home"
LOGIN_FLOW_HINT = "https://x.com/i/flow/login"
def __init__(
self,
*,
profile_key: str = "default",
profiles_root: Path | None = None,
headless: bool = True,
auto_reconnect: bool = True,
max_health_retries: int = 2,
) -> None:
"""
:param profile_key: Subdirectory under ``profiles/x/`` for multi-account switching.
:param profiles_root: Defaults to ``<project>/profiles``.
:param headless: Headless for automated checks; use ``False`` for first manual login.
:param auto_reconnect: If ``start_session`` finds a dead context, restart Playwright.
:param max_health_retries: Retries when validating session after transient failures.
"""
self._profile_key = profile_key
root = profiles_root or _default_profiles_root()
self._profile_home = root / "x" / profile_key
self._user_data_dir = self._profile_home / "chromium_profile"
self._cookies_dir = self._profile_home / "cookies"
self._localstorage_dir = self._profile_home / "localStorage"
self._headless = headless
self._auto_reconnect = auto_reconnect
self._max_health_retries = max(1, max_health_retries)
self._pw: Playwright | None = None
self._context: BrowserContext | None = None
self._primary_page: Page | None = None
self._ensure_storage_layout()
def _ensure_storage_layout(self) -> None:
"""Create profile dirs. Real cookie/localStorage persistence is under ``chromium_profile``."""
self._user_data_dir.mkdir(parents=True, exist_ok=True)
self._cookies_dir.mkdir(parents=True, exist_ok=True)
self._localstorage_dir.mkdir(parents=True, exist_ok=True)
@property
def user_data_dir(self) -> Path:
"""Playwright ``user_data_dir`` passed to ``launch_persistent_context``."""
return self._user_data_dir
def _context_alive(self) -> bool:
if self._context is None:
return False
try:
_ = self._context.pages
return True
except Exception:
return False
async def start_session(self, *, force_new: bool = False) -> BrowserContext:
"""
Launch or reuse persistent Chromium. Session is restored from disk automatically.
:param force_new: Close existing context and open a fresh one (same profile dir).
"""
if self._context_alive() and not force_new:
return self._context # type: ignore[return-value]
if self._context is not None:
await self.close()
self._pw = await async_playwright().start()
try:
self._context = await self._launch_persistent(self._pw)
except Exception:
await self._pw.stop()
self._pw = None
raise
self._primary_page = None
logger.info("XAuthManager: persistent context ready (%s)", self._user_data_dir)
return self._context
async def _launch_persistent(self, pw: Playwright) -> BrowserContext:
"""Single place for launch_persistent_context configuration."""
self._user_data_dir.mkdir(parents=True, exist_ok=True)
configure_playwright_runtime_env()
base_kw: dict[str, Any] = {
"user_data_dir": str(self._user_data_dir),
"headless": self._headless,
"viewport": {"width": 1280, "height": 800},
"locale": "zh-CN",
"ignore_default_args": ["--enable-automation"],
"args": ["--disable-blink-features=AutomationControlled"],
"user_agent": (
"Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 "
"(KHTML, like Gecko) Chrome/131.0.0.0 Safari/537.36"
),
}
import os
env_ch = os.environ.get("X_PLAYWRIGHT_CHANNEL", "").strip()
local_executable = preferred_chromium_executable()
attempts: list[dict[str, Any]] = []
seen: set[str] = set()
for extra in (
{"channel": env_ch} if env_ch else {},
{"executable_path": str(local_executable)} if local_executable else {},
{"channel": "chrome"},
{},
):
key = repr(extra)
if key in seen:
continue
seen.add(key)
attempts.append(extra)
last_err: Exception | None = None
for extra in attempts:
kw = dict(base_kw)
kw.update(extra)
try:
ctx = await pw.chromium.launch_persistent_context(**kw)
await ctx.add_init_script(
"""
try {
Object.defineProperty(navigator, 'webdriver', { get: () => undefined });
} catch (e) {}
"""
)
return ctx
except Exception as exc:
last_err = exc
logger.warning("launch_persistent_context failed (%s): %s", extra, exc)
continue
assert last_err is not None
raise last_err
async def get_context(self) -> BrowserContext:
if not self._context_alive():
if self._auto_reconnect:
return await self.start_session()
raise RuntimeError("No browser session; call start_session() first.")
return self._context # type: ignore[return-value]
async def get_page(self, *, new: bool = False) -> Page:
"""
Return a page for navigation. Default reuses one tab; set ``new=True`` for a fresh tab.
"""
ctx = await self.get_context()
if new:
return await ctx.new_page()
if self._primary_page is not None:
try:
_ = self._primary_page.url
return self._primary_page
except Exception:
self._primary_page = None
if ctx.pages:
self._primary_page = ctx.pages[0]
return self._primary_page
self._primary_page = await ctx.new_page()
return self._primary_page
async def is_logged_in(self, *, page: Page | None = None, navigate: bool = True) -> bool:
"""
Heuristic login detection (DOM + URL). Not guaranteed if X changes UI.
- ``False`` if still on login / signup flow URLs or login UI is visible.
- ``True`` if home timeline / primary column or account UI is present.
"""
p = page if page is not None else await self.get_page()
if navigate:
await p.goto(self.HOME_URL, wait_until="domcontentloaded", timeout=45_000)
await asyncio.sleep(0.6)
url = (p.url or "").lower()
if any(
x in url
for x in (
"/i/flow/login",
"/i/flow/signup",
"/login",
)
):
return False
# Logged-in indicators (best-effort; X changes selectors periodically)
logged_in_selectors = (
'[data-testid="SideNav_AccountSwitcher_Button"]',
'[data-testid="AppTabBar_Home_Link"]',
'[data-testid="primaryColumn"]',
'[data-testid="tweet"]',
)
for sel in logged_in_selectors:
try:
el = await p.wait_for_selector(sel, timeout=2500, state="visible")
if el:
return True
except Exception:
continue
# Login / gate modal text (guest wall)
try:
guest = p.get_by_text(
"Sign in to X",
exact=False,
)
if await guest.count() and await guest.first.is_visible(timeout=800):
return False
except Exception:
pass
return False
async def session_healthy(self) -> bool:
"""True if browser context responds and ``is_logged_in`` succeeds within retries."""
for attempt in range(self._max_health_retries):
try:
if not self._context_alive():
return False
return await self.is_logged_in(navigate=True)
except Exception as exc:
logger.debug("session_healthy attempt %s: %s", attempt + 1, exc)
await asyncio.sleep(1.0)
return False
async def export_playwright_cookies(self, filename: str = "playwright_cookies.json") -> Path:
"""Optional backup: write context cookies to ``profiles/x/<key>/cookies/``."""
ctx = await self.get_context()
cookies = await ctx.cookies()
path = self._cookies_dir / filename
path.write_text(json.dumps(cookies, indent=2), encoding="utf-8")
return path
async def close(self) -> None:
"""Close Playwright; persistent data remains on disk for next run."""
if self._primary_page is not None:
try:
await self._primary_page.close()
except Exception:
pass
self._primary_page = None
if self._context is not None:
try:
await self._context.close()
except Exception:
pass
self._context = None
if self._pw is not None:
try:
await self._pw.stop()
except Exception:
pass
self._pw = None
logger.info("XAuthManager: closed (%s)", self._profile_key)
# --- sync helpers for Tkinter / scripts without asyncio ---
def start_session_sync(self, *, force_new: bool = False) -> BrowserContext:
return asyncio.run(self.start_session(force_new=force_new))
def get_page_sync(self, *, new: bool = False) -> Page:
return asyncio.run(self.get_page(new=new))
def is_logged_in_sync(self, *, navigate: bool = True) -> bool:
return asyncio.run(self.is_logged_in(navigate=navigate))
def close_sync(self) -> None:
asyncio.run(self.close())
async def _demo() -> None:
logging.basicConfig(level=logging.INFO)
auth = XAuthManager(profile_key="default", headless=False)
try:
await auth.start_session()
page = await auth.get_page()
await page.goto(XAuthManager.LOGIN_FLOW_HINT, wait_until="domcontentloaded")
print(
"浏览器已打开登录流程。请在窗口中手动完成登录(含邮箱验证),完成后回到终端按 Enter…",
)
await asyncio.to_thread(input)
ok = await auth.is_logged_in()
print("is_logged_in:", ok)
if ok:
path = await auth.export_playwright_cookies()
print("Cookies 备份:", path)
finally:
await auth.close()
if __name__ == "__main__":
asyncio.run(_demo())
+156
View File
@@ -0,0 +1,156 @@
"""
Helpers for X (Twitter) web login flow in Playwright.
X often opens a combined sheet titled 「注册 X」 even from /i/flow/login; users must switch to
the actual sign-in path or they loop after entering email. These helpers only automate UI clicks
(no credential stuffing, no bypass).
"""
from __future__ import annotations
import asyncio
import re
from playwright.async_api import Page
LOGIN_URL_X = "https://x.com/i/flow/login"
LOGIN_URL_TWITTER = "https://twitter.com/i/flow/login"
# When X shows this, login inside Playwright often cannot proceed (rate limit / bot heuristics).
X_FLOW_ERROR_RECOVERY_HINT_CN = (
"这是 X 端返回的通用错误,常见于自动化浏览器或风控,并不代表本程序损坏。\n\n"
"可行做法(任选):\n"
"1)点击弹窗「重试」后等待几分钟再试,或更换网络(如手机热点);\n"
"2)在插件管理中「清除本 Profile 浏览器数据」后重新「手动登录」;\n"
"3)用本机普通 Chrome 登录 x.com,导出 Cookie 为 JSON,放到 runtime/x_cookies.json,再点「同步到抓取」。\n\n"
"我们无法绕过 X/Google 的风控策略。"
)
async def detect_x_generic_flow_error(page: Page) -> bool:
"""True if the grey modal 「出错了 / Something went wrong」 is likely visible."""
try:
dlg = page.locator('[role="dialog"]')
if await dlg.count() > 0:
text = (await dlg.first.inner_text())[:1200]
low = text.lower()
if ("出错了" in text or "something went wrong" in low) and (
"重试" in text or "reload" in low or "retry" in low or "重新加载" in text
):
return True
except Exception:
pass
return False
async def prefer_login_over_signup(page: Page, *, max_attempts: int = 5) -> None:
"""
If the modal looks like sign-up (注册 X / Sign up for X), try to switch to login.
Safe to call multiple times; no-op when login sheet is already shown.
"""
for _ in range(max_attempts):
await asyncio.sleep(0.35)
# Already looks like login-oriented sheet (best-effort)
try:
login_heading = page.get_by_role("heading", name=re.compile(r"登录.*X|Sign in to X|登录或注册", re.I))
if await login_heading.count() > 0 and await login_heading.first.is_visible(timeout=400):
return
except Exception:
pass
signup_visible = False
try:
su = page.get_by_text(re.compile(r"注册 X|Sign up for X"), exact=False)
if await su.count() > 0:
signup_visible = await su.first.is_visible(timeout=500)
except Exception:
pass
if not signup_visible:
try:
dlg = page.locator('[role="dialog"]')
if await dlg.count() > 0:
t = await dlg.first.inner_text()
if "注册 X" in t or "Sign up for X" in t:
signup_visible = True
except Exception:
pass
if not signup_visible:
return
clicked = await _try_click_login_switch(page)
if not clicked:
break
async def _try_click_login_switch(page: Page) -> bool:
"""Return True if a qualifying control was clicked."""
# Known Twitter/X test ids (may change)
for tid in ("signupSheetSwitchToLogin", "switch_to_login"):
try:
loc = page.locator(f'[data-testid="{tid}"]')
if await loc.count() > 0 and await loc.first.is_visible(timeout=600):
await loc.first.click(timeout=3000)
await asyncio.sleep(0.8)
return True
except Exception:
continue
# Footer / inverse of 「还没有账号?注册」 → often 「登录」 link nearby
try:
inv = page.get_by_text(re.compile(r"已有账号|已有帐户|Already have an account", re.I))
if await inv.count() > 0:
await inv.first.click(timeout=3500)
await asyncio.sleep(0.85)
return True
except Exception:
pass
try:
plain = page.get_by_role("link", name=re.compile(r"^(登录|Sign in)$", re.I))
if await plain.count() > 0:
await plain.first.click(timeout=3500)
await asyncio.sleep(0.85)
return True
except Exception:
pass
try:
href_login = page.locator('a[href*="login"]').filter(has_not_text=re.compile(r"Google|Apple", re.I))
if await href_login.count() > 0 and await href_login.first.is_visible(timeout=600):
await href_login.first.click(timeout=3500)
await asyncio.sleep(0.85)
return True
except Exception:
pass
try:
link = page.locator('a[href="/login"]').first
if await link.is_visible(timeout=600):
await link.click(timeout=3500)
await asyncio.sleep(0.85)
return True
except Exception:
pass
return False
async def goto_x_login_entry(page: Page, *, timeout: float = 90_000) -> None:
"""Navigate to official login flow and try to leave sign-up shell."""
await page.goto(LOGIN_URL_X, wait_until="domcontentloaded", timeout=timeout)
await prefer_login_over_signup(page)
# Some locales keep showing 「注册 X」 on x.com; twitter.com entry occasionally differs.
try:
su = page.get_by_text(re.compile(r"注册 X|Sign up for X"), exact=False)
if await su.count() > 0 and await su.first.is_visible(timeout=700):
await page.goto(LOGIN_URL_TWITTER, wait_until="domcontentloaded", timeout=timeout)
await prefer_login_over_signup(page)
except Exception:
pass