Files
holy-python/database.py

415 lines
17 KiB
Python
Raw Normal View History

2026-08-04 14:02:45 +08:00
import json
import sqlite3
from contextlib import contextmanager
from datetime import datetime
from typing import Any, Iterator
from config import Config
def now_iso() -> str:
return datetime.now().astimezone().isoformat(timespec="seconds")
@contextmanager
def connection() -> Iterator[sqlite3.Connection]:
conn = sqlite3.connect(Config.DATABASE_PATH, timeout=30)
conn.row_factory = sqlite3.Row
conn.execute("PRAGMA foreign_keys = ON")
try:
yield conn
conn.commit()
finally:
conn.close()
def init_database() -> None:
"""仅在服务启动时创建表,接口调用过程中不会重复执行建表。"""
Config.DATABASE_PATH.parent.mkdir(parents=True, exist_ok=True)
Config.FILE_STORAGE_PATH.mkdir(parents=True, exist_ok=True)
with connection() as conn:
conn.executescript(
"""
CREATE TABLE IF NOT EXISTS research_tasks (
id INTEGER PRIMARY KEY AUTOINCREMENT,
public_task_id TEXT NOT NULL UNIQUE,
name TEXT NOT NULL,
platform TEXT NOT NULL,
mode TEXT NOT NULL,
status TEXT NOT NULL,
execution_stage TEXT NOT NULL DEFAULT 'queued',
progress_percent INTEGER NOT NULL DEFAULT 0 CHECK(progress_percent BETWEEN 0 AND 100),
created_by_id TEXT NOT NULL,
created_by_name TEXT NOT NULL,
retry_of_task_id INTEGER,
crawler_task_id TEXT,
error_code TEXT,
error_message TEXT,
result_json TEXT,
cancel_requested_at TEXT,
started_at TEXT,
completed_at TEXT,
created_at TEXT NOT NULL,
updated_at TEXT NOT NULL,
version INTEGER NOT NULL DEFAULT 0,
FOREIGN KEY(retry_of_task_id) REFERENCES research_tasks(id)
);
CREATE INDEX IF NOT EXISTS idx_tasks_status_created
ON research_tasks(status, created_at DESC);
CREATE INDEX IF NOT EXISTS idx_tasks_creator_created
ON research_tasks(created_by_id, created_at DESC);
CREATE TABLE IF NOT EXISTS task_keywords (
id INTEGER PRIMARY KEY AUTOINCREMENT,
task_id INTEGER NOT NULL,
keyword_order INTEGER NOT NULL,
keyword TEXT NOT NULL,
normalized_keyword TEXT NOT NULL,
created_at TEXT NOT NULL,
UNIQUE(task_id, keyword_order),
UNIQUE(task_id, normalized_keyword),
FOREIGN KEY(task_id) REFERENCES research_tasks(id) ON DELETE CASCADE
);
CREATE TABLE IF NOT EXISTS task_events (
id INTEGER PRIMARY KEY AUTOINCREMENT,
task_id INTEGER NOT NULL,
event_type TEXT NOT NULL,
stage TEXT,
progress_percent INTEGER,
actor_type TEXT NOT NULL,
actor_id TEXT,
payload_json TEXT,
occurred_at TEXT NOT NULL,
FOREIGN KEY(task_id) REFERENCES research_tasks(id) ON DELETE CASCADE
);
CREATE TABLE IF NOT EXISTS task_keyword_runs (
id INTEGER PRIMARY KEY AUTOINCREMENT,
task_id INTEGER NOT NULL,
keyword_order INTEGER NOT NULL,
keyword TEXT NOT NULL,
status TEXT NOT NULL DEFAULT 'queued',
progress_percent INTEGER NOT NULL DEFAULT 0,
crawler_task_id TEXT,
result_json TEXT,
error_message TEXT,
started_at TEXT,
completed_at TEXT,
updated_at TEXT NOT NULL,
UNIQUE(task_id, keyword_order),
FOREIGN KEY(task_id) REFERENCES research_tasks(id) ON DELETE CASCADE
);
CREATE INDEX IF NOT EXISTS idx_keyword_runs_task_status
ON task_keyword_runs(task_id, status, keyword_order);
2026-08-04 14:02:45 +08:00
CREATE TABLE IF NOT EXISTS task_files (
id INTEGER PRIMARY KEY AUTOINCREMENT,
public_file_id TEXT NOT NULL UNIQUE,
task_id INTEGER NOT NULL,
name TEXT NOT NULL,
stage TEXT NOT NULL,
category TEXT NOT NULL,
format TEXT NOT NULL,
record_count INTEGER NOT NULL DEFAULT 0,
size_bytes INTEGER NOT NULL DEFAULT 0,
storage_key TEXT NOT NULL,
sha256 TEXT NOT NULL,
schema_version TEXT NOT NULL DEFAULT '1.0',
generated_at TEXT NOT NULL,
available INTEGER NOT NULL DEFAULT 1,
UNIQUE(task_id, name),
FOREIGN KEY(task_id) REFERENCES research_tasks(id) ON DELETE CASCADE
);
CREATE TABLE IF NOT EXISTS auth_users (
id INTEGER PRIMARY KEY AUTOINCREMENT,
ding_user_id TEXT NOT NULL UNIQUE,
union_id TEXT,
open_id TEXT,
name TEXT NOT NULL,
avatar TEXT,
mobile TEXT,
email TEXT,
raw_json TEXT,
last_login_at TEXT NOT NULL,
created_at TEXT NOT NULL,
updated_at TEXT NOT NULL
);
CREATE INDEX IF NOT EXISTS idx_auth_users_name
ON auth_users(name);
CREATE TABLE IF NOT EXISTS auth_sessions (
id INTEGER PRIMARY KEY AUTOINCREMENT,
session_hash TEXT NOT NULL UNIQUE,
user_id INTEGER NOT NULL,
expires_at TEXT NOT NULL,
created_at TEXT NOT NULL,
last_seen_at TEXT NOT NULL,
revoked_at TEXT,
FOREIGN KEY(user_id) REFERENCES auth_users(id) ON DELETE CASCADE
);
CREATE INDEX IF NOT EXISTS idx_auth_sessions_user
ON auth_sessions(user_id, expires_at);
CREATE TABLE IF NOT EXISTS ai_access_tokens (
id INTEGER PRIMARY KEY AUTOINCREMENT,
token_hash TEXT NOT NULL UNIQUE,
user_id INTEGER NOT NULL,
expires_at TEXT NOT NULL,
created_at TEXT NOT NULL,
last_seen_at TEXT NOT NULL,
revoked_at TEXT,
FOREIGN KEY(user_id) REFERENCES auth_users(id) ON DELETE CASCADE
);
CREATE INDEX IF NOT EXISTS idx_ai_access_tokens_user
ON ai_access_tokens(user_id, expires_at);
CREATE TABLE IF NOT EXISTS acceptance_tasks (
id INTEGER PRIMARY KEY AUTOINCREMENT,
public_task_id TEXT NOT NULL UNIQUE,
name TEXT NOT NULL,
acceptance_method TEXT NOT NULL,
status TEXT NOT NULL DEFAULT 'waiting',
progress_percent INTEGER NOT NULL DEFAULT 0,
device_count INTEGER NOT NULL DEFAULT 1,
created_by_id TEXT NOT NULL,
created_by_name TEXT NOT NULL,
error_message TEXT,
created_at TEXT NOT NULL,
updated_at TEXT NOT NULL
);
CREATE INDEX IF NOT EXISTS idx_acceptance_tasks_created
ON acceptance_tasks(created_at DESC);
CREATE TABLE IF NOT EXISTS acceptance_keywords (
id INTEGER PRIMARY KEY AUTOINCREMENT,
task_id INTEGER NOT NULL,
keyword_order INTEGER NOT NULL,
keyword TEXT NOT NULL,
created_at TEXT NOT NULL,
UNIQUE(task_id, keyword_order),
FOREIGN KEY(task_id) REFERENCES acceptance_tasks(id) ON DELETE CASCADE
);
CREATE TABLE IF NOT EXISTS acceptance_feishu_sources (
id INTEGER PRIMARY KEY AUTOINCREMENT,
task_id INTEGER NOT NULL,
source_order INTEGER NOT NULL,
source_url TEXT NOT NULL,
final_url TEXT,
source_type TEXT,
title TEXT,
content_text TEXT,
content_json TEXT,
fetch_status TEXT NOT NULL DEFAULT 'waiting',
http_status INTEGER,
error_message TEXT,
fetched_at TEXT,
created_at TEXT NOT NULL,
updated_at TEXT NOT NULL,
UNIQUE(task_id, source_order),
UNIQUE(task_id, source_url),
FOREIGN KEY(task_id) REFERENCES acceptance_tasks(id) ON DELETE CASCADE
);
CREATE TABLE IF NOT EXISTS acceptance_feishu_records (
id INTEGER PRIMARY KEY AUTOINCREMENT,
source_id INTEGER NOT NULL,
row_number INTEGER NOT NULL,
row_json TEXT NOT NULL,
created_at TEXT NOT NULL,
UNIQUE(source_id, row_number),
FOREIGN KEY(source_id) REFERENCES acceptance_feishu_sources(id) ON DELETE CASCADE
);
CREATE TABLE IF NOT EXISTS acceptance_xhs_notes (
id INTEGER PRIMARY KEY AUTOINCREMENT,
task_id INTEGER NOT NULL,
source_id INTEGER NOT NULL,
source_row_number INTEGER,
source_column_name TEXT,
note_url TEXT NOT NULL,
normalized_url TEXT NOT NULL,
note_id TEXT,
parse_status TEXT NOT NULL DEFAULT 'waiting',
parsed_json TEXT,
error_message TEXT,
parsed_at TEXT,
created_at TEXT NOT NULL,
updated_at TEXT NOT NULL,
UNIQUE(task_id, normalized_url),
FOREIGN KEY(task_id) REFERENCES acceptance_tasks(id) ON DELETE CASCADE,
FOREIGN KEY(source_id) REFERENCES acceptance_feishu_sources(id) ON DELETE CASCADE
);
CREATE INDEX IF NOT EXISTS idx_acceptance_notes_task_status
ON acceptance_xhs_notes(task_id, parse_status);
CREATE TABLE IF NOT EXISTS feishu_notes (
id INTEGER PRIMARY KEY AUTOINCREMENT,
note_url TEXT NOT NULL,
normalized_url TEXT NOT NULL UNIQUE,
note_id TEXT UNIQUE,
title TEXT,
body TEXT,
author_nickname TEXT,
author_avatar TEXT,
likes INTEGER,
collects INTEGER,
comments INTEGER,
images_json TEXT,
tags_json TEXT,
raw_json TEXT NOT NULL,
parsed_at TEXT NOT NULL,
created_at TEXT NOT NULL,
updated_at TEXT NOT NULL
);
CREATE INDEX IF NOT EXISTS idx_feishu_notes_note_id
ON feishu_notes(note_id);
CREATE TABLE IF NOT EXISTS acceptance_keyword_runs (
id INTEGER PRIMARY KEY AUTOINCREMENT,
task_id INTEGER NOT NULL,
keyword TEXT NOT NULL,
run_type TEXT NOT NULL,
run_order INTEGER NOT NULL,
status TEXT NOT NULL DEFAULT 'waiting',
crawler_task_id TEXT,
result_json TEXT,
error_message TEXT,
started_at TEXT,
completed_at TEXT,
created_at TEXT NOT NULL,
updated_at TEXT NOT NULL,
UNIQUE(task_id, run_type, keyword),
FOREIGN KEY(task_id) REFERENCES acceptance_tasks(id) ON DELETE CASCADE
);
CREATE INDEX IF NOT EXISTS idx_acceptance_keyword_runs_task
ON acceptance_keyword_runs(task_id, status);
CREATE TABLE IF NOT EXISTS acceptance_cookie_attempts (
id INTEGER PRIMARY KEY AUTOINCREMENT,
task_id INTEGER NOT NULL,
keyword_run_id INTEGER NOT NULL,
keyword TEXT NOT NULL,
credential_id TEXT NOT NULL,
attempt_number INTEGER NOT NULL,
status TEXT NOT NULL,
crawler_task_id TEXT,
result_json TEXT,
error_message TEXT,
started_at TEXT NOT NULL,
completed_at TEXT NOT NULL,
created_at TEXT NOT NULL,
UNIQUE(keyword_run_id, credential_id, attempt_number),
FOREIGN KEY(task_id) REFERENCES acceptance_tasks(id) ON DELETE CASCADE,
FOREIGN KEY(keyword_run_id) REFERENCES acceptance_keyword_runs(id) ON DELETE CASCADE
);
CREATE INDEX IF NOT EXISTS idx_cookie_attempts_run_status
ON acceptance_cookie_attempts(keyword_run_id, status);
CREATE TABLE IF NOT EXISTS acceptance_task_resources (
id INTEGER PRIMARY KEY AUTOINCREMENT,
task_id INTEGER NOT NULL,
resource_type TEXT NOT NULL,
resource_id TEXT NOT NULL,
resource_order INTEGER NOT NULL,
created_at TEXT NOT NULL,
UNIQUE(task_id, resource_type, resource_id),
FOREIGN KEY(task_id) REFERENCES acceptance_tasks(id) ON DELETE CASCADE
);
CREATE INDEX IF NOT EXISTS idx_acceptance_task_resources
ON acceptance_task_resources(task_id, resource_order);
2026-08-04 14:02:45 +08:00
"""
)
# 将旧版本保存在 acceptance_xhs_notes.parsed_json 中的成功结果迁移到
# 全局去重的笔记业务表;任务关联仍由 acceptance_xhs_notes 保留。
legacy_rows = conn.execute(
"""SELECT n.* FROM acceptance_xhs_notes n
LEFT JOIN feishu_notes f ON f.normalized_url=n.normalized_url
WHERE n.parse_status='completed' AND n.parsed_json IS NOT NULL AND f.id IS NULL"""
).fetchall()
for row in legacy_rows:
try:
payload = json.loads(row["parsed_json"])
except (TypeError, json.JSONDecodeError):
continue
upsert_feishu_note(conn, row, payload)
# 为升级前已有的 Cookie 尝试补建任务资源记录,详情页可立即展示设备进度。
legacy_resources = conn.execute(
"""SELECT task_id,credential_id,MIN(id) AS first_id
FROM acceptance_cookie_attempts
GROUP BY task_id,credential_id ORDER BY task_id,first_id"""
).fetchall()
resource_orders: dict[int, int] = {}
for row in legacy_resources:
order = resource_orders.get(row["task_id"], 0)
conn.execute(
"""INSERT OR IGNORE INTO acceptance_task_resources
(task_id,resource_type,resource_id,resource_order,created_at)
VALUES (?,'cookie',?,?,?)""",
(row["task_id"], row["credential_id"], order, now_iso()),
)
resource_orders[row["task_id"]] = order + 1
def _integer_or_none(value: Any) -> int | None:
if value is None or value == "":
return None
try:
return int(str(value).replace(",", ""))
except (TypeError, ValueError):
return None
def upsert_feishu_note(conn: sqlite3.Connection, note_row: sqlite3.Row, payload: dict[str, Any]) -> None:
"""按规范化链接更新或新增笔记;任务关联保存在 acceptance_xhs_notes。"""
now = now_iso()
note_id = payload.get("noteId") or payload.get("note_id") or note_row["note_id"]
existing = conn.execute(
"""SELECT id FROM feishu_notes
WHERE normalized_url=? OR (? IS NOT NULL AND note_id=?)
ORDER BY CASE WHEN note_id=? THEN 0 ELSE 1 END LIMIT 1""",
(note_row["normalized_url"], note_id, note_id, note_id),
).fetchone()
values = (
note_row["note_url"], note_row["normalized_url"], note_id,
payload.get("title"), payload.get("body") or payload.get("content"),
payload.get("nickname") or payload.get("author_nickname"),
payload.get("avatar") or payload.get("author_avatar"),
_integer_or_none(payload.get("likes")), _integer_or_none(payload.get("collects")),
_integer_or_none(payload.get("comments")),
json.dumps(payload.get("images") or payload.get("images_list") or [], ensure_ascii=False),
json.dumps(payload.get("tags") or [], ensure_ascii=False),
json.dumps(payload, ensure_ascii=False), now, now,
)
if existing:
conn.execute(
"""UPDATE feishu_notes SET
note_url=?,normalized_url=?,note_id=?,title=?,body=?,author_nickname=?,author_avatar=?,
likes=?,collects=?,comments=?,images_json=?,tags_json=?,raw_json=?,parsed_at=?,updated_at=?
WHERE id=?""",
(*values, existing["id"]),
)
else:
conn.execute(
"""INSERT INTO feishu_notes
(note_url,normalized_url,note_id,title,body,author_nickname,author_avatar,
likes,collects,comments,images_json,tags_json,raw_json,parsed_at,created_at,updated_at)
VALUES (?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?)
""",
(*values, now),
)
2026-08-04 14:02:45 +08:00
def json_dumps(value: Any) -> str:
return json.dumps(value, ensure_ascii=False, separators=(",", ":"))
def json_loads(value: str | None) -> Any:
if not value:
return None
return json.loads(value)