refactor: 全量迁云——fam-core 直读甲骨文 SQLite，NAS 只剩推送进程
前一天刚把登录迁到甲骨文，隔天 NAS 上的 fam-core 又挂了导致数据接口 502。 盘点后确认：甲骨文的 SQLite 才是权威数据源（videos 3113 / events 16159 / people 60 / model_calls 9876），NAS 的 MariaDB 全是它的镜像——前端读的数据 本来就产自甲骨文，绕了一圈回家又绕回来。 改动： - db_layer.py 从 725 行重写成 377 行：MySQL 镜像查询改为直读 fam-edge 的 SQLite。5 个 upsert_sync_*（约 300 行去重逻辑，8/29 和 9/3 两次 1062 事故的 发源地）连同 oracle_sync.py 整个删除。SQL 方言：JSON_CONTAINS -> json_each （前置 json_valid，历史脏数据不会把查询搞崩）、LEFT() -> substr()、%s -> ?。 函数名 get_sync_* 一并改掉——已经没有 sync 这回事了 - 新增 edge_client.py：写操作（改名/删除）、帧图头像、服务状态都打给同机 fam-edge，全走 127.0.0.1 - 新增 fam-notifier/：motion_notifier 从 fam-core 拆出独立成服务，游标从 MariaDB 换成本地 JSON 文件。NAS 上从此没有 Flask、没有数据库、没有监听端口 - fam-core 移到甲骨文 /opt/fam-core（systemd，gunicorn -w 2，只绑 127.0.0.1:5401——5400 被 chat-relay 占了）。Caddy 的 /api/* 从"frp 隧道 回源 NAS"改成同机反代，forward_auth 闸门不变 - 前端删掉侧边栏同步面板、统计页同步状态、服务状态页的"NAS 同步"卡片与 "立即同步"按钮（背后的镜像层已不存在）；换成"NAS 运动推送"卡片，读 fam-edge activity 新增的 motion 段（心跳年龄 + 最近事件） - 顺带修掉一个隐蔽 bug：镜像表为保外键稳定用的是 NAS 本地自增 id，而帧图接口 要的是甲骨文的 id，两边在 9/3 那次 id 重排后就对不上了。现在只有一套 id 测试：fam-core 21（新增 12 个 db_layer 用例：脏 JSON 不崩、人物精确匹配不误伤 "人物B"、日期过滤、统计口径、chat_history 懒建表）、fam-notifier 6、 fam-edge 157，全绿。 生产验证：甲骨文 /api/ui/stats 返回 videos 2965 / events 16159 / people 59； NAS 侧 fam-notifier 已推送成功（事件 33070-33072 落库，心跳新鲜）； chat_history 19 条经 scripts/import_chat_history.py 迁移完成。 Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This commit is contained in:
ericwyuan
@@ -1,10 +1,20 @@
|
||||
"""
|
||||
FAM-Core 主应用 - Flask 单进程（新架构 v2）
|
||||
FAM-Core 主应用 - Flask（2026-09-13 从 NAS 迁到甲骨文）
|
||||
|
||||
承载: Oracle-Sync（每 30 分钟拉取增量镜像）+ Member-Manager + Chat-Handler
|
||||
承载: UI-API + Member-Manager + Chat-Handler + Img-Proxy —— 全是查询与转发，
|
||||
**没有任何后台线程**，进程随时可重启，不持有任何状态。
|
||||
|
||||
NAS 不再处理视频：无 Scheduler / Dispatcher / Poller / Event-Receiver / Video-Server。
|
||||
所有视频分析在 Oracle 完成，NAS 仅作管理后台拉取展示，CPU 占用大幅降低。
|
||||
迁云前它跑在 NAS 上，还扛着两个后台线程，现在都不在这里了：
|
||||
- Oracle-Sync（每 30 分钟把甲骨文数据拉一份镜像进 MariaDB）—— 整个删除。
|
||||
本服务现在与 fam-edge 同机，直接读它的 SQLite（见 db_layer.py 开头）。
|
||||
- MotionNotifier（轮询 Surveillance Station 推运动事件）—— 留在 NAS，
|
||||
拆成独立的 fam-notifier（摄像头插在 NAS 上，这部分搬不走）。
|
||||
|
||||
两件跟安全有关的事：
|
||||
- 本服务只监听 127.0.0.1（config.yaml 的 server.host），唯一的客户端是同机
|
||||
Caddy。不像迁云前那样经 frp 把 :8000 暴露到公网。
|
||||
- 登录校验也不在这里：Caddy 用 forward_auth 打 fam-edge 的 /api/auth/verify，
|
||||
通过了才反代进来。
|
||||
"""
|
||||
import os
|
||||
import sys
|
||||
@@ -16,87 +26,52 @@ sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
|
||||
|
||||
from .config_loader import load_config
|
||||
from .logger import setup_logger
|
||||
from .oracle_sync import get_sync
|
||||
from .motion_notifier.motion_notifier import get_motion_notifier
|
||||
from .chat_handler.chat_handler import chat_bp
|
||||
from .member_manager.member_manager import member_bp
|
||||
from .img_proxy import img_bp
|
||||
from .motion_bp import motion_bp
|
||||
from .ui_api import ui_bp
|
||||
from .auth import auth_bp, init_auth
|
||||
from . import db_layer, edge_client
|
||||
|
||||
logger = setup_logger('fam-core.app')
|
||||
|
||||
app = Flask(__name__)
|
||||
|
||||
# 注册蓝图：全部 /api/* 路由（前端已迁云服务器 Caddy:80，NAS 不再托管 SPA 静态文件）
|
||||
app.register_blueprint(chat_bp)
|
||||
app.register_blueprint(member_bp)
|
||||
app.register_blueprint(img_bp)
|
||||
app.register_blueprint(motion_bp)
|
||||
app.register_blueprint(ui_bp)
|
||||
app.register_blueprint(auth_bp)
|
||||
|
||||
# 登录校验全局拦截（页面未登录 302 /login；/api/* 未登录 401）
|
||||
init_auth(app)
|
||||
|
||||
# 健康检查
|
||||
@app.route('/health', methods=['GET'])
|
||||
def health():
|
||||
return jsonify({"status": "ok", "service": "fam-core"}), 200
|
||||
|
||||
|
||||
# 初始化后台同步线程（NAS 唯一常驻线程）
|
||||
_sync = None
|
||||
try:
|
||||
_sync = get_sync()
|
||||
_sync.start()
|
||||
logger.info("Oracle-Sync 已启动")
|
||||
# 启动后立刻拉一次，前端无需等待首个周期即有数据
|
||||
try:
|
||||
_sync.trigger_now()
|
||||
logger.info("启动首次同步完成")
|
||||
except Exception as e:
|
||||
logger.warning(f"启动首次同步失败（后续周期会重试）: {e}")
|
||||
except Exception as e:
|
||||
logger.error(f"Oracle-Sync 启动失败: {e}")
|
||||
|
||||
# 初始化运动监测通知服务（NAS 轮询 SS EventCenter 主路径）
|
||||
_notifier = None
|
||||
try:
|
||||
_notifier = get_motion_notifier()
|
||||
_notifier.start()
|
||||
logger.info("MotionNotifier 已初始化")
|
||||
except Exception as e:
|
||||
logger.error(f"MotionNotifier 初始化失败: {e}")
|
||||
|
||||
|
||||
@app.route('/api/status', methods=['GET'])
|
||||
def status():
|
||||
"""系统状态"""
|
||||
"""服务自检：库能不能读、fam-edge 在不在。
|
||||
|
||||
迁云前这里报告的是镜像同步状态（游标/周期/上次增量条数），镜像没了之后
|
||||
那些字段不再存在，前端侧边栏的同步面板也一并去掉了。
|
||||
"""
|
||||
db_ok, db_err = True, None
|
||||
try:
|
||||
conn = db_layer.get_conn()
|
||||
try:
|
||||
conn.execute("SELECT 1 FROM videos LIMIT 1").fetchone()
|
||||
finally:
|
||||
conn.close()
|
||||
except Exception as e:
|
||||
db_ok, db_err = False, str(e)
|
||||
return jsonify({
|
||||
"service": "fam-core",
|
||||
"sync": _sync.status() if _sync else {"running": False, "error": "未初始化"},
|
||||
"motion": _notifier.status() if _notifier else {"running": False, "error": "未初始化"},
|
||||
"db": {"ok": db_ok, "error": db_err},
|
||||
"edge_base_url": edge_client.base_url(),
|
||||
}), 200
|
||||
|
||||
|
||||
@app.route('/api/sync/trigger', methods=['POST'])
|
||||
def sync_trigger():
|
||||
"""手动立即触发一次甲骨文增量同步（服务状态页"立即同步"按钮）。
|
||||
|
||||
正常情况下后台线程每 30 分钟自动拉一次；这个接口给用户想立刻看到最新数据
|
||||
时用，跟后台线程共用同一把拉取锁（oracle_sync._pull_lock），不会并发重复拉。
|
||||
"""
|
||||
if not _sync:
|
||||
return jsonify({"error": "同步服务未初始化"}), 503
|
||||
ok = _sync.trigger_now()
|
||||
if not ok:
|
||||
return jsonify({"status": "failed", "error": _sync.status().get("last_error")}), 502
|
||||
return jsonify({"status": "ok", **_sync.status()}), 200
|
||||
|
||||
|
||||
if __name__ == '__main__':
|
||||
cfg = load_config()
|
||||
port = cfg.get('server', {}).get('port', 8000)
|
||||
app.run(host='0.0.0.0', port=port, debug=False)
|
||||
server = cfg.get('server', {})
|
||||
app.run(host=server.get('host', '127.0.0.1'),
|
||||
port=server.get('port', 5401), debug=False)
|
||||
|
||||
@@ -47,7 +47,7 @@ def _call_edge_qa(prompt: str) -> str:
|
||||
"""调用 FAM-Edge 问答编排端点（Gemini → NVIDIA → 本地 Ollama 兜底）"""
|
||||
cfg = load_config()
|
||||
qa_url = cfg.get('chat_handler', {}).get(
|
||||
'qa_url', 'http://129.146.26.249:5000/api/edge/chat/ask'
|
||||
'qa_url', 'http://127.0.0.1:5000/api/edge/chat/ask'
|
||||
)
|
||||
timeout = cfg.get('chat_handler', {}).get('timeout', 120)
|
||||
|
||||
@@ -79,7 +79,7 @@ def chat_ask():
|
||||
|
||||
logger.info(f"Chat: person={queried_person}, date={queried_date}, question={question}")
|
||||
|
||||
rows = db_layer.query_sync_events_for_person_date(queried_person, queried_date)
|
||||
rows = db_layer.query_events_for_person_date(queried_person, queried_date)
|
||||
|
||||
if len(rows) == 0:
|
||||
answer = f"今天没有观察到{queried_person}。"
|
||||
@@ -87,7 +87,7 @@ def chat_ask():
|
||||
else:
|
||||
context = _format_events(rows)
|
||||
context_summary = f"查询 sync_events {len(rows)} 条"
|
||||
members = db_layer.get_sync_known_members_context()
|
||||
members = db_layer.get_known_members_context()
|
||||
prompt = build_chat_prompt(
|
||||
context=context,
|
||||
members=members or queried_person,
|
||||
@@ -134,7 +134,7 @@ def chat_ask_stream():
|
||||
return jsonify({"error": "缺少必填字段: question, queried_person, queried_date"}), 400
|
||||
|
||||
logger.info(f"Chat(stream): person={queried_person}, date={queried_date}, question={question}")
|
||||
rows = db_layer.query_sync_events_for_person_date(queried_person, queried_date)
|
||||
rows = db_layer.query_events_for_person_date(queried_person, queried_date)
|
||||
|
||||
def sse(obj):
|
||||
return f"data: {json.dumps(obj, ensure_ascii=False)}\n\n"
|
||||
@@ -157,14 +157,14 @@ def chat_ask_stream():
|
||||
yield sse({"type": "context", "count": len(rows), "summary": context_summary,
|
||||
"preview": context[:800]})
|
||||
|
||||
members = db_layer.get_sync_known_members_context()
|
||||
members = db_layer.get_known_members_context()
|
||||
prompt = build_chat_prompt(
|
||||
context=context, members=members or queried_person,
|
||||
question=question, queried_person=queried_person)
|
||||
|
||||
cfg = load_config()
|
||||
stream_url = cfg.get('chat_handler', {}).get(
|
||||
'qa_stream_url', 'http://129.146.26.249:5000/api/edge/chat/ask/stream')
|
||||
'qa_stream_url', 'http://127.0.0.1:5000/api/edge/chat/ask/stream')
|
||||
timeout = cfg.get('chat_handler', {}).get('timeout', 120)
|
||||
|
||||
full_answer = []
|
||||
|
||||
@@ -1,11 +1,42 @@
|
||||
"""
|
||||
配置加载器 - 从 config.yaml 读取配置
|
||||
配置加载器 - 从 config.yaml 读取配置，支持 ${ENV_VAR} 解析
|
||||
"""
|
||||
import os
|
||||
import re
|
||||
import yaml
|
||||
|
||||
|
||||
def _load_env_file():
|
||||
"""加载部署目录下的 .env（支持 export KEY=VALUE 格式）。
|
||||
|
||||
迁云前是 start_core.sh 负责 source .env 再起 gunicorn；现在由 systemd 拉起，
|
||||
没有那一步，所以在这里兜底加载（跟 fam-edge 的做法一致）。
|
||||
已存在的环境变量不覆盖。
|
||||
"""
|
||||
path = os.environ.get('FAM_ENV_FILE') or os.path.join(
|
||||
os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__)))),
|
||||
'.env')
|
||||
if not os.path.isfile(path):
|
||||
return
|
||||
with open(path, 'r', encoding='utf-8') as f:
|
||||
for line in f:
|
||||
line = line.strip()
|
||||
if not line or line.startswith('#'):
|
||||
continue
|
||||
if line.startswith('export '):
|
||||
line = line[7:].strip()
|
||||
if '=' not in line:
|
||||
continue
|
||||
key, _, value = line.partition('=')
|
||||
key = key.strip()
|
||||
value = value.strip().strip('"').strip("'")
|
||||
if key and key not in os.environ:
|
||||
os.environ[key] = value
|
||||
|
||||
|
||||
_load_env_file()
|
||||
|
||||
|
||||
def _resolve_env_vars(value):
|
||||
"""递归解析字符串中的 ${ENV_VAR} 引用"""
|
||||
if isinstance(value, str):
|
||||
|
||||
@@ -1,300 +1,176 @@
|
||||
"""
|
||||
数据库访问层 - MariaDB 连接管理与同步镜像 CRUD
|
||||
DB-Layer - 直读 fam-edge 的 SQLite 库（2026-09-13 迁云重写）
|
||||
|
||||
新架构 (2026-08-21 重构):
|
||||
NAS 不再处理视频，仅作为管理后台。
|
||||
Oracle (FAM-Edge) 处理整视频分析后存 SQLite，NAS 每 30 分钟拉增量，
|
||||
镜像到本地三张表:
|
||||
sync_videos : 视频会话（全局摘要 + 事件 JSON + 人物 JSON）
|
||||
sync_events : 视频拆出的时间点事件（描述 + 涉及人物 + 是否关注）
|
||||
sync_people : 规范人物表（label + canonical_name，Oracle 维护）
|
||||
sync_cursor : 同步游标（上次成功拉取到的 server_time）
|
||||
**这个文件以前是什么样**：fam-core 跑在 NAS 上，每 30 分钟把甲骨文的数据拉一份
|
||||
镜像进 MariaDB（`sync_videos`/`sync_events`/…），前端读镜像。725 行里有一半是
|
||||
镜像 upsert 的去重逻辑——8/29 和 9/3 两次线上事故（1062 主键冲突、游标卡死）
|
||||
都出在那一半。
|
||||
|
||||
本层只服务同步镜像 + 问答历史，旧 process_tasks/event_details/monitor_events/
|
||||
family_members 相关逻辑已全部移除（视频处理职责已迁移至 Oracle）。
|
||||
**现在**：fam-core 跟 fam-edge 同机，直接打开它的 SQLite 库读，镜像层整个不存在了。
|
||||
连带消失的还有一个隐蔽 bug：镜像表为了保外键稳定用的是 NAS 本地自增 id，而帧图
|
||||
接口要的是甲骨文的 id，两边在 9/3 那次 id 重排后就对不上了。现在只有一套 id。
|
||||
|
||||
要点：
|
||||
- 库文件是 fam-edge 的（`database.path`），**本模块只读它写的表**，唯一写的表是
|
||||
`chat_history`（问答历史，迁云时从 NAS MariaDB 搬过来的，fam-edge 不碰）。
|
||||
- fam-edge 那边开了 WAL，读不会阻塞它的写；这边同样设 busy_timeout 兜底。
|
||||
- 视频/人物的写操作（改名、删除）不在这里做，走 `edge_client` 打给 fam-edge——
|
||||
它除了改库还要合并人物、删磁盘素材。
|
||||
"""
|
||||
import json
|
||||
import pymysql
|
||||
from datetime import datetime
|
||||
from typing import Optional, List, Dict, Any
|
||||
import sqlite3
|
||||
from datetime import datetime, timedelta, timezone
|
||||
from typing import Dict, List, Optional
|
||||
|
||||
from .config_loader import load_config
|
||||
from .logger import setup_logger
|
||||
|
||||
logger = setup_logger('fam-core.db')
|
||||
|
||||
_config = None
|
||||
_db_path = None
|
||||
_chat_schema_ready = False
|
||||
|
||||
|
||||
def get_config():
|
||||
global _config
|
||||
if _config is None:
|
||||
_config = load_config()
|
||||
return _config
|
||||
def _now() -> str:
|
||||
return datetime.now(timezone(timedelta(hours=8))).strftime('%Y-%m-%d %H:%M:%S')
|
||||
|
||||
|
||||
def get_conn():
|
||||
"""获取数据库连接（单 worker gunicorn，无需连接池）"""
|
||||
cfg = get_config().get('database', {})
|
||||
kwargs = dict(
|
||||
host=cfg.get('host', '127.0.0.1'),
|
||||
port=cfg.get('port', 3306),
|
||||
user=cfg.get('user', 'root'),
|
||||
password=cfg.get('password', ''),
|
||||
database=cfg.get('database', 'sentinel_home_ai'),
|
||||
charset='utf8mb4',
|
||||
autocommit=False
|
||||
)
|
||||
unix_socket = cfg.get('unix_socket')
|
||||
if unix_socket:
|
||||
kwargs['unix_socket'] = unix_socket
|
||||
return pymysql.connect(**kwargs)
|
||||
def get_config() -> Dict:
|
||||
return load_config()
|
||||
|
||||
|
||||
def _path() -> str:
|
||||
global _db_path
|
||||
if _db_path is None:
|
||||
_db_path = load_config().get('database', {}).get(
|
||||
'path', '/opt/fam-edge/data/oracle.db')
|
||||
return _db_path
|
||||
|
||||
|
||||
def get_conn() -> sqlite3.Connection:
|
||||
"""每次调用开一个连接（请求级，跟改写前的 MySQL 用法一致）。
|
||||
|
||||
busy_timeout：fam-edge 的写事务提交时会短暂持锁，这里等而不是立刻报
|
||||
`database is locked`（9/3 那次 Oracle 过载时刷过一片这个错）。
|
||||
"""
|
||||
conn = sqlite3.connect(_path(), timeout=10)
|
||||
conn.row_factory = sqlite3.Row
|
||||
conn.execute("PRAGMA busy_timeout=10000")
|
||||
return conn
|
||||
|
||||
|
||||
def _rows(cur) -> List[Dict]:
|
||||
return [dict(r) for r in cur.fetchall()]
|
||||
|
||||
|
||||
def _row(cur) -> Optional[Dict]:
|
||||
r = cur.fetchone()
|
||||
return dict(r) if r else None
|
||||
|
||||
|
||||
# 人物命中判断：person_list_json 是 JSON 数组文本，用 json_each 展开精确匹配。
|
||||
# 必须先 json_valid——历史数据里有非 JSON 的脏值，直接 json_each 会整条查询报错。
|
||||
_PERSON_HIT = ("{col} IS NOT NULL AND json_valid({col}) "
|
||||
"AND EXISTS(SELECT 1 FROM json_each({col}) WHERE json_each.value = ?)")
|
||||
|
||||
# 日期维度：录制时间优先（文件名解析出来的），回退分析时间
|
||||
_DATE_EXPR = "COALESCE(NULLIF({a}.event_start_time,''), {a}.processed_at, {a}.updated_at, {a}.created_at)"
|
||||
|
||||
# 只展示"有内容"的会话：运动片段，或含事件的视频。
|
||||
# 整段素材分割 0 段的空会话（历史素材无运动事件）不展示，避免淹没时间轴。
|
||||
_CONTENT_FILTER = ("substr(v.filename, 1, 7) = 'motion_' "
|
||||
"OR EXISTS(SELECT 1 FROM events se WHERE se.video_id = v.id)")
|
||||
|
||||
|
||||
# ============================================================
|
||||
# 同步镜像：sync_videos
|
||||
# videos / events
|
||||
# ============================================================
|
||||
|
||||
def upsert_sync_videos(rows: List[Dict]) -> int:
|
||||
"""批量 upsert Oracle 传来的 videos 增量。rows 为 Oracle 端 dict 列表。
|
||||
|
||||
v2 修复 (2026-09-03)：不再以 Oracle 的 videos.id 作为 NAS 主键。Oracle 库重建/
|
||||
重排后 id 会复用/跳变（实测 2303 段跳到 3408/4600+ 段），旧写法按 id 插入时与
|
||||
已存在的同 filename 行在 UNIQUE filename 上二次冲突报 1062，整批中止、游标不推进。
|
||||
现改为：以 filename 为业务唯一键去重，命中则就地 UPDATE（保留 NAS 原 id，防止
|
||||
sync_events/model_calls/identity_map 的 video_id 外键失效），Oracle id 仅落
|
||||
oracle_id 列溯源；未命中则插入（优先用 Oracle id 作主键以对齐子表引用，主键冲突时
|
||||
回退本地自增，避免 1062）。
|
||||
"""
|
||||
if not rows:
|
||||
return 0
|
||||
conn = get_conn()
|
||||
n = 0
|
||||
try:
|
||||
cur = conn.cursor()
|
||||
for r in rows:
|
||||
fn = r.get('filename')
|
||||
oracle_id = r.get('id')
|
||||
cur.execute("SELECT id FROM sync_videos WHERE filename=%s", (fn,))
|
||||
row = cur.fetchone()
|
||||
if row:
|
||||
# 命中业务键：就地更新，保留原 NAS id（外键不失效）
|
||||
cur.execute(
|
||||
"""UPDATE sync_videos SET
|
||||
oracle_id=%s, drive_file_id=%s, camera_name=%s,
|
||||
duration_sec=%s, event_start_time=%s, status=%s,
|
||||
summary_json=%s, events_json=%s, people_json=%s,
|
||||
compute_provider=%s, created_at=%s, updated_at=%s,
|
||||
processed_at=%s, synced_at=NOW()
|
||||
WHERE id=%s""",
|
||||
(oracle_id, r.get('drive_file_id'), r.get('camera_name'),
|
||||
r.get('duration_sec') or 0, r.get('event_start_time'),
|
||||
r.get('status'), r.get('summary_json'), r.get('events_json'),
|
||||
r.get('people_json'), r.get('compute_provider'),
|
||||
r.get('created_at'), r.get('updated_at'),
|
||||
r.get('processed_at'), row[0]))
|
||||
else:
|
||||
# 未命中：新文件。优先用 Oracle id 作主键（对齐子表 video_id 引用）
|
||||
try:
|
||||
cur.execute(
|
||||
"""INSERT INTO sync_videos
|
||||
(id, oracle_id, filename, camera_name, duration_sec,
|
||||
event_start_time, status, summary_json, events_json,
|
||||
people_json, compute_provider, created_at, updated_at,
|
||||
processed_at, synced_at)
|
||||
VALUES (%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s, NOW())""",
|
||||
(oracle_id, oracle_id, fn, r.get('camera_name'),
|
||||
r.get('duration_sec') or 0, r.get('event_start_time'),
|
||||
r.get('status'), r.get('summary_json'), r.get('events_json'),
|
||||
r.get('people_json'), r.get('compute_provider'),
|
||||
r.get('created_at'), r.get('updated_at'),
|
||||
r.get('processed_at')))
|
||||
except Exception:
|
||||
# 主键 id 被复用（极端情况）：回退本地自增，避免 1062
|
||||
cur.execute(
|
||||
"""INSERT INTO sync_videos
|
||||
(oracle_id, filename, camera_name, duration_sec,
|
||||
event_start_time, status, summary_json, events_json,
|
||||
people_json, compute_provider, created_at, updated_at,
|
||||
processed_at, synced_at)
|
||||
VALUES (%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s, NOW())""",
|
||||
(oracle_id, fn, r.get('camera_name'),
|
||||
r.get('duration_sec') or 0, r.get('event_start_time'),
|
||||
r.get('status'), r.get('summary_json'), r.get('events_json'),
|
||||
r.get('people_json'), r.get('compute_provider'),
|
||||
r.get('created_at'), r.get('updated_at'),
|
||||
r.get('processed_at')))
|
||||
n += 1
|
||||
conn.commit()
|
||||
return n
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
def get_sync_videos(limit=15, offset=0, date_filter=None) -> List[Dict]:
|
||||
"""获取视频会话列表（已完成优先），支持日期筛选与分页。
|
||||
|
||||
排序/日期维度按视频实际录制时间（event_start_time，文件名解析），
|
||||
为空回退 processed_at/updated_at/created_at。
|
||||
date_filter 形如 '2026-08-21'。
|
||||
只返回"有内容"的会话：运动片段（filename 前缀 motion_）或含事件的视频；
|
||||
整段素材分割 0 段的空会话（历史素材无运动事件）不展示，避免淹没时间轴。
|
||||
"""
|
||||
def get_videos(limit=15, offset=0, date_filter=None) -> List[Dict]:
|
||||
"""视频会话列表（事件时间轴左侧），按录制时间倒序，支持日期筛选与分页。"""
|
||||
d = _DATE_EXPR.format(a='v')
|
||||
sql = f"""SELECT v.id, v.filename, v.camera_name, v.event_start_time, v.status,
|
||||
v.summary_json, v.events_json, v.people_json, v.compute_provider,
|
||||
v.processed_at, v.updated_at,
|
||||
(SELECT COUNT(*) FROM events se WHERE se.video_id = v.id) AS event_count
|
||||
FROM videos v
|
||||
WHERE v.status='done' AND ({_CONTENT_FILTER}) {{extra}}
|
||||
ORDER BY {d} DESC
|
||||
LIMIT ? OFFSET ?"""
|
||||
conn = get_conn()
|
||||
try:
|
||||
cur = conn.cursor(pymysql.cursors.DictCursor)
|
||||
# 录制时间优先，回退分析时间
|
||||
date_expr = "COALESCE(NULLIF(event_start_time,''), processed_at, updated_at, created_at)"
|
||||
# 注意：pymysql execute 用 % 做参数占位符，SQL 字面量不能含 %，故用 LEFT 判断 motion_ 前缀
|
||||
content_filter = ("LEFT(v.filename, 7) = 'motion_' "
|
||||
"OR EXISTS(SELECT 1 FROM sync_events se WHERE se.video_id = v.id)")
|
||||
if date_filter:
|
||||
cur.execute(
|
||||
f"""SELECT v.id, v.filename, v.camera_name, v.event_start_time, v.status,
|
||||
v.summary_json, v.events_json, v.people_json, v.compute_provider,
|
||||
v.processed_at, v.updated_at,
|
||||
(SELECT COUNT(*) FROM sync_events se WHERE se.video_id = v.id) AS event_count
|
||||
FROM sync_videos v
|
||||
WHERE v.status='done' AND ({content_filter}) AND {date_expr} LIKE %s
|
||||
ORDER BY {date_expr} DESC
|
||||
LIMIT %s OFFSET %s""",
|
||||
(f'{date_filter}%', limit, offset))
|
||||
cur = conn.execute(sql.format(extra=f"AND {d} LIKE ?"),
|
||||
(f'{date_filter}%', limit, offset))
|
||||
else:
|
||||
cur.execute(
|
||||
f"""SELECT v.id, v.filename, v.camera_name, v.event_start_time, v.status,
|
||||
v.summary_json, v.events_json, v.people_json, v.compute_provider,
|
||||
v.processed_at, v.updated_at,
|
||||
(SELECT COUNT(*) FROM sync_events se WHERE se.video_id = v.id) AS event_count
|
||||
FROM sync_videos v
|
||||
WHERE v.status='done' AND ({content_filter})
|
||||
ORDER BY {date_expr} DESC
|
||||
LIMIT %s OFFSET %s""",
|
||||
(limit, offset))
|
||||
return cur.fetchall()
|
||||
cur = conn.execute(sql.format(extra=''), (limit, offset))
|
||||
return _rows(cur)
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
def get_sync_video(video_id: int) -> Optional[Dict]:
|
||||
def get_video(video_id: int) -> Optional[Dict]:
|
||||
conn = get_conn()
|
||||
try:
|
||||
cur = conn.cursor(pymysql.cursors.DictCursor)
|
||||
cur.execute("SELECT * FROM sync_videos WHERE id=%s", (video_id,))
|
||||
return cur.fetchone()
|
||||
return _row(conn.execute("SELECT * FROM videos WHERE id=?", (video_id,)))
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
def delete_sync_video(video_id: int):
|
||||
"""删除本地镜像里的这条视频（events 先删，再删 videos）。
|
||||
|
||||
增量同步（get_sync_delta）只做 upsert，感知不到 Oracle 那边的物理删除，
|
||||
所以删除操作不能走"转发 Oracle + trigger_now() 拉增量"这条老路，必须在
|
||||
Oracle 确认删除成功后由调用方显式清理本地镜像。
|
||||
"""
|
||||
def get_events_for_video(video_id: int) -> List[Dict]:
|
||||
conn = get_conn()
|
||||
try:
|
||||
cur = conn.cursor()
|
||||
cur.execute("DELETE FROM sync_events WHERE video_id=%s", (video_id,))
|
||||
cur.execute("DELETE FROM sync_videos WHERE id=%s", (video_id,))
|
||||
conn.commit()
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
# ============================================================
|
||||
# 同步镜像：sync_events
|
||||
# ============================================================
|
||||
|
||||
def upsert_sync_events(rows: List[Dict]) -> int:
|
||||
"""批量 upsert Oracle 传来的 events 增量。"""
|
||||
if not rows:
|
||||
return 0
|
||||
conn = get_conn()
|
||||
n = 0
|
||||
try:
|
||||
cur = conn.cursor()
|
||||
for r in rows:
|
||||
cur.execute(
|
||||
"""INSERT INTO sync_events
|
||||
(id, video_id, ts, description, person_list_json,
|
||||
person_appearances_json, is_attention_event,
|
||||
updated_at, synced_at)
|
||||
VALUES (%s,%s,%s,%s,%s,%s,%s,%s, NOW())
|
||||
ON DUPLICATE KEY UPDATE
|
||||
video_id=VALUES(video_id),
|
||||
ts=VALUES(ts),
|
||||
description=VALUES(description),
|
||||
person_list_json=VALUES(person_list_json),
|
||||
person_appearances_json=VALUES(person_appearances_json),
|
||||
is_attention_event=VALUES(is_attention_event),
|
||||
updated_at=VALUES(updated_at),
|
||||
synced_at=NOW()""",
|
||||
(r.get('id'), r.get('video_id'), r.get('ts'), r.get('description'),
|
||||
r.get('person_list_json'), r.get('person_appearances_json'),
|
||||
1 if r.get('is_attention_event') else 0,
|
||||
r.get('updated_at'))
|
||||
)
|
||||
n += 1
|
||||
conn.commit()
|
||||
return n
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
def get_sync_events_for_video(video_id: int) -> List[Dict]:
|
||||
conn = get_conn()
|
||||
try:
|
||||
cur = conn.cursor(pymysql.cursors.DictCursor)
|
||||
cur.execute(
|
||||
return _rows(conn.execute(
|
||||
"""SELECT id, video_id, ts, description, person_list_json,
|
||||
is_attention_event, updated_at
|
||||
FROM sync_events WHERE video_id=%s ORDER BY ts ASC""",
|
||||
(video_id,))
|
||||
return cur.fetchall()
|
||||
is_attention_event
|
||||
FROM events WHERE video_id=? ORDER BY ts ASC""", (video_id,)))
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
def get_sync_people_clips(label: str, limit: int = 10) -> List[Dict]:
|
||||
"""某人物（label 或 canonical_name）出现过的运动片段列表。
|
||||
|
||||
匹配：label 本身 + 其 canonical_name 下全部 label；命中 sync_events 的
|
||||
person_list_json（JSON_CONTAINS）→ 关联 video（优先运动片段 motion_）。
|
||||
返回按 event_start_time 倒序，含 first_ts（该人物在片段内最早事件 ts，
|
||||
前端缩略图定位用）与 clip_events（片段内该人物事件数）。
|
||||
"""
|
||||
def get_attention_events(limit: int = 200) -> List[Dict]:
|
||||
"""需关注事件（统计图表页用），按录制日期倒序。"""
|
||||
conn = get_conn()
|
||||
try:
|
||||
return _rows(conn.execute(
|
||||
"""SELECT COALESCE(NULLIF(v.event_start_time,''), v.processed_at) AS ev_date,
|
||||
e.person_list_json
|
||||
FROM events e JOIN videos v ON e.video_id=v.id
|
||||
WHERE e.is_attention_event = 1
|
||||
ORDER BY ev_date DESC LIMIT ?""", (limit,)))
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
def get_people_clips(label: str, limit: int = 10) -> List[Dict]:
|
||||
"""某人物出现过的运动片段列表。
|
||||
|
||||
匹配 label 本身 + 同一 canonical_name 下的全部 label；返回含 first_ts
|
||||
（该人物在片段内最早事件时间点，前端缩略图定位用）与 clip_events。
|
||||
"""
|
||||
hit_e = _PERSON_HIT.format(col='e.person_list_json')
|
||||
hit_e2 = _PERSON_HIT.format(col='e2.person_list_json')
|
||||
hit_e3 = _PERSON_HIT.format(col='e3.person_list_json')
|
||||
conn = get_conn()
|
||||
try:
|
||||
cur = conn.cursor(pymysql.cursors.DictCursor)
|
||||
labels = {label}
|
||||
cur.execute(
|
||||
"SELECT label, canonical_name FROM sync_people WHERE label=%s OR canonical_name=%s",
|
||||
(label, label))
|
||||
for r in cur.fetchall():
|
||||
cn = r.get('canonical_name')
|
||||
if cn:
|
||||
cur.execute("SELECT label FROM sync_people WHERE canonical_name=%s", (cn,))
|
||||
labels.update(x['label'] for x in cur.fetchall())
|
||||
clips = []
|
||||
seen = set()
|
||||
for r in _rows(conn.execute(
|
||||
"SELECT label, canonical_name FROM people WHERE label=? OR canonical_name=?",
|
||||
(label, label))):
|
||||
if r.get('canonical_name'):
|
||||
labels.update(x['label'] for x in _rows(conn.execute(
|
||||
"SELECT label FROM people WHERE canonical_name=?", (r['canonical_name'],))))
|
||||
clips, seen = [], set()
|
||||
for lb in sorted(labels):
|
||||
cur.execute(
|
||||
"""SELECT DISTINCT v.id AS video_id, v.filename, v.event_start_time,
|
||||
v.duration_sec, v.summary_json, v.camera_name,
|
||||
(SELECT MIN(e2.ts) FROM sync_events e2
|
||||
WHERE e2.video_id=v.id AND e2.person_list_json IS NOT NULL
|
||||
AND JSON_CONTAINS(e2.person_list_json, JSON_QUOTE(%s), '$')) AS first_ts,
|
||||
(SELECT COUNT(*) FROM sync_events e3
|
||||
WHERE e3.video_id=v.id AND e3.person_list_json IS NOT NULL
|
||||
AND JSON_CONTAINS(e3.person_list_json, JSON_QUOTE(%s), '$')) AS clip_events
|
||||
FROM sync_events e
|
||||
JOIN sync_videos v ON e.video_id=v.id
|
||||
WHERE e.person_list_json IS NOT NULL
|
||||
AND JSON_CONTAINS(e.person_list_json, JSON_QUOTE(%s), '$')
|
||||
AND v.status='done'""",
|
||||
cur = conn.execute(
|
||||
f"""SELECT DISTINCT v.id AS video_id, v.filename, v.event_start_time,
|
||||
v.duration_sec, v.summary_json, v.camera_name,
|
||||
(SELECT MIN(e2.ts) FROM events e2
|
||||
WHERE e2.video_id=v.id AND {hit_e2}) AS first_ts,
|
||||
(SELECT COUNT(*) FROM events e3
|
||||
WHERE e3.video_id=v.id AND {hit_e3}) AS clip_events
|
||||
FROM events e JOIN videos v ON e.video_id=v.id
|
||||
WHERE {hit_e} AND v.status='done'""",
|
||||
(lb, lb, lb))
|
||||
for row in cur.fetchall():
|
||||
for row in _rows(cur):
|
||||
if row['video_id'] not in seen:
|
||||
seen.add(row['video_id'])
|
||||
clips.append(row)
|
||||
@@ -304,397 +180,177 @@ def get_sync_people_clips(label: str, limit: int = 10) -> List[Dict]:
|
||||
conn.close()
|
||||
|
||||
|
||||
def query_sync_events_for_person_date(person: str, date_str: str) -> List[Dict]:
|
||||
def query_events_for_person_date(person: str, date_str: str) -> List[Dict]:
|
||||
"""问答上下文：某人在某天的事件。
|
||||
|
||||
说明: Oracle 事件 ts 为视频内相对时间点（如 00:01:23），不是绝对日期，
|
||||
因此按所属视频的录制日期（event_start_time，回退 processed_at）过滤，
|
||||
再按 person_list_json 命中人名。
|
||||
person 可为真名或抽象标签（Oracle 回灌上下文用真名，但历史标签也保留）。
|
||||
事件 ts 是视频内的相对时间点（如 00:01:23），不是绝对日期，所以按所属视频的
|
||||
录制日期过滤，再按 person_list_json 命中人名。person 可为真名或抽象标签。
|
||||
"""
|
||||
hit = _PERSON_HIT.format(col='e.person_list_json')
|
||||
conn = get_conn()
|
||||
try:
|
||||
cur = conn.cursor(pymysql.cursors.DictCursor)
|
||||
cur.execute(
|
||||
"""SELECT e.ts, e.description, e.person_list_json, e.is_attention_event,
|
||||
v.camera_name, v.filename, v.event_start_time, v.processed_at
|
||||
FROM sync_events e
|
||||
JOIN sync_videos v ON e.video_id = v.id
|
||||
WHERE COALESCE(NULLIF(v.event_start_time,''), v.processed_at) LIKE %s
|
||||
AND e.person_list_json IS NOT NULL
|
||||
AND JSON_CONTAINS(e.person_list_json, JSON_QUOTE(%s), '$')
|
||||
ORDER BY COALESCE(NULLIF(v.event_start_time,''), v.processed_at) ASC, e.ts ASC""",
|
||||
(f'{date_str}%', person))
|
||||
return cur.fetchall()
|
||||
return _rows(conn.execute(
|
||||
f"""SELECT e.ts, e.description, e.person_list_json, e.is_attention_event,
|
||||
v.camera_name, v.filename, v.event_start_time, v.processed_at
|
||||
FROM events e JOIN videos v ON e.video_id = v.id
|
||||
WHERE COALESCE(NULLIF(v.event_start_time,''), v.processed_at) LIKE ?
|
||||
AND {hit}
|
||||
ORDER BY COALESCE(NULLIF(v.event_start_time,''), v.processed_at) ASC, e.ts ASC""",
|
||||
(f'{date_str}%', person)))
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
# ============================================================
|
||||
# 同步镜像：sync_people
|
||||
# people
|
||||
# ============================================================
|
||||
|
||||
def upsert_sync_people(rows: List[Dict]) -> int:
|
||||
"""批量 upsert Oracle 传来的 people 增量。
|
||||
|
||||
v2 修复 (2026-09-03)：同 videos，以 label 为业务唯一键去重，命中就地 UPDATE
|
||||
保留 NAS 原 id（people.id 无外键引用，仅作展示排序），Oracle id 落 oracle_id 列
|
||||
溯源；未命中插入（优先 Oracle id，主键冲突回退自增，避免 1062）。
|
||||
"""
|
||||
if not rows:
|
||||
return 0
|
||||
conn = get_conn()
|
||||
n = 0
|
||||
try:
|
||||
cur = conn.cursor()
|
||||
for r in rows:
|
||||
label = r.get('label')
|
||||
oracle_id = r.get('id')
|
||||
cur.execute("SELECT id FROM sync_people WHERE label=%s", (label,))
|
||||
row = cur.fetchone()
|
||||
if row:
|
||||
cur.execute(
|
||||
"""UPDATE sync_people SET
|
||||
oracle_id=%s, canonical_name=%s, first_seen=%s,
|
||||
appearances=%s, source=%s, features_json=%s,
|
||||
display_uid=%s, updated_at=%s, synced_at=NOW()
|
||||
WHERE id=%s""",
|
||||
(oracle_id, r.get('canonical_name'), r.get('first_seen'),
|
||||
r.get('appearances') or 0, r.get('source'),
|
||||
r.get('features_json'), r.get('display_uid'),
|
||||
r.get('updated_at'), row[0]))
|
||||
else:
|
||||
try:
|
||||
cur.execute(
|
||||
"""INSERT INTO sync_people
|
||||
(id, oracle_id, label, canonical_name, first_seen,
|
||||
appearances, source, features_json, display_uid,
|
||||
updated_at, synced_at)
|
||||
VALUES (%s,%s,%s,%s,%s,%s,%s,%s,%s,%s, NOW())""",
|
||||
(oracle_id, oracle_id, label, r.get('canonical_name'),
|
||||
r.get('first_seen'), r.get('appearances') or 0,
|
||||
r.get('source'), r.get('features_json'),
|
||||
r.get('display_uid'), r.get('updated_at')))
|
||||
except Exception:
|
||||
cur.execute(
|
||||
"""INSERT INTO sync_people
|
||||
(oracle_id, label, canonical_name, first_seen,
|
||||
appearances, source, features_json, display_uid,
|
||||
updated_at, synced_at)
|
||||
VALUES (%s,%s,%s,%s,%s,%s,%s,%s,%s, NOW())""",
|
||||
(oracle_id, label, r.get('canonical_name'),
|
||||
r.get('first_seen'), r.get('appearances') or 0,
|
||||
r.get('source'), r.get('features_json'),
|
||||
r.get('display_uid'), r.get('updated_at')))
|
||||
n += 1
|
||||
conn.commit()
|
||||
return n
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
def get_sync_people() -> List[Dict]:
|
||||
def get_people() -> List[Dict]:
|
||||
conn = get_conn()
|
||||
try:
|
||||
cur = conn.cursor(pymysql.cursors.DictCursor)
|
||||
cur.execute(
|
||||
return _rows(conn.execute(
|
||||
"SELECT id, label, canonical_name, first_seen, appearances, source, "
|
||||
"features_json, display_uid, updated_at "
|
||||
"FROM sync_people ORDER BY id ASC")
|
||||
return cur.fetchall()
|
||||
"features_json, display_uid, updated_at FROM people ORDER BY id ASC"))
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
def upsert_sync_model_calls(rows: List[Dict]) -> int:
|
||||
"""批量 upsert Oracle 传来的 model_calls 增量（幂等，重复覆盖）。"""
|
||||
if not rows:
|
||||
return 0
|
||||
def get_named_members() -> List[str]:
|
||||
"""已命名成员的真名列表（UI 下拉用）。"""
|
||||
conn = get_conn()
|
||||
n = 0
|
||||
try:
|
||||
cur = conn.cursor()
|
||||
for r in rows:
|
||||
cur.execute(
|
||||
"""INSERT INTO sync_model_calls
|
||||
(id, provider, model, video_id, filename, started_at,
|
||||
duration_sec, success, error, created_at, synced_at)
|
||||
VALUES (%s,%s,%s,%s,%s,%s,%s,%s,%s,%s, NOW())
|
||||
ON DUPLICATE KEY UPDATE
|
||||
provider=VALUES(provider),
|
||||
model=VALUES(model),
|
||||
video_id=VALUES(video_id),
|
||||
filename=VALUES(filename),
|
||||
started_at=VALUES(started_at),
|
||||
duration_sec=VALUES(duration_sec),
|
||||
success=VALUES(success),
|
||||
error=VALUES(error),
|
||||
created_at=VALUES(created_at),
|
||||
synced_at=NOW()""",
|
||||
(r.get('id'), r.get('provider'), r.get('model'),
|
||||
r.get('video_id'), r.get('filename'), r.get('started_at'),
|
||||
r.get('duration_sec') or 0, 1 if r.get('success') else 0,
|
||||
(r.get('error') or '')[:500], r.get('created_at')))
|
||||
n += 1
|
||||
conn.commit()
|
||||
return n
|
||||
return [r['canonical_name'] for r in _rows(conn.execute(
|
||||
"SELECT DISTINCT canonical_name FROM people "
|
||||
"WHERE canonical_name IS NOT NULL AND canonical_name != '' "
|
||||
"ORDER BY canonical_name ASC"))]
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
def upsert_sync_identity_map(rows: List[Dict]) -> int:
|
||||
"""批量 upsert Oracle 传来的 person_identity_map 增量（幂等，重复覆盖）。
|
||||
|
||||
v2 修复 (2026-08-29)：不再以 Oracle 的 person_identity_map.id 作为 NAS 主键。
|
||||
该 id 只是 Oracle 端代理主键，一旦 Oracle 库重建/恢复，id 会被复用（实测出现
|
||||
id=744 同时对应 NAS 上的 (1302,'人物E') 与 Oracle 现在的 (1073,'汤圆')），
|
||||
旧写法会先按 id 撞主键、再 UPDATE 成新组合，触发 uq_video_raw_uid 二次冲突
|
||||
报 1062。现改为：NAS 本地自增 id 做主键，业务键 (video_id, raw_uid) 唯一，
|
||||
Oracle 的 id 只落 oracle_id 列作溯源参考。
|
||||
"""
|
||||
if not rows:
|
||||
return 0
|
||||
conn = get_conn()
|
||||
n = 0
|
||||
try:
|
||||
cur = conn.cursor()
|
||||
for r in rows:
|
||||
cur.execute(
|
||||
"""INSERT INTO sync_identity_map
|
||||
(oracle_id, video_id, raw_uid, canonical_name, source, updated_at, synced_at)
|
||||
VALUES (%s,%s,%s,%s,%s,%s, NOW())
|
||||
ON DUPLICATE KEY UPDATE
|
||||
oracle_id=VALUES(oracle_id),
|
||||
canonical_name=VALUES(canonical_name),
|
||||
source=VALUES(source),
|
||||
updated_at=VALUES(updated_at),
|
||||
synced_at=NOW()""",
|
||||
(r.get('id'), r.get('video_id'), r.get('raw_uid'),
|
||||
r.get('canonical_name'), r.get('source'), r.get('updated_at')))
|
||||
n += 1
|
||||
conn.commit()
|
||||
return n
|
||||
finally:
|
||||
conn.close()
|
||||
def get_known_members_context() -> str:
|
||||
"""人物清单文本，注入问答 Prompt，让模型用真名指代。"""
|
||||
rows = get_people()
|
||||
parts = []
|
||||
for r in rows:
|
||||
name = r['canonical_name'] or r['label']
|
||||
if r['canonical_name'] and r['canonical_name'] != r['label']:
|
||||
parts.append(f"- {name}（标识：{r['label']}）")
|
||||
else:
|
||||
parts.append(f"- {name}")
|
||||
return "\n".join(parts)
|
||||
|
||||
|
||||
def get_sync_model_calls(limit: int = 200) -> List[Dict]:
|
||||
"""最近模型调用记录（前端统计展示）。"""
|
||||
# ============================================================
|
||||
# model_calls
|
||||
# ============================================================
|
||||
def get_model_calls(limit: int = 200) -> List[Dict]:
|
||||
conn = get_conn()
|
||||
try:
|
||||
cur = conn.cursor(pymysql.cursors.DictCursor)
|
||||
cur.execute(
|
||||
return _rows(conn.execute(
|
||||
"SELECT id, provider, model, video_id, filename, started_at, "
|
||||
"duration_sec, success, error, created_at "
|
||||
"FROM sync_model_calls ORDER BY id DESC LIMIT %s", (limit,))
|
||||
return cur.fetchall()
|
||||
"FROM model_calls ORDER BY id DESC LIMIT ?", (limit,)))
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
def get_sync_model_calls_stats() -> Dict:
|
||||
"""模型调用统计：按 provider+model 聚合成功/失败/平均耗时。"""
|
||||
def get_model_calls_stats() -> List[Dict]:
|
||||
"""按 provider+model 聚合成功/失败/平均耗时。"""
|
||||
conn = get_conn()
|
||||
try:
|
||||
cur = conn.cursor(pymysql.cursors.DictCursor)
|
||||
cur.execute(
|
||||
return _rows(conn.execute(
|
||||
"""SELECT provider, model,
|
||||
SUM(CASE WHEN success=1 THEN 1 ELSE 0 END) AS ok_cnt,
|
||||
SUM(CASE WHEN success=0 THEN 1 ELSE 0 END) AS fail_cnt,
|
||||
ROUND(AVG(duration_sec), 1) AS avg_duration,
|
||||
MAX(created_at) AS last_call
|
||||
FROM sync_model_calls
|
||||
GROUP BY provider, model ORDER BY provider, model""")
|
||||
return cur.fetchall()
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
def get_sync_named_members() -> List[str]:
|
||||
"""已命名成员的真名列表（供 UI 下拉 / 快捷选择）。"""
|
||||
conn = get_conn()
|
||||
try:
|
||||
cur = conn.cursor(pymysql.cursors.DictCursor)
|
||||
cur.execute(
|
||||
"SELECT DISTINCT canonical_name FROM sync_people "
|
||||
"WHERE canonical_name IS NOT NULL AND canonical_name != '' "
|
||||
"ORDER BY canonical_name ASC")
|
||||
return [r['canonical_name'] for r in cur.fetchall()]
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
def get_sync_known_members_context() -> str:
|
||||
"""获取人物清单文本，注入问答 Prompt，让模型用真名指代。"""
|
||||
conn = get_conn()
|
||||
try:
|
||||
cur = conn.cursor(pymysql.cursors.DictCursor)
|
||||
cur.execute(
|
||||
"SELECT label, canonical_name FROM sync_people ORDER BY id ASC")
|
||||
rows = cur.fetchall()
|
||||
if not rows:
|
||||
return ""
|
||||
parts = []
|
||||
for r in rows:
|
||||
name = r['canonical_name'] or r['label']
|
||||
if r['canonical_name'] and r['canonical_name'] != r['label']:
|
||||
parts.append(f"- {name}（标识：{r['label']}）")
|
||||
else:
|
||||
parts.append(f"- {name}")
|
||||
return "\n".join(parts)
|
||||
FROM model_calls GROUP BY provider, model ORDER BY provider, model"""))
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
# ============================================================
|
||||
# 同步游标
|
||||
# 概览统计
|
||||
# ============================================================
|
||||
|
||||
def get_sync_cursor() -> str:
|
||||
def get_stats(date_str: str = None) -> Dict:
|
||||
"""视频数 / 事件数 / 关注事件数 / 出现人物数（可按日期过滤）。"""
|
||||
d = _DATE_EXPR.format(a='sv')
|
||||
conn = get_conn()
|
||||
try:
|
||||
cur = conn.cursor()
|
||||
cur.execute("SELECT `value` FROM sync_cursor WHERE `key`='last_since'")
|
||||
row = cur.fetchone()
|
||||
return row[0] if row else ''
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
def set_sync_cursor(value: str):
|
||||
conn = get_conn()
|
||||
try:
|
||||
cur = conn.cursor()
|
||||
cur.execute(
|
||||
"""INSERT INTO sync_cursor (`key`, `value`) VALUES ('last_since', %s)
|
||||
ON DUPLICATE KEY UPDATE `value`=VALUES(`value`)""",
|
||||
(value,))
|
||||
conn.commit()
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
# ============================================================
|
||||
# 运动通知游标（MotionNotifier 增量去重用，复用 sync_cursor 表）
|
||||
# ============================================================
|
||||
|
||||
def get_motion_cursor() -> str:
|
||||
"""返回上次推送到甲骨文的最大 SS 事件 id（字符串），无则返回空。"""
|
||||
conn = get_conn()
|
||||
try:
|
||||
cur = conn.cursor()
|
||||
cur.execute("SELECT `value` FROM sync_cursor WHERE `key`='motion_last_event_id'")
|
||||
row = cur.fetchone()
|
||||
return row[0] if row else ''
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
def set_motion_cursor(value: str):
|
||||
conn = get_conn()
|
||||
try:
|
||||
cur = conn.cursor()
|
||||
cur.execute(
|
||||
"""INSERT INTO sync_cursor (`key`, `value`) VALUES ('motion_last_event_id', %s)
|
||||
ON DUPLICATE KEY UPDATE `value`=VALUES(`value`)""",
|
||||
(value,))
|
||||
conn.commit()
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
# ============================================================
|
||||
# 统计
|
||||
# ============================================================
|
||||
|
||||
def get_attention_events(limit: int = 200) -> List[Dict]:
|
||||
"""需关注事件列表（供统计图表页展示日期 + 涉及人物），按录制日期倒序。"""
|
||||
conn = get_conn()
|
||||
try:
|
||||
cur = conn.cursor(pymysql.cursors.DictCursor)
|
||||
cur.execute(
|
||||
"""SELECT COALESCE(NULLIF(v.event_start_time,''), v.processed_at) AS ev_date,
|
||||
e.person_list_json
|
||||
FROM sync_events e JOIN sync_videos v ON e.video_id=v.id
|
||||
WHERE e.is_attention_event = 1
|
||||
ORDER BY ev_date DESC LIMIT %s""",
|
||||
(limit,))
|
||||
return cur.fetchall()
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
def get_sync_stats(date_str: str = None) -> Dict:
|
||||
"""概览统计：视频数 / 事件数 / 关注事件数 / 出现人物数（按 date 可选过滤）。
|
||||
|
||||
人物数通过对 sync_events.person_list_json LIKE 统计（MariaDB 不支持 JSON 数组展开）。
|
||||
"""
|
||||
conn = get_conn()
|
||||
try:
|
||||
cur = conn.cursor(pymysql.cursors.DictCursor)
|
||||
# 视频/事件/关注数（日期维度=录制时间 event_start_time，回退 processed_at）
|
||||
_D = "COALESCE(NULLIF(sv.event_start_time,''), sv.processed_at)"
|
||||
if date_str:
|
||||
cur.execute(
|
||||
"""SELECT
|
||||
(SELECT COUNT(*) FROM sync_videos sv
|
||||
WHERE sv.status='done'
|
||||
AND (LEFT(sv.filename, 7) = 'motion_'
|
||||
OR EXISTS(SELECT 1 FROM sync_events se WHERE se.video_id=sv.id))
|
||||
AND {d} LIKE %s) AS videos,
|
||||
(SELECT COUNT(*) FROM sync_events se
|
||||
JOIN sync_videos sv ON se.video_id=sv.id
|
||||
WHERE {d} LIKE %s) AS events,
|
||||
(SELECT COALESCE(SUM(se.is_attention_event),0) FROM sync_events se
|
||||
JOIN sync_videos sv ON se.video_id=sv.id
|
||||
WHERE {d} LIKE %s) AS attention""".format(d=_D),
|
||||
(f'{date_str}%', f'{date_str}%', f'{date_str}%'))
|
||||
like = f'{date_str}%'
|
||||
stat = _row(conn.execute(
|
||||
f"""SELECT
|
||||
(SELECT COUNT(*) FROM videos sv
|
||||
WHERE sv.status='done'
|
||||
AND (substr(sv.filename,1,7)='motion_'
|
||||
OR EXISTS(SELECT 1 FROM events se WHERE se.video_id=sv.id))
|
||||
AND {d} LIKE ?) AS videos,
|
||||
(SELECT COUNT(*) FROM events se JOIN videos sv ON se.video_id=sv.id
|
||||
WHERE {d} LIKE ?) AS events,
|
||||
(SELECT COALESCE(SUM(se.is_attention_event),0) FROM events se
|
||||
JOIN videos sv ON se.video_id=sv.id
|
||||
WHERE {d} LIKE ?) AS attention""",
|
||||
(like, like, like))) or {}
|
||||
else:
|
||||
cur.execute(
|
||||
stat = _row(conn.execute(
|
||||
"""SELECT
|
||||
(SELECT COUNT(*) FROM sync_videos
|
||||
WHERE status='done'
|
||||
AND (LEFT(filename, 7) = 'motion_'
|
||||
OR EXISTS(SELECT 1 FROM sync_events se WHERE se.video_id=sync_videos.id))) AS videos,
|
||||
(SELECT COUNT(*) FROM sync_events) AS events,
|
||||
(SELECT COALESCE(SUM(is_attention_event),0) FROM sync_events) AS attention""")
|
||||
stat = cur.fetchone() or {}
|
||||
(SELECT COUNT(*) FROM videos sv
|
||||
WHERE sv.status='done'
|
||||
AND (substr(sv.filename,1,7)='motion_'
|
||||
OR EXISTS(SELECT 1 FROM events se WHERE se.video_id=sv.id))) AS videos,
|
||||
(SELECT COUNT(*) FROM events) AS events,
|
||||
(SELECT COALESCE(SUM(is_attention_event),0) FROM events) AS attention""")) or {}
|
||||
|
||||
# 人物数（distinct label 命中 sync_events）
|
||||
cur.execute("SELECT id, label, canonical_name FROM sync_people")
|
||||
people = cur.fetchall()
|
||||
# 基于 person_list_json 命中计数：逐 label 统计命中事件数
|
||||
person_hits = 0
|
||||
for p in people:
|
||||
# 出现人物数：逐 label/真名在事件里查有没有命中（沿用改写前的口径）
|
||||
hits = 0
|
||||
for p in _rows(conn.execute("SELECT label, canonical_name FROM people")):
|
||||
name = p['canonical_name'] or p['label']
|
||||
cur.execute(
|
||||
"SELECT COUNT(*) c FROM sync_events WHERE person_list_json LIKE %s",
|
||||
(f'%{name}%',))
|
||||
if cur.fetchone()['c'] > 0:
|
||||
person_hits += 1
|
||||
stat['people'] = person_hits
|
||||
c = conn.execute("SELECT COUNT(*) AS c FROM events "
|
||||
"WHERE person_list_json LIKE ?", (f'%{name}%',)).fetchone()
|
||||
if c and c['c'] > 0:
|
||||
hits += 1
|
||||
stat['people'] = hits
|
||||
return stat
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
|
||||
# ============================================================
|
||||
# chat_history（保留：问答历史）
|
||||
# chat_history（本模块唯一写的表；2026-09-13 从 NAS MariaDB 迁入）
|
||||
# ============================================================
|
||||
def _ensure_chat_schema(conn):
|
||||
global _chat_schema_ready
|
||||
if _chat_schema_ready:
|
||||
return
|
||||
conn.executescript("""
|
||||
CREATE TABLE IF NOT EXISTS chat_history (
|
||||
chat_id INTEGER PRIMARY KEY AUTOINCREMENT,
|
||||
user_question TEXT NOT NULL,
|
||||
ai_answer TEXT NOT NULL,
|
||||
context_summary TEXT,
|
||||
queried_date TEXT,
|
||||
queried_person TEXT,
|
||||
created_at TEXT
|
||||
);
|
||||
CREATE INDEX IF NOT EXISTS idx_chat_created ON chat_history(created_at);
|
||||
""")
|
||||
conn.commit()
|
||||
_chat_schema_ready = True
|
||||
|
||||
|
||||
def insert_chat_history(user_question: str, ai_answer: str,
|
||||
context_summary: str, queried_date: str,
|
||||
queried_person: str) -> int:
|
||||
"""插入对话记录"""
|
||||
conn = get_conn()
|
||||
try:
|
||||
cur = conn.cursor()
|
||||
cur.execute(
|
||||
_ensure_chat_schema(conn)
|
||||
cur = conn.execute(
|
||||
"""INSERT INTO chat_history
|
||||
(user_question, ai_answer, context_summary, queried_date, queried_person)
|
||||
VALUES (%s, %s, %s, %s, %s)""",
|
||||
(user_question, ai_answer, context_summary, queried_date, queried_person)
|
||||
)
|
||||
(user_question, ai_answer, context_summary, queried_date,
|
||||
queried_person, created_at)
|
||||
VALUES (?, ?, ?, ?, ?, ?)""",
|
||||
(user_question, ai_answer, context_summary, queried_date,
|
||||
queried_person, _now()))
|
||||
conn.commit()
|
||||
return cur.lastrowid
|
||||
finally:
|
||||
@@ -702,24 +358,20 @@ def insert_chat_history(user_question: str, ai_answer: str,
|
||||
|
||||
|
||||
def get_chat_history(limit=20, offset=0, date_filter=None, person_filter=None) -> List[Dict]:
|
||||
"""获取对话历史（分页）"""
|
||||
conn = get_conn()
|
||||
try:
|
||||
cur = conn.cursor(pymysql.cursors.DictCursor)
|
||||
conditions = []
|
||||
params = []
|
||||
_ensure_chat_schema(conn)
|
||||
conds, params = [], []
|
||||
if date_filter:
|
||||
conditions.append("queried_date = %s")
|
||||
conds.append("queried_date = ?")
|
||||
params.append(date_filter)
|
||||
if person_filter:
|
||||
conditions.append("queried_person = %s")
|
||||
conds.append("queried_person = ?")
|
||||
params.append(person_filter)
|
||||
where = f"WHERE {' AND '.join(conditions)}" if conditions else ""
|
||||
where = f"WHERE {' AND '.join(conds)}" if conds else ""
|
||||
params.extend([limit, offset])
|
||||
cur.execute(
|
||||
f"SELECT * FROM chat_history {where} ORDER BY created_at DESC LIMIT %s OFFSET %s",
|
||||
params
|
||||
)
|
||||
return cur.fetchall()
|
||||
return _rows(conn.execute(
|
||||
f"SELECT * FROM chat_history {where} ORDER BY created_at DESC LIMIT ? OFFSET ?",
|
||||
params))
|
||||
finally:
|
||||
conn.close()
|
||||
|
||||
102
fam-core/src/fam_core/edge_client.py
Normal file
102
fam-core/src/fam_core/edge_client.py
Normal file
@@ -0,0 +1,102 @@
|
||||
"""
|
||||
Edge-Client - 访问同机 fam-edge 的薄客户端（2026-09-13 迁云后新增）
|
||||
|
||||
迁云之前这些调用散在 `oracle_sync.py` 里（NAS 跨公网访问甲骨文），随镜像层一起
|
||||
删掉了。现在 fam-core 与 fam-edge 同机，全部走 127.0.0.1：不出公网、无 TLS、
|
||||
无隧道，token 仍然带着（fam-edge 侧的鉴权没变，且它同时对公网监听 :5000）。
|
||||
|
||||
只保留三类调用：
|
||||
- push_name_correct / push_video_delete —— 写操作仍由 fam-edge 处理，因为它
|
||||
除了改库还要做人物合并、删磁盘素材，不是单纯一条 SQL
|
||||
- get_activity —— 服务状态页
|
||||
- fetch_image —— 事件帧图 / 人物头像（图源和裁剪都在 fam-edge）
|
||||
"""
|
||||
import requests
|
||||
|
||||
from .config_loader import load_config
|
||||
from .logger import setup_logger
|
||||
|
||||
logger = setup_logger('fam-core.edge_client')
|
||||
|
||||
_cfg = None
|
||||
|
||||
|
||||
def _conf():
|
||||
global _cfg
|
||||
if _cfg is None:
|
||||
c = load_config().get('edge', {})
|
||||
_cfg = {
|
||||
'base_url': (c.get('base_url') or 'http://127.0.0.1:5000').rstrip('/'),
|
||||
'token': c.get('token') or '',
|
||||
'timeout': int(c.get('timeout', 60)),
|
||||
}
|
||||
return _cfg
|
||||
|
||||
|
||||
def base_url():
|
||||
return _conf()['base_url']
|
||||
|
||||
|
||||
def token():
|
||||
return _conf()['token']
|
||||
|
||||
|
||||
def _post(rel: str, payload: dict):
|
||||
"""返回 (ok, err)。写操作统一用这个，失败时把原因带回给前端。"""
|
||||
c = _conf()
|
||||
body = dict(payload)
|
||||
if c['token']:
|
||||
body['token'] = c['token']
|
||||
try:
|
||||
resp = requests.post(f"{c['base_url']}{rel}", json=body, timeout=(5, c['timeout']))
|
||||
except requests.RequestException as e:
|
||||
logger.error(f"fam-edge {rel} 请求失败: {e}")
|
||||
return False, str(e)
|
||||
if resp.status_code != 200:
|
||||
logger.warning(f"fam-edge {rel} 返回 {resp.status_code}: {resp.text[:160]}")
|
||||
return False, f"HTTP {resp.status_code} {resp.text[:120]}"
|
||||
return True, ''
|
||||
|
||||
|
||||
def push_name_correct(label: str, canonical_name: str):
|
||||
return _post('/api/oracle/people/correct',
|
||||
{'label': label, 'canonical_name': canonical_name})
|
||||
|
||||
|
||||
def push_identity_correct(video_id: int, current_name: str, new_name: str):
|
||||
return _post('/api/oracle/identity/correct',
|
||||
{'video_id': video_id, 'current_name': current_name, 'new_name': new_name})
|
||||
|
||||
|
||||
def push_video_delete(video_id: int):
|
||||
return _post('/api/oracle/video/delete', {'video_id': video_id})
|
||||
|
||||
|
||||
def get_activity():
|
||||
"""服务状态页用：fam-edge 的队列 / 分割 / 模型活动快照。返回 (data, err)。"""
|
||||
c = _conf()
|
||||
try:
|
||||
r = requests.get(f"{c['base_url']}/api/oracle/activity",
|
||||
params={'token': c['token']}, timeout=(5, 15))
|
||||
except requests.RequestException as e:
|
||||
return None, f"连接 fam-edge 失败: {e}"
|
||||
if r.status_code != 200:
|
||||
return None, f"fam-edge activity HTTP {r.status_code}"
|
||||
return r.json(), None
|
||||
|
||||
|
||||
def fetch_image(rel: str, params: dict):
|
||||
"""帧图 / 头像透传，失败返回 None（调用方给 404）。"""
|
||||
c = _conf()
|
||||
p = dict(params)
|
||||
if c['token']:
|
||||
p['token'] = c['token']
|
||||
try:
|
||||
resp = requests.get(f"{c['base_url']}{rel}", params=p, timeout=(5, c['timeout']))
|
||||
except requests.RequestException as e:
|
||||
logger.error(f"fam-edge {rel} 请求失败: {e}")
|
||||
return None
|
||||
if resp.status_code == 200 and resp.content:
|
||||
return resp.content
|
||||
logger.warning(f"fam-edge {rel} 返回 {resp.status_code}")
|
||||
return None
|
||||
@@ -1,14 +1,12 @@
|
||||
"""
|
||||
Img-Proxy - NAS 端图片代理（图源在 Oracle，计算全部在 Oracle）
|
||||
Img-Proxy - 图片代理（图源与裁剪都在 fam-edge）
|
||||
|
||||
浏览器不直连 Oracle（避免公网暴露 5000 端口与 token 外泄），
|
||||
而是访问 NAS fam-core 的 /api/proxy/*，由 NAS 出网到 Oracle 拉取 jpeg 回传。
|
||||
|
||||
Oracle 侧已有磁盘缓存 / VLM 人物定位裁剪，NAS 端仅透传，不做图像计算。
|
||||
浏览器不直连 fam-edge（避免 token 外泄），而是访问 /api/proxy/*，由本服务转一手。
|
||||
迁云后这一跳是同机 127.0.0.1，纯透传，不做图像计算。
|
||||
"""
|
||||
from flask import Blueprint, Response, request
|
||||
|
||||
from .oracle_sync import get_sync
|
||||
from . import edge_client
|
||||
from .logger import setup_logger
|
||||
|
||||
logger = setup_logger('fam-core.img_proxy')
|
||||
@@ -17,21 +15,7 @@ img_bp = Blueprint('img_proxy', __name__)
|
||||
|
||||
|
||||
def _fetch(rel: str, params: dict):
|
||||
import requests
|
||||
sync = get_sync() # 复用 oracle_sync 已解析好的 base_url/token/timeout，避免两处配置各读一份
|
||||
params = dict(params)
|
||||
if sync.token:
|
||||
params['token'] = sync.token
|
||||
try:
|
||||
resp = requests.get(f"{sync.base_url}{rel}", params=params,
|
||||
timeout=(10, sync.timeout))
|
||||
if resp.status_code == 200 and resp.content:
|
||||
return resp.content
|
||||
logger.warning(f"Oracle {rel} 返回 {resp.status_code}: {resp.text[:120]}")
|
||||
return None
|
||||
except Exception as e:
|
||||
logger.error(f"Oracle {rel} 请求失败: {e}")
|
||||
return None
|
||||
return edge_client.fetch_image(rel, params)
|
||||
|
||||
|
||||
@img_bp.route('/api/proxy/frame', methods=['GET'])
|
||||
|
||||
@@ -14,7 +14,7 @@ from flask import Blueprint, request, jsonify
|
||||
|
||||
from ..logger import setup_logger
|
||||
from .. import db_layer
|
||||
from ..oracle_sync import get_sync
|
||||
from .. import edge_client
|
||||
|
||||
logger = setup_logger('fam-core.member_manager')
|
||||
|
||||
@@ -24,7 +24,7 @@ member_bp = Blueprint('member_manager', __name__)
|
||||
@member_bp.route('/api/member/unnamed', methods=['GET'])
|
||||
def list_unnamed():
|
||||
"""列出未命名人物（canonical_name 为空）"""
|
||||
members = db_layer.get_sync_people()
|
||||
members = db_layer.get_people()
|
||||
result = []
|
||||
for m in members:
|
||||
canonical = m.get('canonical_name')
|
||||
@@ -40,7 +40,7 @@ def list_unnamed():
|
||||
@member_bp.route('/api/member/list', methods=['GET'])
|
||||
def list_members():
|
||||
"""列出所有人物（按 canonical_name 或 label 展示）"""
|
||||
members = db_layer.get_sync_people()
|
||||
members = db_layer.get_people()
|
||||
result = []
|
||||
for m in members:
|
||||
canonical = m.get('canonical_name')
|
||||
@@ -58,7 +58,7 @@ def list_members():
|
||||
|
||||
@member_bp.route('/api/member/name', methods=['POST'])
|
||||
def name_member():
|
||||
"""命名人物（回推 Oracle + 立即拉回本地镜像）
|
||||
"""命名人物（交给 fam-edge 落库 + 合并人物）
|
||||
|
||||
请求: {"label": "人物A", "canonical_name": "张三"}
|
||||
"""
|
||||
@@ -71,18 +71,13 @@ def name_member():
|
||||
if not label or not canonical_name:
|
||||
return jsonify({"error": "缺少必填字段: label, canonical_name"}), 400
|
||||
|
||||
logger.info(f"命名: {label} -> {canonical_name}（回推 Oracle）")
|
||||
ok, err = get_sync().push_name_correct(label, canonical_name)
|
||||
logger.info(f"命名: {label} -> {canonical_name}（回推 fam-edge）")
|
||||
ok, err = edge_client.push_name_correct(label, canonical_name)
|
||||
if not ok:
|
||||
return jsonify({"error": f"回推 Oracle 失败: {err}"}), 502
|
||||
return jsonify({"error": f"回推 fam-edge 失败: {err}"}), 502
|
||||
|
||||
# 立即拉回最新 people 镜像，前端无需等待下一个 30 分钟周期
|
||||
try:
|
||||
get_sync().trigger_now()
|
||||
except Exception as e:
|
||||
logger.warning(f"命名后即时拉回失败（下一个周期会自动同步）: {e}")
|
||||
|
||||
members = db_layer.get_sync_people()
|
||||
members = db_layer.get_people()
|
||||
return jsonify({
|
||||
"status": "ok",
|
||||
"label": label,
|
||||
@@ -114,24 +109,20 @@ def merge_member():
|
||||
return jsonify({"error": "source 与 target 不能相同"}), 400
|
||||
|
||||
# 解析 target 的规范名
|
||||
members = {m['label']: m for m in db_layer.get_sync_people()}
|
||||
members = {m['label']: m for m in db_layer.get_people()}
|
||||
target_row = members.get(target)
|
||||
if target_row and target_row.get('canonical_name'):
|
||||
canonical = target_row['canonical_name']
|
||||
else:
|
||||
canonical = target # target 未命名 -> 以 label 作为规范名
|
||||
|
||||
logger.info(f"合并: {source} -> {canonical}（回推 Oracle）")
|
||||
ok, err = get_sync().push_name_correct(source, canonical)
|
||||
logger.info(f"合并: {source} -> {canonical}（回推 fam-edge）")
|
||||
ok, err = edge_client.push_name_correct(source, canonical)
|
||||
if not ok:
|
||||
return jsonify({"error": f"回推 Oracle 失败: {err}"}), 502
|
||||
return jsonify({"error": f"回推 fam-edge 失败: {err}"}), 502
|
||||
|
||||
try:
|
||||
get_sync().trigger_now()
|
||||
except Exception as e:
|
||||
logger.warning(f"合并后即时拉回失败（下一个周期会自动同步）: {e}")
|
||||
|
||||
members = db_layer.get_sync_people()
|
||||
members = db_layer.get_people()
|
||||
return jsonify({
|
||||
"status": "ok",
|
||||
"source": source,
|
||||
@@ -163,15 +154,11 @@ def identity_correct():
|
||||
if not video_id or not current_name or not new_name:
|
||||
return jsonify({"error": "缺少必填字段: video_id, current_name, new_name"}), 400
|
||||
|
||||
logger.info(f"人物纠错: video_id={video_id} {current_name} -> {new_name}（回推 Oracle）")
|
||||
ok, err = get_sync().push_identity_correct(video_id, current_name, new_name)
|
||||
logger.info(f"人物纠错: video_id={video_id} {current_name} -> {new_name}（回推 fam-edge）")
|
||||
ok, err = edge_client.push_identity_correct(video_id, current_name, new_name)
|
||||
if not ok:
|
||||
return jsonify({"error": f"回推 Oracle 失败: {err}"}), 502
|
||||
return jsonify({"error": f"回推 fam-edge 失败: {err}"}), 502
|
||||
|
||||
try:
|
||||
get_sync().trigger_now()
|
||||
except Exception as e:
|
||||
logger.warning(f"纠错后即时拉回失败（下一个周期会自动同步）: {e}")
|
||||
|
||||
return jsonify({
|
||||
"status": "ok", "video_id": video_id,
|
||||
|
||||
@@ -1,20 +0,0 @@
|
||||
"""
|
||||
运动监测状态接口（fam-core）
|
||||
|
||||
运动事件采集走 MotionNotifier 轮询 SS EventCenter.Event.List（真实
|
||||
event_id/start_time/duration），不再提供 /api/ss/webhook 接收端点
|
||||
（SS Webhook 行动规则已弃用，2026-08-25 移除）。
|
||||
"""
|
||||
from flask import Blueprint, jsonify
|
||||
|
||||
from .logger import setup_logger
|
||||
from .motion_notifier.motion_notifier import get_motion_notifier
|
||||
|
||||
logger = setup_logger('fam-core.motion_bp')
|
||||
|
||||
motion_bp = Blueprint('motion_bp', __name__)
|
||||
|
||||
|
||||
@motion_bp.route('/api/ss/status', methods=['GET'])
|
||||
def ss_status():
|
||||
return jsonify(get_motion_notifier().status()), 200
|
||||
@@ -1,4 +0,0 @@
|
||||
"""MotionNotifier 包：NAS 端运动监测通知服务。"""
|
||||
from .motion_notifier import MotionNotifier, get_motion_notifier
|
||||
|
||||
__all__ = ["MotionNotifier", "get_motion_notifier"]
|
||||
@@ -1,370 +0,0 @@
|
||||
"""
|
||||
MotionNotifier - NAS 端运动监测通知服务（轮询主路径，2026-08-22 定型）
|
||||
|
||||
职责:
|
||||
在 NAS 本机**轮询**群晖 Surveillance Station 的运动侦测事件
|
||||
（SYNO.SurveillanceStation.EventCenter.Event method=List，参数名下划线风格
|
||||
camera_ids/start_time/end_time，event_type=10 即运动），增量推送到甲骨文
|
||||
FAM-Edge 的 /api/ss/motion 接口。轮询能拿到真实 event_id/start_time/duration。
|
||||
|
||||
数据方向（关键约束）: NAS -> Oracle，单向。甲骨文不再反向访问 NAS。
|
||||
- 原 fam-edge 的 dsm_motion_client（甲骨文主动查 SS API）已停用；
|
||||
- 推送后由甲骨文本地落库 ss_motion_events，供 video_processor 本地预过滤使用。
|
||||
|
||||
可靠性设计:
|
||||
- 游标（MariaDB sync_cursor.motion_last_event_id）持久化；重启优先续用 DB 游标，
|
||||
停机期间的事件由窗口回看补推；仅首次部署时才初始化为 SS 当前最大 id。
|
||||
- 推送失败的批次不推进游标，下一轮窗口回看重试（不丢事件）。
|
||||
- 心跳线程（enabled 即跑，与轮询无关）：定期空 POST /api/ss/motion，证明
|
||||
NAS->Oracle 推送链路存活，供甲骨文侧判断"无运动"结论是否可信。
|
||||
|
||||
"""
|
||||
import os
|
||||
import re
|
||||
import time
|
||||
import threading
|
||||
from datetime import datetime, timezone
|
||||
|
||||
import requests
|
||||
|
||||
from ..logger import setup_logger
|
||||
from ..config_loader import load_config
|
||||
from .. import db_layer
|
||||
|
||||
logger = setup_logger('fam-core.motion_notifier')
|
||||
|
||||
_MOTION_NOTIFIER = None
|
||||
|
||||
|
||||
def get_motion_notifier():
|
||||
"""模块级单例（app.py 启动时创建并 start，状态接口经此获取）。"""
|
||||
global _MOTION_NOTIFIER
|
||||
if _MOTION_NOTIFIER is None:
|
||||
_MOTION_NOTIFIER = MotionNotifier()
|
||||
return _MOTION_NOTIFIER
|
||||
|
||||
|
||||
class MotionNotifier:
|
||||
def __init__(self):
|
||||
cfg = load_config().get('motion_notifier', {})
|
||||
self.enabled = bool(cfg.get('enabled', False))
|
||||
self.dsm_host = cfg.get('dsm_host', '192.168.50.64')
|
||||
self.dsm_port = int(cfg.get('dsm_port', 5000))
|
||||
self.dsm_account = self._resolve(cfg.get('dsm_account', ''))
|
||||
self.dsm_password = self._resolve(cfg.get('dsm_password', ''))
|
||||
self.camera_ids = cfg.get('camera_ids', [2])
|
||||
self.oracle_base_url = cfg.get('oracle_base_url',
|
||||
'http://129.146.26.249:5000').rstrip('/')
|
||||
self.oracle_token = self._resolve(cfg.get('oracle_token', '${ORACLE_SYNC_TOKEN}'))
|
||||
self.poll_interval_sec = int(cfg.get('poll_interval_sec', 60))
|
||||
self.poll_window_hours = int(cfg.get('poll_window_hours', 2))
|
||||
self.batch_size = int(cfg.get('batch_size', 100))
|
||||
self.timeout = int(cfg.get('timeout_sec', 10))
|
||||
# 是否启用轮询（默认开启）：轮询 EventCenter.Event.List 为主数据源，拿到
|
||||
# 真实 event_id/start_time/duration；Webhook 仅作为可选的低延迟补充
|
||||
# （SS 行动规则未配置时不会触发，不影响主路径）
|
||||
self.poll_enabled = bool(cfg.get('poll_enabled', True))
|
||||
# 心跳：跟"轮询 SS"是两回事——不查 SS，只是定期空 POST 一下甲骨文，证明
|
||||
# NAS->Oracle 这条推送链路本身还活着。enabled=true 时始终跑（不受
|
||||
# poll_enabled 影响），供甲骨文侧 has_motion_in_range_local() 判断
|
||||
# "这段时间没收到运动事件"是真的没运动，还是推送链路已经挂了。
|
||||
self.heartbeat_interval_sec = int(cfg.get('heartbeat_interval_sec', 300))
|
||||
# 轮询路径（poll_enabled=true 时）关注的摄像头
|
||||
self.camera_ids = cfg.get('camera_ids', [2])
|
||||
self._base = f"http://{self.dsm_host}:{self.dsm_port}/webapi"
|
||||
self._sid = None
|
||||
self._running = False
|
||||
self._thread = None
|
||||
self._last_event_id = None
|
||||
# 推送过但 duration<=0（动作进行中，SS 尚未回填真实时长）的事件 id。
|
||||
# 下轮窗口回查时若已结束（duration>0）补推覆盖 Oracle，避免分割永远跳过。
|
||||
self._zero_dur_ids = set()
|
||||
self._last_poll_at = None
|
||||
self._last_error = None
|
||||
self._pushed_total = 0
|
||||
self._hb_running = False
|
||||
self._hb_thread = None
|
||||
self._last_heartbeat_at = None
|
||||
self._last_heartbeat_error = None
|
||||
|
||||
@staticmethod
|
||||
def _resolve(v):
|
||||
"""解析 ${ENV} 引用；非字符串或不含 ${...} 原样返回。"""
|
||||
if isinstance(v, str) and v.startswith('${') and v.endswith('}'):
|
||||
return os.environ.get(v[2:-1], '')
|
||||
return v
|
||||
|
||||
# ------------------------------------------------------------------
|
||||
# Surveillance Station 登录 / 事件查询
|
||||
# ------------------------------------------------------------------
|
||||
def _login(self):
|
||||
try:
|
||||
resp = requests.get(
|
||||
f"{self._base}/auth.cgi",
|
||||
params={"api": "SYNO.API.Auth", "version": 6, "method": "login",
|
||||
"account": self.dsm_account, "passwd": self.dsm_password,
|
||||
"session": "SurveillanceStation", "format": "sid"},
|
||||
timeout=self.timeout)
|
||||
data = resp.json()
|
||||
if data.get('success'):
|
||||
return data['data']['sid']
|
||||
logger.warning(f"SS 登录失败: {data.get('error')}")
|
||||
except Exception as e:
|
||||
logger.warning(f"SS 登录异常: {e}")
|
||||
return None
|
||||
|
||||
def _fetch_events(self, start_ts: int, end_ts: int):
|
||||
"""查询 [start_ts, end_ts] 窗口内的 SS 事件。返回事件列表或 None（查询失败）。"""
|
||||
if self._sid is None:
|
||||
self._sid = self._login()
|
||||
if self._sid is None:
|
||||
return None
|
||||
for attempt in range(2):
|
||||
try:
|
||||
resp = requests.get(
|
||||
f"{self._base}/entry.cgi",
|
||||
params={"api": "SYNO.SurveillanceStation.EventCenter.Event",
|
||||
"version": 1, "method": "List",
|
||||
"camera_ids": ",".join(str(c) for c in self.camera_ids),
|
||||
"start_time": start_ts, "end_time": end_ts,
|
||||
"limit": 1000, "_sid": self._sid},
|
||||
timeout=self.timeout)
|
||||
data = resp.json()
|
||||
except Exception as e:
|
||||
logger.warning(f"SS 事件查询异常: {e}")
|
||||
return None
|
||||
if not data.get('success'):
|
||||
code = (data.get('error') or {}).get('code')
|
||||
if code in (106, 107, 119) and attempt == 0:
|
||||
# session 过期/被顶掉，重新登录重试一次
|
||||
self._sid = self._login()
|
||||
if self._sid is None:
|
||||
return None
|
||||
continue
|
||||
logger.warning(f"SS 事件查询失败: {data.get('error')}")
|
||||
return None
|
||||
# 响应按 ds_id（CMS 多机场景的服务器 id，单机固定 "0"）分组，拉平
|
||||
events = [e for grp in (data.get('data') or {}).values()
|
||||
for e in (grp or [])]
|
||||
return events
|
||||
return None
|
||||
|
||||
# ------------------------------------------------------------------
|
||||
# 推送到甲骨文
|
||||
# ------------------------------------------------------------------
|
||||
def push_events_to_oracle(self, events) -> int:
|
||||
"""把标准化后的事件列表推送到 Oracle /api/ss/motion。返回成功推送条数。"""
|
||||
if not events:
|
||||
return 0
|
||||
norm = []
|
||||
for e in events:
|
||||
eid = e.get('id') or e.get('event_id')
|
||||
if eid is None:
|
||||
continue
|
||||
norm.append({
|
||||
"event_id": int(eid),
|
||||
"camera_id": e.get('camera_id'),
|
||||
"event_type": e.get('event_type'),
|
||||
"start_time": e.get('start_time'),
|
||||
"duration": e.get('duration'),
|
||||
"thumbnail_url": e.get('thumbnail_url') or e.get('thumbnail_dir'),
|
||||
})
|
||||
if not norm:
|
||||
return 0
|
||||
try:
|
||||
resp = requests.post(
|
||||
f"{self.oracle_base_url}/api/ss/motion",
|
||||
json={"token": self.oracle_token, "events": norm},
|
||||
timeout=(10, 30))
|
||||
except requests.RequestException as e:
|
||||
logger.error(f"推送运动事件到 Oracle 失败: {e}")
|
||||
self._last_error = str(e)
|
||||
return 0
|
||||
if resp.status_code != 200:
|
||||
logger.error(f"推送运动事件到 Oracle 返回 {resp.status_code}: {resp.text[:200]}")
|
||||
self._last_error = f"HTTP {resp.status_code}"
|
||||
return 0
|
||||
try:
|
||||
stored = resp.json().get('stored', 0)
|
||||
except ValueError:
|
||||
stored = len(norm)
|
||||
self._pushed_total += stored
|
||||
self._last_error = None
|
||||
logger.info(f"运动事件推送成功: {len(norm)} 条 -> Oracle 存储 {stored}")
|
||||
return stored
|
||||
|
||||
def send_heartbeat(self) -> bool:
|
||||
"""空 events 调一次 /api/ss/motion，只为证明 NAS->Oracle 推送链路还活着。
|
||||
|
||||
跟 push_events_to_oracle 分开一个方法，是因为那个方法 events 为空时直接
|
||||
return 0（不发请求）——心跳恰恰就是要在没有真实事件时也发一次请求。
|
||||
"""
|
||||
try:
|
||||
resp = requests.post(
|
||||
f"{self.oracle_base_url}/api/ss/motion",
|
||||
json={"token": self.oracle_token, "events": []},
|
||||
timeout=(10, 30))
|
||||
except requests.RequestException as e:
|
||||
logger.warning(f"运动心跳推送失败: {e}")
|
||||
self._last_heartbeat_error = str(e)
|
||||
return False
|
||||
if resp.status_code != 200:
|
||||
logger.warning(f"运动心跳推送返回 {resp.status_code}: {resp.text[:200]}")
|
||||
self._last_heartbeat_error = f"HTTP {resp.status_code}"
|
||||
return False
|
||||
self._last_heartbeat_at = datetime.now()
|
||||
self._last_heartbeat_error = None
|
||||
return True
|
||||
|
||||
def _heartbeat_run(self):
|
||||
logger.info(f"运动心跳线程启动，间隔 {self.heartbeat_interval_sec}s")
|
||||
while self._hb_running:
|
||||
try:
|
||||
self.send_heartbeat()
|
||||
except Exception as e:
|
||||
self._last_heartbeat_error = str(e)
|
||||
logger.error(f"运动心跳异常: {e}", exc_info=True)
|
||||
for _ in range(self.heartbeat_interval_sec):
|
||||
if not self._hb_running:
|
||||
break
|
||||
time.sleep(1)
|
||||
|
||||
# ------------------------------------------------------------------
|
||||
# 增量轮询主循环
|
||||
# ------------------------------------------------------------------
|
||||
def _init_cursor(self):
|
||||
"""启动时初始化 last_event_id。
|
||||
|
||||
优先使用 DB 已保存的游标：服务停机/重启期间产生的新事件，会在重启后
|
||||
被下一轮窗口回看捞到并补推（不丢事件）。
|
||||
仅当 DB 无有效游标（首次部署）时，取 SS 当前最大事件 id 作为起点，
|
||||
避免把历史事件全部回灌一遍。
|
||||
"""
|
||||
saved = db_layer.get_motion_cursor()
|
||||
if saved and int(saved) > 0:
|
||||
self._last_event_id = int(saved)
|
||||
logger.info(f"运动通知游标初始化（DB）: last_event_id={self._last_event_id}")
|
||||
return
|
||||
now = int(datetime.now(timezone.utc).timestamp())
|
||||
events = self._fetch_events(now - 3600, now) # 最近 1h 用于定位最大 id
|
||||
if events:
|
||||
self._last_event_id = max(
|
||||
int(e.get('id', 0)) for e in events if e.get('id'))
|
||||
else:
|
||||
self._last_event_id = 0
|
||||
logger.info(f"运动通知游标初始化（SS 当前最大）: last_event_id={self._last_event_id}")
|
||||
|
||||
def _poll_once(self):
|
||||
now = int(datetime.now(timezone.utc).timestamp())
|
||||
start_ts = now - int(self.poll_window_hours * 3600)
|
||||
events = self._fetch_events(start_ts, now)
|
||||
if events is None:
|
||||
return # 查询失败，下一轮重试
|
||||
by_id = {int(e.get('id')): e for e in events if e.get('id')}
|
||||
|
||||
# 1) 补推：先前推送时 duration<=0（动作进行中）的事件，现在若已结束
|
||||
# （SS 回填真实 duration>0），重推覆盖 Oracle（ON CONFLICT UPDATE）。
|
||||
# 重启兜底：游标 DB 续用 → 窗口回看会把历史事件整体重推，届时同样覆盖。
|
||||
if self._zero_dur_ids:
|
||||
recheck = []
|
||||
for eid in list(self._zero_dur_ids):
|
||||
e = by_id.get(eid)
|
||||
if e and int(e.get('duration') or 0) > 0:
|
||||
recheck.append(e)
|
||||
self._zero_dur_ids.discard(eid)
|
||||
if recheck:
|
||||
self.push_events_to_oracle(recheck)
|
||||
logger.info(f"补推 {len(recheck)} 条已结束事件的真实 duration")
|
||||
|
||||
# 2) 增量推送新事件（id 大于游标）
|
||||
new = [e for e in events
|
||||
if e.get('id') and int(e.get('id')) > (self._last_event_id or 0)]
|
||||
if not new:
|
||||
return
|
||||
new.sort(key=lambda e: int(e.get('id', 0)))
|
||||
cursor = self._last_event_id
|
||||
for i in range(0, len(new), self.batch_size):
|
||||
batch = new[i:i + self.batch_size]
|
||||
stored = self.push_events_to_oracle(batch)
|
||||
if stored > 0:
|
||||
# 只有推送成功的批次才推进游标；失败批次保持原地，
|
||||
# 下一轮窗口回看会重新捞到并重试（不丢事件）
|
||||
cursor = max(int(e.get('id', 0)) for e in batch)
|
||||
# 记录推送时仍在进行中的事件（duration<=0），待下轮补推真实时长
|
||||
for e in batch:
|
||||
if int(e.get('duration') or 0) <= 0:
|
||||
self._zero_dur_ids.add(int(e.get('id')))
|
||||
self._last_event_id = cursor
|
||||
db_layer.set_motion_cursor(str(self._last_event_id))
|
||||
|
||||
def _run(self):
|
||||
logger.info(f"MotionNotifier 启动，轮询间隔 {self.poll_interval_sec}s，"
|
||||
f"目标 SS {self.dsm_host}:{self.dsm_port}")
|
||||
try:
|
||||
self._init_cursor()
|
||||
except Exception as e:
|
||||
logger.warning(f"运动通知游标初始化失败（从 0 开始）: {e}")
|
||||
self._last_event_id = 0
|
||||
while self._running:
|
||||
try:
|
||||
self._poll_once()
|
||||
self._last_poll_at = datetime.now()
|
||||
except Exception as e:
|
||||
self._last_error = str(e)
|
||||
logger.error(f"运动通知轮询异常: {e}", exc_info=True)
|
||||
# 分段休眠，便于 stop 快速唤醒
|
||||
for _ in range(self.poll_interval_sec):
|
||||
if not self._running:
|
||||
break
|
||||
time.sleep(1)
|
||||
|
||||
# ------------------------------------------------------------------
|
||||
def start(self):
|
||||
if not self.enabled:
|
||||
logger.info("MotionNotifier 未启用（motion_notifier.enabled=false）")
|
||||
return
|
||||
# 心跳线程跟轮询是否开启无关：只要 MotionNotifier 整体 enabled，就该
|
||||
# 持续证明推送链路活着，哪怕当前正好没有真实运动事件可推送。
|
||||
if not self._hb_running:
|
||||
self._hb_running = True
|
||||
self._hb_thread = threading.Thread(
|
||||
target=self._heartbeat_run, daemon=True, name='motion-heartbeat')
|
||||
self._hb_thread.start()
|
||||
if not self.poll_enabled:
|
||||
logger.info("MotionNotifier 轮询已禁用（poll_enabled=false），仅作为 "
|
||||
"Webhook 推送客户端 + 心跳 + 摄像头名映射使用")
|
||||
return
|
||||
if self._running:
|
||||
return
|
||||
self._running = True
|
||||
self._thread = threading.Thread(target=self._run, daemon=True,
|
||||
name='motion-notifier')
|
||||
self._thread.start()
|
||||
|
||||
def is_alive(self):
|
||||
return self._thread is not None and self._thread.is_alive()
|
||||
|
||||
def stop(self):
|
||||
self._running = False
|
||||
if self._thread:
|
||||
self._thread.join(timeout=5)
|
||||
self._hb_running = False
|
||||
if self._hb_thread:
|
||||
self._hb_thread.join(timeout=5)
|
||||
|
||||
def status(self) -> dict:
|
||||
return {
|
||||
"heartbeat_running": self._hb_thread is not None and self._hb_thread.is_alive(),
|
||||
"last_heartbeat_at": self._last_heartbeat_at.isoformat() if self._last_heartbeat_at else None,
|
||||
"last_heartbeat_error": self._last_heartbeat_error,
|
||||
"running": self.is_alive(),
|
||||
"enabled": self.enabled,
|
||||
"poll_enabled": self.poll_enabled,
|
||||
"last_event_id": self._last_event_id,
|
||||
"zero_dur_pending": len(self._zero_dur_ids),
|
||||
"last_poll_at": self._last_poll_at.isoformat() if self._last_poll_at else None,
|
||||
"last_error": self._last_error,
|
||||
"pushed_total": self._pushed_total,
|
||||
"oracle": self.oracle_base_url,
|
||||
"dsm": f"{self.dsm_host}:{self.dsm_port}",
|
||||
}
|
||||
@@ -1,234 +0,0 @@
|
||||
"""
|
||||
Oracle-Sync - NAS 端唯一后台线程
|
||||
|
||||
职责:
|
||||
1. 每 interval_sec（默认 1800s = 30 分钟）从甲骨文 FAM-Edge 拉取增量:
|
||||
GET {base_url}/api/oracle/sync?since=<cursor>&token=<token>
|
||||
返回 {videos, events, people, server_time}，写入本地 MariaDB 镜像表
|
||||
(sync_videos / sync_events / sync_people)，并推进 sync_cursor。
|
||||
2. 接收命名校正回推: POST {base_url}/api/oracle/people/correct
|
||||
{label, canonical_name, token} —— 手动命名（manual 优先，不被 LLM 覆盖）。
|
||||
|
||||
数据流向（新架构 v2）:
|
||||
Google 硬盘 --rclone--> 甲骨文本地 --> 整视频分析 --> Oracle SQLite
|
||||
--> [本线程每 30 分钟拉增量] --> NAS MariaDB 镜像 --> fam-ui 读取展示
|
||||
|
||||
NAS 不再处理任何视频，CPU 占用显著降低。
|
||||
"""
|
||||
import time
|
||||
import threading
|
||||
import requests
|
||||
from datetime import datetime
|
||||
|
||||
from .logger import setup_logger
|
||||
from .config_loader import load_config
|
||||
from . import db_layer
|
||||
|
||||
logger = setup_logger('fam-core.oracle_sync')
|
||||
|
||||
_SYNC_INSTANCE = None
|
||||
|
||||
|
||||
def get_sync():
|
||||
"""模块级单例（app.py 启动时创建并 start，其余模块经此获取）"""
|
||||
global _SYNC_INSTANCE
|
||||
if _SYNC_INSTANCE is None:
|
||||
_SYNC_INSTANCE = OracleSync()
|
||||
return _SYNC_INSTANCE
|
||||
|
||||
|
||||
class OracleSync:
|
||||
def __init__(self):
|
||||
cfg = load_config().get('oracle_sync', {})
|
||||
self.base_url = cfg.get('base_url', 'http://129.146.26.249:5000').rstrip('/')
|
||||
self.token = cfg.get('token', '')
|
||||
self.interval_sec = int(cfg.get('interval_sec', 1800))
|
||||
self.timeout = int(cfg.get('timeout', 120))
|
||||
self._running = False
|
||||
self._thread = None
|
||||
self._last_sync_at = None
|
||||
self._last_error = None
|
||||
self._last_count = None
|
||||
# 拉取互斥锁：trigger_now（命名后即时拉回）与后台 _run 并发时只允许一个执行
|
||||
self._pull_lock = threading.Lock()
|
||||
|
||||
# ------------------------------------------------------------------
|
||||
def _pull_once(self) -> bool:
|
||||
"""执行一次增量拉取（加锁防并发双拉）。返回是否成功。"""
|
||||
with self._pull_lock:
|
||||
return self._pull_once_locked()
|
||||
|
||||
def _pull_once_locked(self) -> bool:
|
||||
since = db_layer.get_sync_cursor() or ''
|
||||
params = {'since': since, 'token': self.token}
|
||||
try:
|
||||
resp = requests.get(
|
||||
f"{self.base_url}/api/oracle/sync",
|
||||
params=params, timeout=(10, self.timeout))
|
||||
except requests.RequestException as e:
|
||||
self._last_error = f"请求失败: {e}"
|
||||
logger.error(f"拉取同步失败: {e}")
|
||||
return False
|
||||
|
||||
if resp.status_code == 401:
|
||||
self._last_error = "token 校验失败"
|
||||
logger.error("同步 token 校验失败 (401)，请检查 oracle_sync.token 配置")
|
||||
return False
|
||||
if resp.status_code != 200:
|
||||
self._last_error = f"HTTP {resp.status_code}"
|
||||
logger.error(f"同步返回异常: {resp.status_code} {resp.text[:200]}")
|
||||
return False
|
||||
|
||||
try:
|
||||
data = resp.json()
|
||||
except ValueError:
|
||||
self._last_error = "非 JSON 响应"
|
||||
logger.error("同步返回非 JSON 响应")
|
||||
return False
|
||||
|
||||
videos = data.get('videos', []) or []
|
||||
events = data.get('events', []) or []
|
||||
people = data.get('people', []) or []
|
||||
model_calls = data.get('model_calls', []) or []
|
||||
identity_map = data.get('identity_map', []) or []
|
||||
server_time = data.get('server_time', '') or ''
|
||||
|
||||
n_videos = db_layer.upsert_sync_videos(videos)
|
||||
n_events = db_layer.upsert_sync_events(events)
|
||||
n_people = db_layer.upsert_sync_people(people)
|
||||
n_calls = db_layer.upsert_sync_model_calls(model_calls)
|
||||
n_identity = db_layer.upsert_sync_identity_map(identity_map)
|
||||
|
||||
if server_time:
|
||||
db_layer.set_sync_cursor(server_time)
|
||||
|
||||
self._last_sync_at = datetime.now()
|
||||
self._last_error = None
|
||||
self._last_count = (n_videos, n_events, n_people, n_calls, n_identity)
|
||||
logger.info(
|
||||
f"同步完成: videos+{n_videos} events+{n_events} people+{n_people} "
|
||||
f"model_calls+{n_calls} identity_map+{n_identity} since={since!r} "
|
||||
f"-> server_time={server_time}")
|
||||
return True
|
||||
|
||||
# ------------------------------------------------------------------
|
||||
def push_name_correct(self, label: str, canonical_name: str):
|
||||
"""回推命名校正到 Oracle（手动命名优先级最高，不被 LLM 覆盖）。
|
||||
|
||||
返回 (success: bool, error: str)
|
||||
"""
|
||||
label = (label or '').strip()
|
||||
canonical_name = (canonical_name or '').strip()
|
||||
if not label or not canonical_name:
|
||||
return False, "缺少 label / canonical_name"
|
||||
try:
|
||||
resp = requests.post(
|
||||
f"{self.base_url}/api/oracle/people/correct",
|
||||
json={"label": label, "canonical_name": canonical_name,
|
||||
"token": self.token},
|
||||
timeout=(10, 30))
|
||||
except requests.RequestException as e:
|
||||
logger.error(f"命名校正回推失败: {e}")
|
||||
return False, str(e)
|
||||
if resp.status_code == 200:
|
||||
return True, ""
|
||||
msg = f"HTTP {resp.status_code}: {resp.text[:200]}"
|
||||
logger.error(f"命名校正回推失败: {msg}")
|
||||
return False, msg
|
||||
|
||||
# ------------------------------------------------------------------
|
||||
def push_identity_correct(self, video_id: int, current_name: str, new_name: str):
|
||||
"""回推事件时间轴/人物管理"这个人识别错了"纠错到 Oracle。
|
||||
|
||||
跟 push_name_correct 的区别：这个按 (video_id, 当前展示名) 定位，只改这
|
||||
一段视频里错认的那个人，不影响同名字符串在其他视频里的映射（人物 uid
|
||||
只在单次视频分析内稳定，跨视频复用同一字符串完全可能是不同真人）。
|
||||
|
||||
返回 (success: bool, error: str)
|
||||
"""
|
||||
try:
|
||||
video_id = int(video_id)
|
||||
except (TypeError, ValueError):
|
||||
return False, "video_id 必须是数字"
|
||||
current_name = (current_name or '').strip()
|
||||
new_name = (new_name or '').strip()
|
||||
if not current_name or not new_name:
|
||||
return False, "缺少 current_name / new_name"
|
||||
try:
|
||||
resp = requests.post(
|
||||
f"{self.base_url}/api/oracle/identity/correct",
|
||||
json={"video_id": video_id, "current_name": current_name,
|
||||
"new_name": new_name, "token": self.token},
|
||||
timeout=(10, 30))
|
||||
except requests.RequestException as e:
|
||||
logger.error(f"人物纠错回推失败: {e}")
|
||||
return False, str(e)
|
||||
if resp.status_code == 200:
|
||||
return True, ""
|
||||
msg = f"HTTP {resp.status_code}: {resp.text[:200]}"
|
||||
logger.error(f"人物纠错回推失败: {msg}")
|
||||
return False, msg
|
||||
|
||||
def push_video_delete(self, video_id: int):
|
||||
"""回推事件时间轴"删除该视频"到 Oracle。返回 (success: bool, error: str)。"""
|
||||
try:
|
||||
video_id = int(video_id)
|
||||
except (TypeError, ValueError):
|
||||
return False, "video_id 必须是数字"
|
||||
try:
|
||||
resp = requests.post(
|
||||
f"{self.base_url}/api/oracle/video/delete",
|
||||
json={"video_id": video_id, "token": self.token},
|
||||
timeout=(10, 30))
|
||||
except requests.RequestException as e:
|
||||
logger.error(f"视频删除回推失败: {e}")
|
||||
return False, str(e)
|
||||
if resp.status_code == 200:
|
||||
return True, ""
|
||||
msg = f"HTTP {resp.status_code}: {resp.text[:200]}"
|
||||
logger.error(f"视频删除回推失败: {msg}")
|
||||
return False, msg
|
||||
|
||||
# ------------------------------------------------------------------
|
||||
def _run(self):
|
||||
logger.info(f"OracleSync 线程启动，间隔 {self.interval_sec}s，目标 {self.base_url}")
|
||||
while self._running:
|
||||
try:
|
||||
self._pull_once()
|
||||
except Exception as e:
|
||||
self._last_error = str(e)
|
||||
logger.error(f"同步异常: {e}", exc_info=True)
|
||||
# 分段休眠，便于 stop 快速唤醒
|
||||
for _ in range(self.interval_sec):
|
||||
if not self._running:
|
||||
break
|
||||
time.sleep(1)
|
||||
|
||||
def start(self):
|
||||
if self._running:
|
||||
return
|
||||
self._running = True
|
||||
self._thread = threading.Thread(target=self._run, daemon=True, name='oracle-sync')
|
||||
self._thread.start()
|
||||
|
||||
def is_alive(self):
|
||||
return self._thread is not None and self._thread.is_alive()
|
||||
|
||||
def stop(self):
|
||||
self._running = False
|
||||
if self._thread:
|
||||
self._thread.join(timeout=5)
|
||||
|
||||
def trigger_now(self) -> bool:
|
||||
"""立即触发一次同步（命名后即时拉回 / 手动）。"""
|
||||
return self._pull_once()
|
||||
|
||||
def status(self) -> dict:
|
||||
return {
|
||||
"running": self.is_alive(),
|
||||
"last_sync_at": self._last_sync_at.isoformat() if self._last_sync_at else None,
|
||||
"last_error": self._last_error,
|
||||
"last_count": self._last_count,
|
||||
"cursor": db_layer.get_sync_cursor(),
|
||||
"interval_sec": self.interval_sec,
|
||||
}
|
||||
@@ -7,9 +7,8 @@ UI-API - Vue 前端数据接口（新架构 v3：Vue SPA 取代 Streamlit）
|
||||
<id>`（删除视频会话），因为语义上属于 videos 这个资源，比塞进 member_manager.py
|
||||
更清晰；写法沿用项目里"先回推 Oracle，成功后处理本地状态"的既有模式。
|
||||
|
||||
/api/ui/service-status 需要代理 Oracle 的 /api/oracle/activity（浏览器不直连
|
||||
Oracle，避免 token 暴露），写法照抄 img_proxy.py 的模式：复用 oracle_sync.get_sync()
|
||||
已解析好的 base_url/token，不再单独存一份配置。
|
||||
/api/ui/service-status 代理同机 fam-edge 的 /api/oracle/activity（浏览器不直连
|
||||
fam-edge，避免 token 暴露），走 edge_client 统一读 base_url/token。
|
||||
"""
|
||||
import re
|
||||
from datetime import datetime
|
||||
@@ -19,7 +18,7 @@ from flask import Blueprint, request, jsonify
|
||||
|
||||
from .logger import setup_logger
|
||||
from . import db_layer
|
||||
from .oracle_sync import get_sync
|
||||
from . import edge_client
|
||||
|
||||
logger = setup_logger('fam-core.ui_api')
|
||||
|
||||
@@ -50,7 +49,7 @@ def videos():
|
||||
date_filter = request.args.get('date') or None
|
||||
page = max(0, request.args.get('page', 0, type=int))
|
||||
page_size = 15
|
||||
rows = db_layer.get_sync_videos(limit=page_size, offset=page * page_size,
|
||||
rows = db_layer.get_videos(limit=page_size, offset=page * page_size,
|
||||
date_filter=date_filter)
|
||||
return jsonify({"videos": _ser(rows), "page": page, "page_size": page_size}), 200
|
||||
|
||||
@@ -58,10 +57,10 @@ def videos():
|
||||
@ui_bp.route('/api/ui/videos/<int:video_id>', methods=['GET'])
|
||||
def video_detail(video_id):
|
||||
"""单个视频会话详情 + 时间线事件列表（事件时间轴右侧）。"""
|
||||
video = db_layer.get_sync_video(video_id)
|
||||
video = db_layer.get_video(video_id)
|
||||
if not video:
|
||||
return jsonify({"error": "视频不存在"}), 404
|
||||
events = db_layer.get_sync_events_for_video(video_id)
|
||||
events = db_layer.get_events_for_video(video_id)
|
||||
return jsonify({"video": _ser(video), "events": _ser(events)}), 200
|
||||
|
||||
|
||||
@@ -69,16 +68,13 @@ def video_detail(video_id):
|
||||
def video_delete(video_id):
|
||||
"""删除视频会话（事件时间轴"删除"入口）。
|
||||
|
||||
先回推 Oracle 物理删除（events + videos 行 + 磁盘文件），成功后再清理本地
|
||||
MariaDB 镜像——增量同步（get_sync_delta）只做 upsert 感知不到删除，不能像
|
||||
命名纠错那样靠 trigger_now() 拉增量顺带清理，必须显式调 delete_sync_video。
|
||||
Oracle 回推失败时不清理本地镜像，避免"Oracle 还留着、NAS 却以为删了"的
|
||||
数据不一致，让用户看错误提示后重试。
|
||||
交给 fam-edge 处理：它删 events + videos 行，还要删磁盘上的运动片段文件。
|
||||
迁云前这里还要再清一次 NAS 本地镜像（增量同步只 upsert，感知不到删除），
|
||||
现在没有镜像了，fam-edge 删完就是最终状态。
|
||||
"""
|
||||
ok, err = get_sync().push_video_delete(video_id)
|
||||
ok, err = edge_client.push_video_delete(video_id)
|
||||
if not ok:
|
||||
return jsonify({"error": f"删除失败: {err}"}), 502
|
||||
db_layer.delete_sync_video(video_id)
|
||||
return jsonify({"status": "ok", "video_id": video_id}), 200
|
||||
|
||||
|
||||
@@ -86,7 +82,7 @@ def video_delete(video_id):
|
||||
def stats():
|
||||
"""统计卡：视频/事件/人物/关注数（可选按日期过滤）。"""
|
||||
date_filter = request.args.get('date') or None
|
||||
return jsonify(_ser(db_layer.get_sync_stats(date_filter))), 200
|
||||
return jsonify(_ser(db_layer.get_stats(date_filter))), 200
|
||||
|
||||
|
||||
def _clean_person(s: str) -> str:
|
||||
@@ -100,7 +96,7 @@ def people():
|
||||
|
||||
对应旧 Streamlit 版本 app.py 里的分组逻辑，原样搬到服务端。
|
||||
"""
|
||||
rows = db_layer.get_sync_people()
|
||||
rows = db_layer.get_people()
|
||||
groups = {}
|
||||
for p in rows:
|
||||
key = p.get('canonical_name') or p['label']
|
||||
@@ -128,7 +124,7 @@ def people():
|
||||
def people_clips():
|
||||
"""某人物出现过的运动片段列表（人物卡「运动片段」区块）。
|
||||
|
||||
按 label/canonical_name 匹配 sync_events.person_list_json → 关联视频
|
||||
按 label/canonical_name 匹配 events.person_list_json → 关联视频
|
||||
（运动片段优先）。返回片段 video_id/filename/event_start_time/duration/
|
||||
summary/camera_name + first_ts（缩略图定位）+ clip_events（片段内事件数）。
|
||||
"""
|
||||
@@ -136,7 +132,7 @@ def people_clips():
|
||||
if not label:
|
||||
return jsonify({"error": "缺少 label 参数"}), 400
|
||||
limit = min(20, request.args.get('limit', 10, type=int) or 10)
|
||||
clips = db_layer.get_sync_people_clips(label, limit)
|
||||
clips = db_layer.get_people_clips(label, limit)
|
||||
return jsonify({"label": label, "clips": _ser(clips)}), 200
|
||||
|
||||
|
||||
@@ -163,41 +159,25 @@ def attention_events():
|
||||
@ui_bp.route('/api/ui/named-members', methods=['GET'])
|
||||
def named_members():
|
||||
"""已命名成员真名列表（AI 对话页快捷选择）。"""
|
||||
return jsonify({"members": db_layer.get_sync_named_members()}), 200
|
||||
return jsonify({"members": db_layer.get_named_members()}), 200
|
||||
|
||||
|
||||
@ui_bp.route('/api/ui/model-stats', methods=['GET'])
|
||||
def model_stats():
|
||||
"""云端模型调用统计：按模型聚合 + 最近调用明细。"""
|
||||
agg = db_layer.get_sync_model_calls_stats()
|
||||
calls = db_layer.get_sync_model_calls(limit=100)
|
||||
agg = db_layer.get_model_calls_stats()
|
||||
calls = db_layer.get_model_calls(limit=100)
|
||||
return jsonify({"aggregate": _ser(agg), "recent_calls": _ser(calls)}), 200
|
||||
|
||||
|
||||
@ui_bp.route('/api/ui/service-status', methods=['GET'])
|
||||
def service_status():
|
||||
"""服务状态页：NAS 同步状态 + Oracle 实时活动（代理，token 不下发浏览器）。"""
|
||||
sync = get_sync()
|
||||
nas_status = sync.status()
|
||||
"""服务状态页：fam-edge 的队列/分割/模型活动（代理，token 不下发浏览器）。
|
||||
|
||||
oracle_data = None
|
||||
oracle_error = None
|
||||
if sync.base_url and sync.token:
|
||||
import requests
|
||||
try:
|
||||
r = requests.get(f"{sync.base_url}/api/oracle/activity",
|
||||
params={'token': sync.token}, timeout=15)
|
||||
if r.status_code == 200:
|
||||
oracle_data = r.json()
|
||||
else:
|
||||
oracle_error = f"Oracle activity HTTP {r.status_code}"
|
||||
except Exception as e:
|
||||
oracle_error = f"连接 Oracle 失败: {e}"
|
||||
else:
|
||||
oracle_error = "未配置 oracle_sync.base_url/token"
|
||||
|
||||
return jsonify({
|
||||
"nas_sync": nas_status,
|
||||
"oracle": oracle_data,
|
||||
"oracle_error": oracle_error,
|
||||
}), 200
|
||||
迁云前这里还有一块 "NAS 同步状态"（镜像拉取的游标/周期/上次条数），随镜像层
|
||||
一起删了——现在前端读的就是 fam-edge 写的那份库，没有"同步"这个中间状态。
|
||||
NAS 侧只剩运动事件推送，它的心跳在 fam-edge 的 service_activity 里，
|
||||
已经包含在 activity 快照中。
|
||||
"""
|
||||
data, err = edge_client.get_activity()
|
||||
return jsonify({"oracle": data, "oracle_error": err}), 200
|
||||
|
||||
Reference in New Issue
Block a user