feat(startup): 服务端启动 bootstrap 与缓存预热

- 新增 STARTUP_BOOTSTRAP_ENABLED / STARTUP_CACHE_WARMUP_ENABLED 配置开关 - lifespan 拆分 bootstrap 步骤并后台线程预热缓存，失败可降级继续启动 - server_start.sh / web_start.sh 扩展 SERVER_PORT、启动与调度开关的 env 覆盖 - bootstrap_paddleocr_mobile.sh 改用 python3 并补充 poppler-utils 依赖 - 补充启动 bootstrap 与 env 覆盖优先级测试
2026-06-18 22:11:37 +08:00
parent 35372c6661
commit 59ba76c74a
7 changed files with 236 additions and 8 deletions
--- a/server/scripts/bootstrap_paddleocr_mobile.sh
+++ b/server/scripts/bootstrap_paddleocr_mobile.sh
@@ -3,17 +3,17 @@ set -euo pipefail

 ROOT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
 OCR_VENV_DIR="${ROOT_DIR}/.venv-ocr312"
-PYTHON_BIN="${PYTHON_BIN:-python3.12}"
-PADDLEPADDLE_VERSION="${PADDLEPADDLE_VERSION:-3.3.1}"
+PYTHON_BIN="${PYTHON_BIN:-python3}"
+PADDLEPADDLE_VERSION="${PADDLEPADDLE_VERSION:-3.2.2}"
 PADDLEOCR_VERSION="${PADDLEOCR_VERSION:-3.6.0}"

 if ! command -v "${PYTHON_BIN}" >/dev/null 2>&1; then
-  echo "python3.12 不存在，请先安装 Python 3.12。" >&2
+  echo "${PYTHON_BIN} 不存在，请先安装 Python 3。" >&2
  exit 1
 fi

 apt-get update
-apt-get install -y libgl1 libglib2.0-0
+apt-get install -y --no-install-recommends libgl1 libglib2.0-0 poppler-utils

 "${PYTHON_BIN}" -m venv "${OCR_VENV_DIR}"
 "${OCR_VENV_DIR}/bin/pip" install --upgrade pip
--- a/server/server_start.sh
+++ b/server/server_start.sh
@@ -88,8 +88,11 @@ if [ ! -f "$ROOT_ENV_FILE" ]; then
 fi

 ENV_OVERRIDE_SERVER_HOST_SET=false
+ENV_OVERRIDE_SERVER_PORT_SET=false
 ENV_OVERRIDE_POSTGRES_HOST_SET=false
 ENV_OVERRIDE_DATABASE_URL_SET=false
+ENV_OVERRIDE_STARTUP_BOOTSTRAP_ENABLED_SET=false
+ENV_OVERRIDE_BACKGROUND_SCHEDULERS_ENABLED_SET=false
 ENV_OVERRIDE_ONLYOFFICE_ENABLED_SET=false
 ENV_OVERRIDE_ONLYOFFICE_PUBLIC_URL_SET=false
 ENV_OVERRIDE_ONLYOFFICE_BACKEND_URL_SET=false
@@ -107,6 +110,11 @@ if [ "${SERVER_HOST+x}" = x ]; then
    ENV_OVERRIDE_SERVER_HOST="$SERVER_HOST"
 fi

+if [ "${SERVER_PORT+x}" = x ]; then
+    ENV_OVERRIDE_SERVER_PORT_SET=true
+    ENV_OVERRIDE_SERVER_PORT="$SERVER_PORT"
+fi
+
 if [ "${POSTGRES_HOST+x}" = x ]; then
    ENV_OVERRIDE_POSTGRES_HOST_SET=true
    ENV_OVERRIDE_POSTGRES_HOST="$POSTGRES_HOST"
@@ -117,6 +125,16 @@ if [ "${DATABASE_URL+x}" = x ]; then
    ENV_OVERRIDE_DATABASE_URL="$DATABASE_URL"
 fi

+if [ "${STARTUP_BOOTSTRAP_ENABLED+x}" = x ]; then
+    ENV_OVERRIDE_STARTUP_BOOTSTRAP_ENABLED_SET=true
+    ENV_OVERRIDE_STARTUP_BOOTSTRAP_ENABLED="$STARTUP_BOOTSTRAP_ENABLED"
+fi
+
+if [ "${BACKGROUND_SCHEDULERS_ENABLED+x}" = x ]; then
+    ENV_OVERRIDE_BACKGROUND_SCHEDULERS_ENABLED_SET=true
+    ENV_OVERRIDE_BACKGROUND_SCHEDULERS_ENABLED="$BACKGROUND_SCHEDULERS_ENABLED"
+fi
+
 if [ "$PREFER_ENV_FILE_FOR_ONLYOFFICE" != true ] && [ "${ONLYOFFICE_ENABLED+x}" = x ]; then
    ENV_OVERRIDE_ONLYOFFICE_ENABLED_SET=true
    ENV_OVERRIDE_ONLYOFFICE_ENABLED="$ONLYOFFICE_ENABLED"
@@ -145,6 +163,10 @@ if [ "$ENV_OVERRIDE_SERVER_HOST_SET" = true ]; then
    SERVER_HOST="$ENV_OVERRIDE_SERVER_HOST"
 fi

+if [ "$ENV_OVERRIDE_SERVER_PORT_SET" = true ]; then
+    SERVER_PORT="$ENV_OVERRIDE_SERVER_PORT"
+fi
+
 if [ "$ENV_OVERRIDE_POSTGRES_HOST_SET" = true ]; then
    POSTGRES_HOST="$ENV_OVERRIDE_POSTGRES_HOST"
 fi
@@ -153,6 +175,14 @@ if [ "$ENV_OVERRIDE_DATABASE_URL_SET" = true ]; then
    DATABASE_URL="$ENV_OVERRIDE_DATABASE_URL"
 fi

+if [ "$ENV_OVERRIDE_STARTUP_BOOTSTRAP_ENABLED_SET" = true ]; then
+    STARTUP_BOOTSTRAP_ENABLED="$ENV_OVERRIDE_STARTUP_BOOTSTRAP_ENABLED"
+fi
+
+if [ "$ENV_OVERRIDE_BACKGROUND_SCHEDULERS_ENABLED_SET" = true ]; then
+    BACKGROUND_SCHEDULERS_ENABLED="$ENV_OVERRIDE_BACKGROUND_SCHEDULERS_ENABLED"
+fi
+
 if [ "$ENV_OVERRIDE_ONLYOFFICE_ENABLED_SET" = true ]; then
    ONLYOFFICE_ENABLED="$ENV_OVERRIDE_ONLYOFFICE_ENABLED"
 fi
--- a/server/src/app/core/config.py
+++ b/server/src/app/core/config.py
@@ -43,6 +43,8 @@ class Settings(BaseSettings):
    app_port: int = Field(default=8000, alias="SERVER_PORT")
    server_workers: int = Field(default=1, alias="SERVER_WORKERS")
    web_concurrency: int | None = Field(default=None, alias="WEB_CONCURRENCY")
+    startup_bootstrap_enabled: bool = Field(default=True, alias="STARTUP_BOOTSTRAP_ENABLED")
+    startup_cache_warmup_enabled: bool = Field(default=False, alias="STARTUP_CACHE_WARMUP_ENABLED")
    background_schedulers_enabled: bool = Field(
        default=True,
        alias="BACKGROUND_SCHEDULERS_ENABLED",
--- a/server/src/app/main.py
+++ b/server/src/app/main.py
@@ -2,6 +2,8 @@ from __future__ import annotations

 from collections.abc import AsyncIterator
 from contextlib import asynccontextmanager
+from logging import Logger
+import threading

 from fastapi import FastAPI
 from fastapi.middleware.cors import CORSMiddleware
@@ -10,11 +12,13 @@ from app.api.router import api_router
 from app.core.config import get_settings
 from app.core.logging import get_logger, setup_logging
 from app.core.openapi import API_DESCRIPTION, OPENAPI_TAGS
+from app.db.session import get_session_factory
 from app.middleware.logging import AccessLogMiddleware
 from app.schemas.common import RootStatusRead
 from app.services.agent_foundation import prepare_agent_foundation
 from app.services.digital_employee_reminder_scheduler import digital_employee_reminder_scheduler
 from app.services.employee import prepare_employee_directory
+from app.services.employee import EmployeeService
 from app.services.employee_profile_scheduler import employee_profile_scheduler
 from app.services.finance_dashboard_scheduler import finance_dashboard_scheduler
 from app.services.finance_report_scheduler import finance_report_scheduler
@@ -23,6 +27,8 @@ from app.services.knowledge import prepare_knowledge_library
 from app.services.knowledge_index_tasks import knowledge_index_task_manager
 from app.services.knowledge_rag import shutdown_knowledge_rag_runtime
 from app.services.knowledge_scheduler import knowledge_index_scheduler
+from app.services.settings import SettingsService
+from app.services.user_session_metrics import UserSessionMetricService


 def _effective_server_workers(settings: object) -> int:
@@ -42,15 +48,55 @@ def _should_start_background_schedulers(settings: object) -> bool:
    return _effective_server_workers(settings) <= 1


+def _run_startup_bootstrap(logger: Logger) -> None:
+    steps = (
+        ("employee_directory", prepare_employee_directory),
+        ("agent_foundation", prepare_agent_foundation),
+        ("knowledge_library", prepare_knowledge_library),
+        ("hermes_skills", sync_repository_hermes_skills),
+    )
+    for name, step in steps:
+        try:
+            step()
+        except Exception:
+            logger.exception("Startup bootstrap step failed; continuing degraded name=%s", name)
+
+
+def _warm_startup_caches(logger: Logger) -> None:
+    try:
+        session_factory = get_session_factory()
+        with session_factory() as db:
+            SettingsService(db).ensure_settings_ready()
+            EmployeeService(db).ensure_directory_ready()
+            UserSessionMetricService(db).ensure_storage_ready()
+        logger.info("Startup cache warmup complete")
+    except Exception:
+        logger.exception("Startup cache warmup failed; continuing without warm cache")
+
+
+def _start_cache_warmup_thread(logger: Logger) -> None:
+    thread = threading.Thread(
+        target=_warm_startup_caches,
+        args=(logger,),
+        name="x-financial-startup-cache-warmup",
+        daemon=True,
+    )
+    thread.start()
+
+
@asynccontextmanager
 async def lifespan(_: FastAPI) -> AsyncIterator[None]:
    settings = get_settings()
    logger = get_logger("app.main")

-    prepare_employee_directory()
-    prepare_agent_foundation()
-    prepare_knowledge_library()
-    sync_repository_hermes_skills()
+    if settings.startup_bootstrap_enabled:
+        _run_startup_bootstrap(logger)
+    else:
+        logger.warning("Startup bootstrap skipped because STARTUP_BOOTSTRAP_ENABLED=false")
+
+    if settings.startup_cache_warmup_enabled:
+        _start_cache_warmup_thread(logger)
+
    schedulers_started = _should_start_background_schedulers(settings)
    if schedulers_started:
        knowledge_index_scheduler.start()
--- a/server/tests/test_env_file_precedence.py
+++ b/server/tests/test_env_file_precedence.py
@@ -79,3 +79,63 @@ def test_root_start_can_prefer_env_file_over_inherited_onlyoffice_values(tmp_pat
    assert result.returncode == 0, result.stderr
    assert "ONLYOFFICE_ENABLED=true" in result.stdout
    assert "ONLYOFFICE_PUBLIC_URL=http://10.10.10.122:8082" in result.stdout
+
+
+def test_web_start_preserves_inherited_runtime_ports(tmp_path: Path) -> None:
+    result = _run_script_prefix(
+        tmp_path,
+        "web/web_start.sh",
+        env_file_content=(
+            "WEB_HOST=10.10.10.122\n"
+            "WEB_PORT=5273\n"
+            "SERVER_HOST=10.10.10.122\n"
+            "SERVER_PORT=9000\n"
+            "POSTGRES_HOST=10.10.10.189\n"
+        ),
+        env={
+            "WEB_HOST": "0.0.0.0",
+            "WEB_PORT": "5173",
+            "SERVER_HOST": "0.0.0.0",
+            "SERVER_PORT": "8000",
+            "POSTGRES_HOST": "www.caoxiaozhu.com",
+        },
+        output_vars=["WEB_HOST", "WEB_PORT", "SERVER_HOST", "SERVER_PORT", "POSTGRES_HOST"],
+    )
+
+    assert result.returncode == 0, result.stderr
+    assert "WEB_HOST=0.0.0.0" in result.stdout
+    assert "WEB_PORT=5173" in result.stdout
+    assert "SERVER_HOST=0.0.0.0" in result.stdout
+    assert "SERVER_PORT=8000" in result.stdout
+    assert "POSTGRES_HOST=www.caoxiaozhu.com" in result.stdout
+
+
+def test_server_start_preserves_inherited_runtime_guards(tmp_path: Path) -> None:
+    result = _run_script_prefix(
+        tmp_path,
+        "server/server_start.sh",
+        env_file_content=(
+            "SERVER_HOST=10.10.10.122\n"
+            "SERVER_PORT=9000\n"
+            "STARTUP_BOOTSTRAP_ENABLED=true\n"
+            "BACKGROUND_SCHEDULERS_ENABLED=true\n"
+        ),
+        env={
+            "SERVER_HOST": "0.0.0.0",
+            "SERVER_PORT": "8000",
+            "STARTUP_BOOTSTRAP_ENABLED": "false",
+            "BACKGROUND_SCHEDULERS_ENABLED": "false",
+        },
+        output_vars=[
+            "SERVER_HOST",
+            "SERVER_PORT",
+            "STARTUP_BOOTSTRAP_ENABLED",
+            "BACKGROUND_SCHEDULERS_ENABLED",
+        ],
+    )
+
+    assert result.returncode == 0, result.stderr
+    assert "SERVER_HOST=0.0.0.0" in result.stdout
+    assert "SERVER_PORT=8000" in result.stdout
+    assert "STARTUP_BOOTSTRAP_ENABLED=false" in result.stdout
+    assert "BACKGROUND_SCHEDULERS_ENABLED=false" in result.stdout
--- a/server/tests/test_startup_bootstrap.py
+++ b/server/tests/test_startup_bootstrap.py
@@ -0,0 +1,70 @@
+from __future__ import annotations
+
+from fastapi.testclient import TestClient
+
+from app.core.config import get_settings
+import app.main as main_module
+
+
+def _set_bootstrap_step(
+    monkeypatch,
+    name: str,
+    calls: list[str],
+    *,
+    should_fail: bool = False,
+) -> None:
+    def step() -> None:
+        calls.append(name)
+        if should_fail:
+            raise RuntimeError(f"{name} failed")
+
+    monkeypatch.setattr(main_module, name, step)
+
+
+def test_lifespan_can_skip_startup_bootstrap(monkeypatch) -> None:
+    calls: list[str] = []
+    monkeypatch.setenv("STARTUP_BOOTSTRAP_ENABLED", "false")
+    monkeypatch.setenv("BACKGROUND_SCHEDULERS_ENABLED", "false")
+    get_settings.cache_clear()
+
+    for name in (
+        "prepare_employee_directory",
+        "prepare_agent_foundation",
+        "prepare_knowledge_library",
+        "sync_repository_hermes_skills",
+    ):
+        _set_bootstrap_step(monkeypatch, name, calls)
+
+    try:
+        with TestClient(main_module.create_app()) as client:
+            response = client.get("/")
+    finally:
+        get_settings.cache_clear()
+
+    assert response.status_code == 200
+    assert calls == []
+
+
+def test_lifespan_continues_when_startup_bootstrap_fails(monkeypatch) -> None:
+    calls: list[str] = []
+    monkeypatch.setenv("STARTUP_BOOTSTRAP_ENABLED", "true")
+    monkeypatch.setenv("BACKGROUND_SCHEDULERS_ENABLED", "false")
+    get_settings.cache_clear()
+
+    steps: tuple[tuple[str, bool], ...] = (
+        ("prepare_employee_directory", True),
+        ("prepare_agent_foundation", False),
+        ("prepare_knowledge_library", False),
+        ("sync_repository_hermes_skills", False),
+    )
+    for name, should_fail in steps:
+        _set_bootstrap_step(monkeypatch, name, calls, should_fail=should_fail)
+
+    try:
+        with TestClient(main_module.create_app()) as client:
+            response = client.get("/")
+    finally:
+        get_settings.cache_clear()
+
+    assert response.status_code == 200
+    assert calls == [name for name, _ in steps]
--- a/web/web_start.sh
+++ b/web/web_start.sh
@@ -29,7 +29,9 @@ error() { printf '%b\n' "${RED}[ERROR]${NC} $*"; exit 1; }

 if [ -f "$ROOT_ENV_FILE" ]; then
    ENV_OVERRIDE_WEB_HOST_SET=false
+    ENV_OVERRIDE_WEB_PORT_SET=false
    ENV_OVERRIDE_SERVER_HOST_SET=false
+    ENV_OVERRIDE_SERVER_PORT_SET=false
    ENV_OVERRIDE_POSTGRES_HOST_SET=false

    if [ "${WEB_HOST+x}" = x ]; then
@@ -37,11 +39,21 @@ if [ -f "$ROOT_ENV_FILE" ]; then
        ENV_OVERRIDE_WEB_HOST="$WEB_HOST"
    fi

+    if [ "${WEB_PORT+x}" = x ]; then
+        ENV_OVERRIDE_WEB_PORT_SET=true
+        ENV_OVERRIDE_WEB_PORT="$WEB_PORT"
+    fi
+
    if [ "${SERVER_HOST+x}" = x ]; then
        ENV_OVERRIDE_SERVER_HOST_SET=true
        ENV_OVERRIDE_SERVER_HOST="$SERVER_HOST"
    fi

+    if [ "${SERVER_PORT+x}" = x ]; then
+        ENV_OVERRIDE_SERVER_PORT_SET=true
+        ENV_OVERRIDE_SERVER_PORT="$SERVER_PORT"
+    fi
+
    if [ "${POSTGRES_HOST+x}" = x ]; then
        ENV_OVERRIDE_POSTGRES_HOST_SET=true
        ENV_OVERRIDE_POSTGRES_HOST="$POSTGRES_HOST"
@@ -55,10 +67,18 @@ if [ -f "$ROOT_ENV_FILE" ]; then
        WEB_HOST="$ENV_OVERRIDE_WEB_HOST"
    fi

+    if [ "$ENV_OVERRIDE_WEB_PORT_SET" = true ]; then
+        WEB_PORT="$ENV_OVERRIDE_WEB_PORT"
+    fi
+
    if [ "$ENV_OVERRIDE_SERVER_HOST_SET" = true ]; then
        SERVER_HOST="$ENV_OVERRIDE_SERVER_HOST"
    fi

+    if [ "$ENV_OVERRIDE_SERVER_PORT_SET" = true ]; then
+        SERVER_PORT="$ENV_OVERRIDE_SERVER_PORT"
+    fi
+
    if [ "$ENV_OVERRIDE_POSTGRES_HOST_SET" = true ]; then
        POSTGRES_HOST="$ENV_OVERRIDE_POSTGRES_HOST"
    fi