feat(api): add web search and geocoding

app/api/v1/endpoints/geocoding.py

@@ -0,0 +1,29 @@
+from typing import Any
+
+from fastapi import APIRouter, HTTPException, status
+
+from app.services.geocoding import (
+    TiandituGeocodeRequest,
+    TiandituGeocodingAPIError,
+    TiandituGeocodingConfigError,
+    geocode_tianditu,
+)
+
+router = APIRouter()
+
+
+@router.post(
+    "/tianditu/geocode",
+    summary="Tianditu Geocoding",
+    description="调用天地图地理编码服务，将结构化地址转换为经纬度",
+)
+async def tianditu_geocode(request: TiandituGeocodeRequest) -> dict[str, Any]:
+    try:
+        return await geocode_tianditu(request)
+    except TiandituGeocodingConfigError as exc:
+        raise HTTPException(
+            status_code=status.HTTP_503_SERVICE_UNAVAILABLE,
+            detail=str(exc),
+        ) from exc
+    except TiandituGeocodingAPIError as exc:
+        raise HTTPException(status_code=exc.status_code, detail=exc.detail) from exc

app/api/v1/endpoints/web_search.py

@@ -0,0 +1,29 @@
+from typing import Any
+
+from fastapi import APIRouter, HTTPException, status
+
+from app.services.web_search import (
+    BochaSearchAPIError,
+    BochaSearchConfigError,
+    WebSearchRequest,
+    search_bocha_web,
+)
+
+router = APIRouter()
+
+
+@router.post(
+    "/web-search",
+    summary="Web Search",
+    description="调用 Bocha Web Search API 获取实时网页搜索结果",
+)
+async def web_search(request: WebSearchRequest) -> dict[str, Any]:
+    try:
+        return await search_bocha_web(request)
+    except BochaSearchConfigError as exc:
+        raise HTTPException(
+            status_code=status.HTTP_503_SERVICE_UNAVAILABLE,
+            detail=str(exc),
+        ) from exc
+    except BochaSearchAPIError as exc:
+        raise HTTPException(status_code=exc.status_code, detail=exc.detail) from exc

app/api/v1/router.py

@@ -17,6 +17,8 @@ from app.api.v1.endpoints import (
     user_management,  # 新增：用户管理
     audit,  # 新增：审计日志
     meta,
+    web_search,
+    geocoding,
 )
 from app.api.v1.endpoints.network import (
     general,
@@ -89,6 +91,8 @@ api_router.include_router(schemes.router, tags=["Schemes"])
 api_router.include_router(misc.router, tags=["Misc"])
 api_router.include_router(risk.router, tags=["Risk"])
 api_router.include_router(cache.router, tags=["Cache"])
+api_router.include_router(web_search.router, tags=["Web Search"])
+api_router.include_router(geocoding.router, tags=["Geocoding"])
 api_router.include_router(leakage.router, prefix="/leakage", tags=["Leakage"])
 api_router.include_router(
     burst_detection.router, prefix="/burst-detection", tags=["Burst Detection"]

app/core/config.py

@@ -57,6 +57,16 @@ class Settings(BaseSettings):
     KEYCLOAK_ALGORITHM: str = "RS256"
     KEYCLOAK_AUDIENCE: str = ""
 
+    # Bocha Web Search API
+    BOCHA_API_KEY: str = ""
+    BOCHA_WEB_SEARCH_URL: str = "https://api.bochaai.com/v1/web-search"
+    BOCHA_WEB_SEARCH_TIMEOUT_SECONDS: float = 30.0
+
+    # Tianditu Geocoding API
+    TIANDITU_GEOCODER_TOKEN: str = ""
+    TIANDITU_GEOCODER_URL: str = "https://api.tianditu.gov.cn/geocoder"
+    TIANDITU_GEOCODER_TIMEOUT_SECONDS: float = 30.0
+
     @property
     def SQLALCHEMY_DATABASE_URI(self) -> str:
         db_password = quote_plus(self.DB_PASSWORD)

app/services/geocoding.py

@@ -0,0 +1,76 @@
+import json
+from typing import Any
+
+import httpx
+from pydantic import AliasChoices, BaseModel, Field
+
+from app.core.config import settings
+
+
+class TiandituGeocodeRequest(BaseModel):
+    keyword: str = Field(
+        ...,
+        min_length=1,
+        validation_alias=AliasChoices("keyword", "keyWord"),
+        description="地理编码地址关键字",
+    )
+
+
+class TiandituGeocodingConfigError(RuntimeError):
+    pass
+
+
+class TiandituGeocodingAPIError(RuntimeError):
+    def __init__(self, status_code: int, detail: Any):
+        super().__init__("Tianditu Geocoding API request failed")
+        self.status_code = status_code
+        self.detail = detail
+
+
+async def geocode_tianditu(
+    request: TiandituGeocodeRequest,
+    *,
+    client: httpx.AsyncClient | None = None,
+) -> dict[str, Any]:
+    if not settings.TIANDITU_GEOCODER_TOKEN:
+        raise TiandituGeocodingConfigError("TIANDITU_GEOCODER_TOKEN is not configured")
+
+    params = {
+        "ds": json.dumps({"keyWord": request.keyword}, ensure_ascii=False),
+        "tk": settings.TIANDITU_GEOCODER_TOKEN,
+    }
+
+    if client is not None:
+        response = await client.get(settings.TIANDITU_GEOCODER_URL, params=params)
+        return _parse_response(response)
+
+    async with httpx.AsyncClient(
+        timeout=settings.TIANDITU_GEOCODER_TIMEOUT_SECONDS
+    ) as managed_client:
+        response = await managed_client.get(
+            settings.TIANDITU_GEOCODER_URL,
+            params=params,
+        )
+        return _parse_response(response)
+
+
+def _parse_response(response: httpx.Response) -> dict[str, Any]:
+    try:
+        response.raise_for_status()
+    except httpx.HTTPStatusError as exc:
+        raise TiandituGeocodingAPIError(
+            exc.response.status_code,
+            _response_detail(exc.response),
+        ) from exc
+
+    data = response.json()
+    if str(data.get("status")) != "0":
+        raise TiandituGeocodingAPIError(502, data)
+    return data
+
+
+def _response_detail(response: httpx.Response) -> Any:
+    try:
+        return response.json()
+    except ValueError:
+        return response.text

app/services/web_search.py

@@ -0,0 +1,93 @@
+from typing import Any, Literal
+
+import httpx
+from pydantic import BaseModel, Field
+
+from app.core.config import settings
+
+
+Freshness = Literal["noLimit", "oneDay", "oneWeek", "oneMonth", "oneYear"]
+
+
+class WebSearchRequest(BaseModel):
+    query: str = Field(..., min_length=1, description="搜索关键词")
+    freshness: Freshness | str = Field(
+        default="noLimit",
+        description="时间范围：noLimit、oneDay、oneWeek、oneMonth、oneYear 或日期范围",
+    )
+    summary: bool = Field(default=True, description="是否返回网页摘要")
+    count: int = Field(default=10, ge=1, le=50, description="返回结果数量")
+    include: list[str] | None = Field(default=None, description="限定搜索域名")
+    exclude: list[str] | None = Field(default=None, description="排除搜索域名")
+
+
+class BochaSearchConfigError(RuntimeError):
+    pass
+
+
+class BochaSearchAPIError(RuntimeError):
+    def __init__(self, status_code: int, detail: Any):
+        super().__init__("Bocha Web Search API request failed")
+        self.status_code = status_code
+        self.detail = detail
+
+
+def _build_payload(request: WebSearchRequest) -> dict[str, Any]:
+    payload = request.model_dump(exclude_none=True)
+    if request.include:
+        payload["include"] = ",".join(request.include)
+    if request.exclude:
+        payload["exclude"] = ",".join(request.exclude)
+    return payload
+
+
+async def search_bocha_web(
+    request: WebSearchRequest,
+    *,
+    client: httpx.AsyncClient | None = None,
+) -> dict[str, Any]:
+    if not settings.BOCHA_API_KEY:
+        raise BochaSearchConfigError("BOCHA_API_KEY is not configured")
+
+    headers = {
+        "Authorization": f"Bearer {settings.BOCHA_API_KEY}",
+        "Content-Type": "application/json",
+    }
+    payload = _build_payload(request)
+
+    if client is not None:
+        response = await client.post(
+            settings.BOCHA_WEB_SEARCH_URL,
+            headers=headers,
+            json=payload,
+        )
+        return _parse_response(response)
+
+    async with httpx.AsyncClient(
+        timeout=settings.BOCHA_WEB_SEARCH_TIMEOUT_SECONDS
+    ) as managed_client:
+        response = await managed_client.post(
+            settings.BOCHA_WEB_SEARCH_URL,
+            headers=headers,
+            json=payload,
+        )
+        return _parse_response(response)
+
+
+def _parse_response(response: httpx.Response) -> dict[str, Any]:
+    try:
+        response.raise_for_status()
+    except httpx.HTTPStatusError as exc:
+        raise BochaSearchAPIError(
+            exc.response.status_code,
+            _response_detail(exc.response),
+        ) from exc
+
+    return response.json()
+
+
+def _response_detail(response: httpx.Response) -> Any:
+    try:
+        return response.json()
+    except ValueError:
+        return response.text