feat: OCR 面单 enrichment flow (api/ocr/ + flows/sds/label_ocr.py + db OCR 方法)

This commit is contained in:
yeuimu
2026-07-23 16:15:24 +00:00
parent 8ab0aededa
commit 51dbde4538
6 changed files with 266 additions and 0 deletions
+43
View File
@@ -0,0 +1,43 @@
"""OCR 面单识别 Inference API 客户端(httpx,无鉴权)。
端点: POST /recognize_url JSON {url, country}
"""
from __future__ import annotations
import httpx
from api.ocr.models import OcrResult
DEFAULT_BASE_URL = "http://192.168.124.92:9801"
DEFAULT_TIMEOUT = 60.0
class OcrClient:
"""OCR 推理服务客户端。"""
def __init__(self, base_url: str = DEFAULT_BASE_URL, timeout: float = DEFAULT_TIMEOUT) -> None:
self._http = httpx.Client(base_url=base_url, timeout=timeout)
def recognize_url(self, url: str, country: str = "US") -> OcrResult:
"""通过 PDF URL 识别面单,返回 OcrResult。"""
resp = self._http.post("/recognize_url", json={"url": url, "country": country})
resp.raise_for_status()
return OcrResult.model_validate(resp.json())
def health(self) -> bool:
"""健康检查。"""
try:
resp = self._http.get("/health")
return resp.status_code == 200
except Exception: # noqa: BLE001
return False
def close(self) -> None:
self._http.close()
def __enter__(self) -> "OcrClient":
return self
def __exit__(self, *exc: object) -> None:
self.close()