54 lines
1.7 KiB
Python
54 lines
1.7 KiB
Python
"""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 recognize_batch(self, urls: list[dict[str, str]], country: str = "US") -> list[OcrResult]:
|
|
"""v2 批量识别:urls=[{"url":"...", "filename":"orderNo"}], 最多 200 项/批。
|
|
|
|
返回 OcrResult 列表,file 字段 = filename(用于回匹配 orderNo)。
|
|
"""
|
|
resp = self._http.post("/api/v2/batch", json={"urls": urls, "country": country}, timeout=300.0)
|
|
resp.raise_for_status()
|
|
data = resp.json()
|
|
return [OcrResult.model_validate(r) for r in data.get("results", [])]
|
|
|
|
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()
|