inkreachhub/inkreach-prefect
4
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

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

Browse Source
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
Download Patch File Download Diff File Expand all files Collapse all files
api/ocr/__init__.py
+9
View File
@@ -0,0 +1,9 @@
"""OCR 面单识别服务客户端(无鉴权内部服务)。
Inference API: http://192.168.124.92:9801
"""
from api.ocr.client import OcrClient
from api.ocr.models import OcrResult, SUPPORTED_COUNTRIES, country_to_ocr_code
__all__ = ["OcrClient", "OcrResult", "SUPPORTED_COUNTRIES", "country_to_ocr_code"]
api/ocr/client.py
+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()
api/ocr/models.py
+56
View File
@@ -0,0 +1,56 @@
"""OCR 面单识别响应模型 + 国家代码映射。"""
from __future__ import annotations
from typing import Any
from pydantic import BaseModel, ConfigDict
class OcrResult(BaseModel):
"""单张面单 OCR 识别结果。"""
model_config = ConfigDict(extra="allow")
file: str | None = None
layout: str | None = None
courier: str | None = None
courier_name: str | None = None
country: str | None = None
error: str | None = None
tracking: str | None = None
s_name: str | None = None
s_street: str | None = None
s_city: str | None = None
s_state: str | None = None
s_zip: str | None = None
r_name: str | None = None
r_street: str | None = None
r_city: str | None = None
r_state: str | None = None
r_zip: str | None = None
marker: str | None = None
@property
def ok(self) -> bool:
return not self.error
# OCR 服务已有分类器模型的国家(版面见 api.md §4)
SUPPORTED_COUNTRIES = {"US", "UK", "DE", "MX", "PL"}
# ISO 国家代码 / 中文国名 → OCR 国家代码
_COUNTRY_MAP: dict[str, str] = {
# 代码
"GB": "UK", "UK": "UK", "US": "US", "DE": "DE", "MX": "MX", "PL": "PL",
# 中文名(SDS countryName 可能用)
"英国": "UK", "美国": "US", "德国": "DE", "墨西哥": "MX", "波兰": "PL",
}
def country_to_ocr_code(country: Any) -> str | None:
"""将 ISO 代码或中文国名转为 OCR 服务支持的国家代码,不支持则返回 None。"""
if not country or not isinstance(country, str):
return None
code = _COUNTRY_MAP.get(country.strip().upper()) or _COUNTRY_MAP.get(country.strip())
return code if code in SUPPORTED_COUNTRIES else None
flows/sds/label_ocr.py
+98
View File
@@ -0,0 +1,98 @@
"""SDS 面单 OCR enrichment:对已发货的 SDS 订单,用面单 PDF URL 调 OCR 服务回填标签字段。
对应 sds-label-ocr。独立于同步 flow,每 4 小时跑一批。
"""
from __future__ import annotations
from concurrent.futures import ThreadPoolExecutor, as_completed
from prefect import flow, task
from api.ocr import OcrClient, country_to_ocr_code
from flows.shared import SyncDB, build_stats_rows, emit_sync_summary
OCR_CONCURRENCY = 5
SHIPPED_STATUSES = ["已完成", "部分发货", "备货中", "待确认"]
_OCR_FIELD_MAP = [
("layout", "ocr_layout"), ("courier", "ocr_courier"), ("tracking", "ocr_tracking"),
("marker", "ocr_marker"), ("s_name", "ocr_s_name"), ("s_street", "ocr_s_street"),
("s_city", "ocr_s_city"), ("s_state", "ocr_s_state"), ("s_zip", "ocr_s_zip"),
("r_name", "ocr_r_name"), ("r_street", "ocr_r_street"), ("r_city", "ocr_r_city"),
("r_state", "ocr_r_state"), ("r_zip", "ocr_r_zip"),
]
def _process_one(db: SyncDB, ocr: OcrClient, order_no: str, created_at, pdf_url: str, dest_country: str | None) -> str:
country_code = country_to_ocr_code(dest_country)
if not country_code:
db.update_ocr_fields(order_no, created_at, {"ocr_layout": "UNSUPPORTED"})
return "unsupported"
try:
r = ocr.recognize_url(pdf_url, country_code)
except Exception as err: # noqa: BLE001
print(f"[OCR] {order_no} 请求失败: {str(err)[:100]}")
db.update_ocr_fields(order_no, created_at, {"ocr_layout": "ERROR"})
return "error"
if not r.ok:
print(f"[OCR] {order_no} 识别失败: {(r.error or '')[:80]}")
db.update_ocr_fields(order_no, created_at, {"ocr_layout": "ERROR"})
return "error"
fields = {col: getattr(r, attr) or "" for attr, col in _OCR_FIELD_MAP}
db.update_ocr_fields(order_no, created_at, fields)
return "success"
@task
def process_batch(pending: list) -> dict[str, int]:
db = SyncDB()
ocr = OcrClient()
stats = {"success": 0, "error": 0, "unsupported": 0}
try:
done = 0
total = len(pending)
with ThreadPoolExecutor(max_workers=min(OCR_CONCURRENCY, total)) as pool:
futures = {
pool.submit(_process_one, db, ocr, no, ca, url, dc): no
for no, ca, url, dc in pending
}
for fut in as_completed(futures):
result = fut.result()
stats[result] += 1
done += 1
if done % 50 == 0 or done == total:
print(f"[OCR] 进度: {done}/{total} 成功 {stats['success']} / 失败 {stats['error']} / 不支持 {stats['unsupported']}")
stats["processed"] = total
return stats
finally:
ocr.close()
@flow(name="sds-label-ocr", log_prints=True)
def sds_label_ocr(batch_size: int = 200) -> dict:
db = SyncDB()
before = db.count_ocr_status()
print(f"OCR 状态(开始): {before}")
pending = db.get_ocr_pending(batch_size)
print(f"待 OCR: {len(pending)} 单 (batch_size={batch_size})")
if not pending:
emit_sync_summary("SDS 面单 OCR", build_stats_rows("SDS", {}, [
("待处理", 0), ("累计已完成", before.get("done", 0)), ("累计失败", before.get("error", 0)),
]))
return {"processed": 0, **before}
stats = process_batch(pending)
after = db.count_ocr_status()
print(f"完成: {stats} 累计: {after}")
emit_sync_summary("SDS 面单 OCR", build_stats_rows("SDS", {}, [
("本批处理", stats["processed"]), ("本批成功", stats["success"]),
("本批失败", stats["error"]), ("本批不支持", stats["unsupported"]),
("累计已完成", after.get("done", 0)), ("累计待处理", after.get("pending", 0)),
]))
return {**stats, **after}
if __name__ == "__main__":
sds_label_ocr.serve(name="sds-label-ocr", cron="0 */4 * * *")
flows/shared/db.py
+53
View File
@@ -98,6 +98,59 @@ class SyncDB:
rows = cur.fetchall()
return [(r[0], r[1]) for r in rows]
# ---------- OCR 面单 enrichment ----------
def get_ocr_pending(self, limit: int = 200) -> list[tuple[str, datetime, str, str | None]]:
"""查待 OCR 的 SDS 订单:(orderNo, createdAt, labelPdfUrl, destinationCountry)。
条件:SDS、有 labelPdfUrl、ocr_layout 未处理(Null)、已发货/已完成。
"""
with self._conn() as conn:
with conn.cursor() as cur:
cur.execute(
'''SELECT "orderNo","createdAt","labelPdfUrl","destinationCountry"
FROM "Orders"
WHERE "systemName"=%s AND "labelPdfUrl" IS NOT NULL
AND "ocr_layout" IS NULL
AND "systemStatus"=ANY(%s)
ORDER BY "orderedAt" DESC LIMIT %s''',
("SDS", ["已完成", "部分发货", "备货中", "待确认"], limit),
)
return [(r[0], r[1], r[2], r[3]) for r in cur.fetchall()]
def update_ocr_fields(
self, order_no: str, created_at: datetime, fields: dict[str, str | None]
) -> None:
"""更新单个订单的 ocr_* 字段。fields 为 {ocr_xxx: value} 字典。"""
if not fields:
return
assignments = ", ".join(f'"{k}"=%s' for k in fields)
params = list(fields.values()) + [order_no, created_at]
with self._conn() as conn:
with conn.cursor() as cur:
cur.execute(
f'''UPDATE "Orders" SET {assignments}
WHERE "orderNo"=%s AND "createdAt"=%s''',
params,
)
conn.commit()
def count_ocr_status(self) -> dict[str, int]:
"""统计 OCR 处理状态(done/error/unsupported/pending)。"""
with self._conn() as conn:
with conn.cursor() as cur:
cur.execute(
'''SELECT
COUNT(*) FILTER (WHERE "ocr_layout" IS NOT NULL AND "ocr_layout" NOT IN ('ERROR','UNSUPPORTED')) AS done,
COUNT(*) FILTER (WHERE "ocr_layout"='ERROR') AS error,
COUNT(*) FILTER (WHERE "ocr_layout"='UNSUPPORTED') AS unsupported,
COUNT(*) FILTER (WHERE "ocr_layout" IS NULL AND "labelPdfUrl" IS NOT NULL) AS pending
FROM "Orders" WHERE "systemName"=%s''',
("SDS",),
)
row = cur.fetchone()
return {"done": row[0], "error": row[1], "unsupported": row[2], "pending": row[3]} if row else {}
# ---------- 主 upsert ----------
def sync_orders_and_goods(
prefect.yaml
+7
View File
@@ -80,3 +80,10 @@ deployments:
timezone: Asia/Shanghai
work_pool:
name: inkreach-pool
- name: sds-label-ocr
entrypoint: flows/sds/label_ocr.py:sds_label_ocr
schedules:
- cron: 0 */4 * * *
timezone: Asia/Shanghai
work_pool:
name: inkreach-pool