在自动识别时将图片转正并上传到云端

README.md

@@ -105,4 +105,8 @@
     1. 照片涂抹重新添加方向识别与矫正，因为照片审核人员暂无法保证所有图片的方向正确
     2. 添加照片涂抹错误分析
     3. 优化图片分割和拓展，对横竖图片分别进行优化
-    4. 优化镜像构建，将依赖与内容分开，显著提高构建速度
+    4. 优化镜像构建，将依赖与内容分开，显著提高构建速度
+18. 版本号：1.11.0
+    1. 修正含旋转的信息抽取结果的绘制
+    2. 调整zx_ocr表名为zx_ie_result
+    3. 在自动识别时将图片转正并上传到云端

docker-compose.yml

@@ -1,6 +1,6 @@
 x-env:
   &template
-  image: fcb_photo_review:1.10.4
+  image: fcb_photo_review:1.11.0
   restart: always
 
 services:

photo_review/photo_review.py

@@ -3,6 +3,7 @@ import logging
 import os
 import tempfile
 import time
+from collections import defaultdict
 from time import sleep
 
 import cv2
@@ -67,25 +68,25 @@ def request_ie_result(task_enum, phrecs):
 
 
 # 关键信息提取
-def information_extraction(ie, phrecs):
+def information_extraction(ie, phrecs, identity):
     result = {}
-    # 同一批图的标识
-    identity = int(time.time())
     for phrec in phrecs:
         img_path = ufile.get_private_url(phrec.cfjaddress)
         if not img_path:
             continue
-        split_results = image_util.split(img_path)
+        image = image_util.read(img_path)
-
+        angles = image_util.parse_rotation_angles(image)
+        angle_count = defaultdict(int, {"0": 0})
+        split_results = image_util.split(image)
         for split_result in split_results:
-            angles = image_util.parse_rotation_angles(split_result["img"])
             rotated_img = image_util.rotate(split_result["img"], int(angles[0]))
             ie_results = [{"result": ie_temp_image(ie, OCR, rotated_img), "angle": angles[0]}]
-            if not ie_results[0] or len(ie_results[0]) < len(ie.kwargs.get("schema")):
+            if not ie_results[0]["result"] or len(ie_results[0]["result"]) < len(ie.kwargs.get("schema")):
                 rotated_img = image_util.rotate(split_result["img"], int(angles[1]))
                 ie_results.append({"result": ie_temp_image(ie, OCR, rotated_img), "angle": angles[1]})
 
             now = util.get_default_datetime()
+            best_angle = ["0", 0]
             for ie_result in ie_results:
                 result_json = json.dumps(ie_result["result"], ensure_ascii=False)
                 if len(result_json) > 5000:
@@ -101,6 +102,23 @@ def information_extraction(ie, phrecs):
 
                 result = merge_result(result, ie_result["result"])
 
+                if len(ie_result["result"]) > best_angle[1]:
+                    best_angle = [ie_result["angle"], len(ie_result["result"])]
+
+            angle_count[best_angle[0]] += 1
+
+        img_angle = max(angle_count, key=angle_count.get)
+        if img_angle != "0":
+            image = image_util.rotate(image, int(img_angle))
+            with tempfile.NamedTemporaryFile(delete=False, suffix=".jpg") as temp_file:
+                cv2.imwrite(temp_file.name, image)
+            try:
+                ufile.upload_file(phrec.cfjaddress, temp_file.name)
+            except Exception as e:
+                logging.error(f"上传图片({phrec.cfjaddress})失败", exc_info=e)
+            finally:
+                util.delete_temp_file(temp_file.name)
+
     return result
 
 
@@ -159,8 +177,8 @@ def save_or_update_ie(table, pk_phhd, data):
     session.close()
 
 
-def settlement_task(pk_phhd, settlement_list):
+def settlement_task(pk_phhd, settlement_list, identity):
-    settlement_list_ie_result = information_extraction(SETTLEMENT_IE, settlement_list)
+    settlement_list_ie_result = information_extraction(SETTLEMENT_IE, settlement_list, identity)
     settlement_data = {
         "pk_phhd": pk_phhd,
         "name": handle_name(get_best_value_in_keys(settlement_list_ie_result, PATIENT_NAME)),
@@ -192,8 +210,8 @@ def settlement_task(pk_phhd, settlement_list):
     save_or_update_ie(ZxIeSettlement, pk_phhd, settlement_data)
 
 
-def discharge_task(pk_phhd, discharge_record):
+def discharge_task(pk_phhd, discharge_record, identity):
-    discharge_record_ie_result = information_extraction(DISCHARGE_IE, discharge_record)
+    discharge_record_ie_result = information_extraction(DISCHARGE_IE, discharge_record, identity)
     discharge_data = {
         "pk_phhd": pk_phhd,
         "hospital": handle_hospital(get_best_value_in_keys(discharge_record_ie_result, HOSPITAL)),
@@ -233,8 +251,8 @@ def discharge_task(pk_phhd, discharge_record):
     save_or_update_ie(ZxIeDischarge, pk_phhd, discharge_data)
 
 
-def cost_task(pk_phhd, cost_list):
+def cost_task(pk_phhd, cost_list, identity):
-    cost_list_ie_result = information_extraction(COST_IE, cost_list)
+    cost_list_ie_result = information_extraction(COST_IE, cost_list, identity)
     cost_data = {
         "pk_phhd": pk_phhd,
         "name": handle_name(get_best_value_in_keys(cost_list_ie_result, PATIENT_NAME)),
@@ -266,9 +284,11 @@ def photo_review(pk_phhd):
         elif phrec.cRectype == "4":
             cost_list.append(phrec)
 
-    settlement_task(pk_phhd, settlement_list)
+    # 同一批图的标识
-    discharge_task(pk_phhd, discharge_record)
+    identity = int(time.time())
-    cost_task(pk_phhd, cost_list)
+    settlement_task(pk_phhd, settlement_list, identity)
+    discharge_task(pk_phhd, discharge_record, identity)
+    cost_task(pk_phhd, cost_list, identity)
 
 
 def main():