Merge branch 'feature/add_log_20240924' into feature/uat-tmp

冯轩
Showing 3 changed files with 27 additions and 8 deletions
src/apps/doc/management/commands/ocr_process.py
src/common/fsm_econtract/retriever.py
src/common/tools/pdf_to_img.py
--- a/src/apps/doc/management/commands/ocr_process.py
View file @83e1571
+++ b/src/apps/doc/management/commands/ocr_process.py
View file @83e1571
@@ -1015,6 +1015,13 @@ class Command(BaseCommand, LoggerMixin):
                            tmp_res = page_info_dict.get(str(pno2), {}).get(key1, '')
                            img_pno = pno1
                        res[key] = tmp_res
+                        # 添加处理, 
+                        # [售后回租合同] - 如果 key 是 "承租人签字", 且内容中包含 签署日期:XXXX, 则将签署日期去除
+                        # [车辆租赁抵押合同] - 如果 key 是 ""
+                        if key == '承租人签字' and '签署日期' in tmp_res:
+                            res[key] = tmp_res.split('签署日期')[0]
+                        if key == "抵押人签字" and "签署日期" in tmp_res:
+                            res[key] = tmp_res.split("签署日期")[0]
                        res.setdefault(consts.IMG_PATH_KEY, dict())[key] = page_info_dict.get(str(img_pno), {}).get(
                            consts.IMG_PATH_KEY, '')
                    else:
@@ -1668,16 +1675,20 @@ class Command(BaseCommand, LoggerMixin):
                        path_split = img_path.split('/')
                        task_str = consts.SPLIT_STR.join((path_split[-5], path_split[-3]))
+                        self.online_log.info('{0} [before lock] [img={1}] '.format(self.log_base, img_path))
                        with lock:
+                            self.online_log.info('{0} [get lock] [img={1}] '.format(self.log_base, img_path))
                            doc_res_dict = res_dict.setdefault(task_str, {})
                            doc_res_dict[img_path] = ocr_1_res
                            res_dict[task_str] = doc_res_dict
                            todo_count = todo_count_dict.get(task_str)
                            if todo_count == 1:
                                finish_queue.put(task_str)
+                                self.online_log.info('{0} [ocr_1 to finish_queue] [img={1}] '.format(self.log_base, img_path))
                                del todo_count_dict[task_str]
                            else:
                                todo_count_dict[task_str] = todo_count - 1
+                        self.online_log.info('{0} [after lock] [img={1}] '.format(self.log_base, img_path))
                    except Exception as e:
                        self.online_log.error('{0} [process error (store ocr res)] [img_path={1}] [error={2}]'.format(
                            self.log_base, img_path, traceback.format_exc()))
--- a/src/common/fsm_econtract/retriever.py
View file @83e1571
+++ b/src/common/fsm_econtract/retriever.py
View file @83e1571
@@ -19,10 +19,18 @@ class HMHRetriever:
    def get_target_fields(self, pdf_text_list):
        result = dict()
        is_find_name_id_company, is_find_application_no, is_find_name_date = False, False, False
-        for bbox, text in pdf_text_list.pop(str(0), []):
+        # for bbox, text in pdf_text_list.pop(str(0), []):
-            # print(text)
+        pdf_text_items = pdf_text_list.pop(str(0), [])
+        for i in range(len(pdf_text_items)):
+            bbox, text = pdf_text_items[i]
+            combined_text = text
+            if i < len(pdf_text_items) - 1:
+                combined_text += pdf_text_items[i + 1][1]
            if not is_find_name_id_company:
-                name_id_company_list = re.findall(r'姓名(.*)证件号码(.*)与(.*公司)', text)
+                # name_id_company_list = re.findall(r'姓名(.*?)证件号码(.*?)与(.*?公司|.*)', combined_text)
+                name_id_company_list = re.findall(r'姓名(.*)证件号码(.*)与(.*公司)', combined_text)
                for name_id_company_tuple in name_id_company_list:
                    if len(name_id_company_tuple) == 3: 
                        result[self.search_fields_list[0][0]] = {
@@ -40,7 +48,7 @@ class HMHRetriever:
                        is_find_name_id_company = True
                        break
            if not is_find_application_no:
-                application_no_list = re.findall(r'合同编号.*(CH-B\d*-\d*).*', text)
+                application_no_list = re.findall(r'合同编号.*(CH-B\d*-\d*).*', combined_text)
                if len(application_no_list) == 1:
                    result[self.search_fields_list[3][0]] = {
                        self.words_str: application_no_list[0],
@@ -48,7 +56,7 @@ class HMHRetriever:
                    }
                    is_find_application_no = True
            if not is_find_name_date:
-                name_date_list = re.findall(r'(.*).*签署日期.*(\d{4}-\d{2}-\d{2})', text)
+                name_date_list = re.findall(r'(.*).*签署日期.*(\d{4}-\d{2}-\d{2})', combined_text)
                for name_date_tuple in name_date_list:
                    if len(name_date_tuple) == 2: 
                        result[self.search_fields_list[4][0]] = {
--- a/src/common/tools/pdf_to_img.py
View file @83e1571
+++ b/src/common/tools/pdf_to_img.py
View file @83e1571
@@ -10,9 +10,9 @@ from io import BytesIO
 from unicodedata import normalize
 # 页面保存为png图片参数
-ZOOM_X_1 = ZOOM_Y_1 = 3.0
+ZOOM_X_1 = ZOOM_Y_1 = 1.0
-ZOOM_X_2 = ZOOM_Y_2 = 5.0
+ZOOM_X_2 = ZOOM_Y_2 = 2.0
-ZOOM_X_3 = ZOOM_Y_3 = 7.0
+ZOOM_X_3 = ZOOM_Y_3 = 3.0
 trans_1 = fitz.Matrix(ZOOM_X_1, ZOOM_X_1).preRotate(0)  # zoom factor 1 in each dimension
 trans_2 = fitz.Matrix(ZOOM_X_2, ZOOM_X_2).preRotate(0)  # zoom factor 2 in each dimension
 trans_3 = fitz.Matrix(ZOOM_X_3, ZOOM_X_3).preRotate(0)  # zoom factor 3 in each dimension