1.10

flask_app/general/投标人须知正文提取指定内容.py

@@ -5,8 +5,9 @@ from flask_app.general.json_utils import clean_json_string
 from flask_app.general.model_continue_query import process_continue_answers
 from flask_app.general.通义千问long import upload_file, qianwen_long_stream
 
+
 # 提取两个大标题之间的内容
-def extract_between_sections(data, target_values):
+def extract_between_sections(data, target_values, flag=False):
     target_found = False
     extracted_data = {}
     current_section_title = ""
@@ -27,7 +28,8 @@ def extract_between_sections(data, target_values):
                 target_found = False
 
             # 检查当前标题是否包含 target_values 中的任意关键词
-            if any(tv in value for tv in target_values) and not file_pattern.search(value):
+            if any(tv in value for tv in target_values):
+                if (flag and not file_pattern.search(value)) or not flag:
                     target_found = True  # 找到了目标章节，开始捕获后续内容
                     current_section_title = value  # 保存章节标题内容
 
@@ -40,6 +42,7 @@ def extract_between_sections(data, target_values):
 
     return extracted_data
 
+
 # 生成无结构的数据货物标
 def postprocess_formatted1(section_content):
     # print(json.dumps(section_content, ensure_ascii=False, indent=4))
@@ -77,6 +80,7 @@ def postprocess_formatted1(section_content):
 
     return concatenated
 
+
 # -----------以下为直接从序号中提取，无大标题的相关函数----------
 
 # 对于每个target_value元素，如果有完美匹配json_data中的键，那就加入这个完美匹配的键名，否则，把全部模糊匹配到的键名都加入
@@ -92,6 +96,8 @@ def find_keys_by_value(target_value, json_data):
 def find_keys_with_prefix(key_prefix, json_data):
     subheadings = [k for k in json_data if k.startswith(key_prefix)]
     return subheadings
+
+
 def extract_json(data, target_values):
     results = {}
     for target_value in target_values:
@@ -113,6 +119,7 @@ def extract_json(data, target_values):
                 results[subkey] = data[subkey]
     return results
 
+
 # 生成无结构的数据工程标,对提取出的若干键值对，生成外键为target_value，值为列表的新键值对
 def postprocess_formatted2(data, target_values):
     """
@@ -183,6 +190,7 @@ def postprocess_formatted2(data, target_values):
 
     return result
 
+
 def get_requirements_with_gpt(merged_baseinfo_path, selection):
     """
     根据 selection 的值选择相应的用户查询，并调用大模型获取要求。
@@ -338,6 +346,7 @@ def get_requirements_with_gpt(merged_baseinfo_path, selection):
     except Exception as e:
         return {"error": "调用大模型失败"}
 
+
 # 读取JSON数据，提取内容，转换结构，并打印结果
 def extract_from_notice(merged_baseinfo_path, clause_path, type):
     """
@@ -365,8 +374,8 @@ def extract_from_notice(merged_baseinfo_path, clause_path, type):
         3: ["重新招标、不再招标和终止招标", "重新招标", "重新采购", "不再招标", "不再采购", "终止招标", "终止采购"],
         4: ["评标"]  # 测试
     }
-
     # 获取对应 type 的 target_values
+    flag = (type == 2)
     target_values = type_target_map.get(type)
     if not target_values:
         print(f"Error: Invalid type specified: {type}. Use 1, 2, 3, or 4.")
@@ -378,7 +387,7 @@ def extract_from_notice(merged_baseinfo_path, clause_path, type):
                 data = json.load(file)
             if len(data) >= 60:
                 # 尝试使用大章节筛选
-                extracted_data = extract_between_sections(data, target_values)
+                extracted_data = extract_between_sections(data, target_values, flag)
                 if extracted_data:
                     # 后处理并返回结果
                     extracted_data_concatenated = {
@@ -402,12 +411,13 @@ def extract_from_notice(merged_baseinfo_path, clause_path, type):
         print(f"Error occurred while processing clause_path '{clause_path}': {e}")
         return DEFAULT_RESULT
 
+
 if __name__ == "__main__":
     # file_path = 'C:\\Users\\Administrator\\Desktop\\fsdownload\\3bffaa84-2434-4bd0-a8ee-5c234ccd7fa0\\clause1.json'
-    merged_baseinfo_path=r"C:\Users\Administrator\Desktop\fsdownload\ec7d5328-9c57-450f-baf4-2e5a6f90ed1d\tmp\merged_baseinfo_path_more.pdf"
-    clause_path=r"C:\Users\Administrator\Desktop\fsdownload\ec7d5328-9c57-450f-baf4-2e5a6f90ed1d\tmp\clause1.json"
+    merged_baseinfo_path = r"C:\Users\Administrator\Desktop\fsdownload\b29de31a-297e-42cf-b9ba-6859e530a472\ztbfile_merged_baseinfo.pdf"
+    clause_path = r"C:\Users\Administrator\Desktop\fsdownload\b29de31a-297e-42cf-b9ba-6859e530a472\clause1.json"
     try:
-        res = extract_from_notice(merged_baseinfo_path,clause_path, 2)  # 可以改变此处的 type 参数测试不同的场景
+        res = extract_from_notice(merged_baseinfo_path, clause_path, 1)  # 可以改变此处的 type 参数测试不同的场景
         res2 = json.dumps(res, ensure_ascii=False, indent=4)
         print(res2)
     except ValueError as e: