11.5货物标截取优化

flask_app/general/清除file_id.py

@@ -0,0 +1,23 @@
+import os
+from openai import OpenAI
+import json
+
+# 初始化 OpenAI 客户端
+client = OpenAI(
+    api_key=os.getenv("DASHSCOPE_API_KEY"),
+    base_url="https://dashscope.aliyuncs.com/compatible-mode/v1",
+)
+
+# 获取文件列表
+file_stk = client.files.list()
+
+# 将文件信息解析为 JSON 格式
+file_data = json.loads(file_stk.model_dump_json())
+
+# 提取所有文件的 id
+file_ids = [file["id"] for file in file_data["data"]]
+
+# 循环删除每个文件
+for file_id in file_ids:
+    file_object = client.files.delete(file_id)
+    print(file_object.model_dump_json())

flask_app/main/截取pdf.py

@@ -337,17 +337,17 @@ def truncate_pdf_main(input_path, output_folder, selection):
             # Selection 2: 评标办法
             pattern_pairs = [
                 (
-                    re.compile(r'^第[一二三四五六七八九十]+(?:章|部分)\s*(磋商|谈判|评标|评定|评审)\s*(办法|方法)'),
+                    re.compile(r'^第[一二三四五六七八九十]+(?:章|部分)\s*(?=.*(?:磋商|谈判|评标|评定|评审))(?=.*(?:办法|方法))'),
                     # Alternative begin pattern
                     re.compile(r'评标办法正文|评标办法|^第[一二三四五六七八九1-9]+(?:章|部分)\s*[\u4e00-\u9fff]+')
                 # Alternative end pattern
                 ),
                 (
                     re.compile(
-                        r'第[一二三四五六七八九1-9]+(?:章|部分)\s*[\u4e00-\u9fff、]*?(磋商|谈判|评标|评定|评审)\s*(办法|方法)[\u4e00-\u9fff、]*\s*$|\s*评标(办法|方法)前附表\s*$',
+                        r'第[一二三四五六七八九十1-9]+(?:章|部分)\s*[\u4e00-\u9fff、()（）]*?(?=.*(?:磋商|谈判|评标|评定|评审))(?=.*(?:办法|方法))[\u4e00-\u9fff、()（）]*\s*$|\s*评标(办法|方法)前附表\s*$',
                         re.MULTILINE
                     ),
-                    re.compile(r'第[一二三四五六七八九1-9]+(?:章|部分)\s*[\u4e00-\u9fff]+\s*$', re.MULTILINE)
+                    re.compile(r'第[一二三四五六七八九十1-9]+(?:章|部分)\s*[\u4e00-\u9fff]+\s*$', re.MULTILINE)
                 )
             ]
             output_suffix = "evaluation_method"
@@ -559,22 +559,23 @@ def truncate_pdf_specific_engineering(pdf_path, output_folder, selections, uniqu
 
 
 # TODO:需要完善二次请求。目前invalid一定能返回 前附表  须知正文如果为空的话要额外处理一下，比如说就不进行跳转（见xx表） 开评定标这里也要考虑  如果评分表为空，也要处理。
-# TODO:zbtest8 zbtest18有问题  后期需要完善，截取需要截两次，第一次严格第二次宽松
 # TODO:目前merged_baseinfo没有包含投标人须知正文。
+
+#TODO:zbtest20有问题
 # 投标人须知前附表改为货物标一样的
 if __name__ == "__main__":
     start_time = time.time()
     # input_path = "C:\\Users\\Administrator\\Desktop\\new招标文件\\工程标\\HBDL-2024-0017-001-招标文件.pdf"
-    input_path="C:\\Users\\Administrator\\Desktop\\fsdownload\\0b1861e6-c7f6-4541-9182-b1384ba84f3b\\ztbfile.pdf"
+    # input_path="C:\\Users\\Administrator\\Desktop\\fsdownload\\0b1861e6-c7f6-4541-9182-b1384ba84f3b\\ztbfile.pdf"
     # input_path = "C:\\Users\\Administrator\\Desktop\\货物标\\zbfiles\\2-招标文件.pdf"
-    # input_path="C:\\Users\\Administrator\\Desktop\\招标文件\\招标test文件夹\\zbtest1.pdf"
+    input_path="C:\\Users\\Administrator\\Desktop\\招标文件\\招标test文件夹"
     output_folder = "C:\\Users\\Administrator\\Desktop\\fsdownload\\0b1861e6-c7f6-4541-9182-b1384ba84f3b\\tmp"
     # files=truncate_pdf_multiple(input_path,output_folder)
-    selections = [4, 1]  # 仅处理 selection 4、1
+    # selections = [4, 1]  # 仅处理 selection 4、1
-    files=truncate_pdf_specific_engineering(input_path,output_folder,selections)
+    # files=truncate_pdf_specific_engineering(input_path,output_folder,selections)
     # print(files)
-    # selection = 2  # 例如：1 - 投标人须知前附表+正文, 2 - 评标办法, 3 -资格审查条件 4-招标公告 5-无效标
+    selection = 2  # 例如：1 - 投标人须知前附表+正文, 2 - 评标办法, 3 -资格审查条件 4-招标公告 5-无效标
-    # generated_files = truncate_pdf_main(input_path, output_folder, selection)
+    generated_files = truncate_pdf_main(input_path, output_folder, selection)
     # print(generated_files)
     # print("生成的文件:", generated_files)
     end_time = time.time()

flask_app/货物标/截取pdf货物标版.py

@@ -151,11 +151,12 @@ def extract_pages(pdf_path, output_folder, begin_pattern, begin_page, end_patter
 
 def get_patterns_for_procurement():
     begin_pattern = re.compile(
-        r'^第[一二三四五六七八九十百千]+(?:章|部分).*?(?:服务|项目|商务).*?要求|'
+        r'^第[一二三四五六七八九十百千]+(?:章|部分).*?(?:服务|项目|商务|技术).*?要求|^第[一二三四五六七八九十百千]+(?:章|部分).*(?:采购|需求).*',
-        r'^第[一二三四五六七八九十百千]+(?:章|部分).*?(?:采购|技术标准).*|'
+        re.MULTILINE)
-        r'^[一二三四五六七八九十百千]+、\s*采购清单', re.MULTILINE)
+    end_pattern = re.compile(r'^第[一二三四五六七八九十百千]+(?:章|部分)\s*[\u4e00-\u9fff]+', re.MULTILINE)
-    end_pattern = re.compile(
+    # begin_pattern=re.compile(
-        r'^第[一二三四五六七八九十百千]+(?:章|部分)\s*[\u4e00-\u9fff]+', re.MULTILINE)
+    #     r'第[一二三四五六七八九1-9]+(?:章|部分)\s*[\u4e00-\u9fff、]*?(?:服务|项目|商务|技术)\s*(办法|方法)[\u4e00-\u9fff、]*\s*$'
+    # )
     return begin_pattern, end_pattern
 
 
@@ -176,21 +177,6 @@ def get_patterns_for_evaluation_method():
         r'第[一二三四五六七八九1-9]+(?:章|部分)\s*[\u4e00-\u9fff]+\s*$', re.MULTILINE)
     return begin_pattern, end_pattern
 
-
-# def get_patterns_for_qualification():
-#     # # 原始匹配逻辑
-#     # begin_pattern_original = re.compile(
-#     #     r'^第[一二三四五六七八九十百千]+(?:章|部分).*?(资格审查).*', re.MULTILINE)
-#     # end_pattern_original = re.compile(
-#     #     r'^第[一二三四五六七八九十百千]+(?:章|部分)\s*[\u4e00-\u9fff]+', re.MULTILINE)
-#
-#     # 新匹配逻辑
-#     begin_pattern_new = re.compile(
-#         r'^资格性检查', re.MULTILINE)
-#     end_pattern_new = re.compile(
-#         r'^附件\s*\d+|^第[一二三四五六七八九十百千]+(?:章|部分)\s*[\u4e00-\u9fff]+', re.MULTILINE)
-#
-#     return begin_pattern_new, end_pattern_new
 def get_patterns_for_qualification():
     # begin_pattern = re.compile(
     #     r'^(?:附录(?:一|1)?[：:]|附件(?:一|1)?[：:]|附表(?:一|1)?[：:]|资格性检查|资格审查表).*(?:资格|符合性)?.*$',
@@ -570,20 +556,6 @@ def extract_pages_twice(pdf_path, output_folder, output_suffix, common_header,be
         return ""
 
 
-# def save_extracted_pages(pdf_document, start_page, end_page, pdf_path, output_folder, output_suffix):
-#     if output_suffix=='notice':
-#         print(start_page)
-#     base_file_name = os.path.splitext(os.path.basename(pdf_path))[0]
-#     output_pdf_path = os.path.join(output_folder, f"{base_file_name}_{output_suffix}.pdf")
-#     output_doc = PdfWriter()
-#     for page_num in range(start_page, end_page + 1):
-#         output_doc.add_page(pdf_document.pages[page_num])
-#     with open(output_pdf_path, 'wb') as f:
-#         output_doc.write(f)
-#     print(f"{output_suffix} 已截取并保存页面从 {start_page} 到 {end_page} 为 {output_pdf_path}")
-#     return output_pdf_path
-
-
 def save_extracted_pages(pdf_document, start_page, end_page, pdf_path, output_folder, output_suffix):
     try:
         # 检查 start_page 和 end_page 是否为 None
@@ -701,9 +673,17 @@ def process_input(input_path, output_folder, selection, output_suffix):
             local_output_suffix = "tobidders_notice"
         elif selection == 5:
             begin_pattern = re.compile(
-                r'^第[一二三四五六七八九十百千]+(?:章|部分).*?(?:服务|项目|商务).*?要求|^第[一二三四五六七八九十百千]+(?:章|部分).*?采购.*|^第[一二三四五六七八九十百千]+(?:章|部分).*?需求.*')
+                r'^第[一二三四五六七八九十百千]+(?:章|部分).*?(?:服务|项目|商务|技术).*?要求|^第[一二三四五六七八九十百千]+(?:章|部分).*(?:采购|需求).*')
             end_pattern = re.compile(r'^第[一二三四五六七八九十百千]+(?:章|部分)\s*[\u4e00-\u9fff]+')
             local_output_suffix = "procurement"
+
+            # begin_pattern = re.compile(
+            #     r'^第[一二三四五六七八九十百千]+(?:章|部分).*?(?:服务|项目|商务).*?要求|'
+            #     r'^第[一二三四五六七八九十百千]+(?:章|部分).*?(?:采购|技术标准).*|'
+            #     r'^[一二三四五六七八九十百千]+、\s*采购清单', re.MULTILINE)
+            # end_pattern = re.compile(
+            #     r'^第[一二三四五六七八九十百千]+(?:章|部分)\s*[\u4e00-\u9fff]+', re.MULTILINE)
+
         elif selection == 6:
             begin_pattern = re.compile(r'^第[一二三四五六七八九十百千]+(?:章|部分).*?(?:格式).*')
             end_pattern = re.compile(r'^第[一二三四五六七八九十百千]+(?:章|部分)\s*[\u4e00-\u9fff]+', re.MULTILINE)
@@ -825,11 +805,11 @@ if __name__ == "__main__":
     # input_path = "C:\\Users\\Administrator\\Desktop\\货物标\\output1\\2-招标文件_procurement.pdf"
     # input_path="C:\\Users\\Administrator\\Desktop\\fsdownload\\a091d107-805d-4e28-b8b2-0c7327737238\\ztbfile.pdf"
     # output_folder = "C:\\Users\\Administrator\\Desktop\\fsdownload\\a091d107-805d-4e28-b8b2-0c7327737238\\tmp"
-    output_folder="C:\\Users\\Administrator\\Desktop\\货物标\\zbfiles\\新建文件夹"
+    output_folder="C:\\Users\\Administrator\\Desktop\\货物标\\output1"
     # files = truncate_pdf_multiple(input_path, output_folder)
     # selections = [1,4]
     # files=truncate_pdf_specific_goods(input_path,output_folder,selections)
     # print(files)
-    selection = 3# 例如：1 - 公告, 2 - 评标办法, 3 - 资格审查后缀有qualification1或qualification2（与评标办法一致）  4.投标人须知前附表part1 投标人须知正文part2   5-采购需求
+    selection = 5# 例如：1 - 公告, 2 - 评标办法, 3 - 资格审查后缀有qualification1或qualification2（与评标办法一致）  4.投标人须知前附表part1 投标人须知正文part2   5-采购需求
     generated_files = truncate_pdf_main(input_path, output_folder, selection)
     print(generated_files)

flask_app/货物标/资格审查main.py

@@ -434,7 +434,7 @@ def combine_qualification_review(invalid_path, output_folder, qualification_path
         else:
             file_to_process = qualification_path
 
-        combined_res = process_file(file_to_process)
+        combined_res = process_file(file_to_process,invalid_path)
         match_keys = find_chapter_clause_references(combined_res)
 
         if not match_keys: