zbparse/flask_app/货物标/投标人须知正文提取指定内容货物标版.py

100 lines
4.4 KiB
Python
Raw Normal View History

2024-09-30 16:23:39 +08:00
import json
import re
2024-10-22 21:02:54 +08:00
from functools import cmp_to_key
2024-12-11 17:42:51 +08:00
from flask_app.general.投标人须知正文提取指定内容 import get_requirements_with_gpt, concatenate_keys_values, extract_between_sections
2024-09-30 16:23:39 +08:00
2024-11-07 16:22:52 +08:00
# def process_with_outer_key(data):
# processed_data = {}
#
# # 遍历外层的键值对
# for outer_key, inner_data in data.items():
# # 调用 transform_json 函数对内层数据进行处理
# processed_inner_data = transform_json(inner_data)
#
# # 将处理后的数据保留在外层键下
# processed_data[outer_key] = processed_inner_data
#
# return processed_data
2024-09-30 16:23:39 +08:00
2024-10-19 12:53:25 +08:00
"""
递归处理嵌套的数据结构字典和列表
对最内层的字符串值应用 post_process 函数
post_process 函数尝试将长字符串按特定模式分割成块每块至少包含50个中英文字符
如果字典中所有值都是 """/" 或空列表则返回''的列表
"""
2024-09-30 16:23:39 +08:00
# 读取JSON数据提取内容转换结构并打印结果
2024-11-07 17:01:14 +08:00
def extract_from_notice(merged_baseinfo_path, clause_path, type):
2024-10-17 15:33:58 +08:00
"""
从公告中提取特定类型的内容
2024-11-07 17:01:14 +08:00
2024-10-17 15:33:58 +08:00
Args:
2024-11-07 17:01:14 +08:00
merged_baseinfo_path (str): 合并后的基础信息路径
2024-10-17 15:33:58 +08:00
clause_path (str): 包含条款的JSON文件路径
type (int): 提取的类型
1 - ["投标文件","响应文件","响应性文件"]
2 - ["开标", "评标", "定标","磋商程序","中标","程序","步骤"]
3 - ["重新招标、不再招标和终止招标", "重新招标", "不再招标", "终止招标"]
4 - ["评标"] # 测试
Returns:
dict str: 提取并处理后的数据或在 `clause_path` 为空或发生错误时返回空字符串 `""`
"""
# 定义默认的返回结果
DEFAULT_RESULT = ""
2024-11-07 17:01:14 +08:00
# 映射 type 到 target_values
type_target_map = {
1: ["投标文件", "响应文件", "响应性文件"],
2024-12-25 14:35:52 +08:00
2: ["开标", "评标", "定标","评审","成交","合同","磋商","谈判", "中标", "程序", "步骤"],
2024-11-07 17:01:14 +08:00
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
}
target_values = type_target_map.get(type)
if not target_values:
2024-10-17 15:33:58 +08:00
print(f"Error: Invalid type specified: {type}. Use 1, 2, 3, or 4.")
return DEFAULT_RESULT
2024-11-06 17:29:45 +08:00
2024-11-07 17:01:14 +08:00
try:
# 如果 clause_path 不为空,尝试从 JSON 文件中提取数据
if clause_path:
with open(clause_path, 'r', encoding='utf-8') as file:
data = json.load(file)
# 提取目标部分
extracted_data = extract_between_sections(data, target_values) # 读取json截取大标题之间的内容
# 如果成功提取到数据,进行后续处理并返回
if extracted_data:
# 合并键值对,启用结构化
extracted_data_concatenated = {
section: concatenate_keys_values(content)
for section, content in extracted_data.items()
}
# transformed_data = process_with_outer_key(extracted_data) #取消注释这三行
# final_result = process_nested_data(transformed_data)
# return final_result
return extracted_data_concatenated
# 如果 clause_path 为空或提取数据失败,调用回退函数
final_result = get_requirements_with_gpt(merged_baseinfo_path, type)
return final_result
2024-09-30 16:23:39 +08:00
2024-10-17 15:33:58 +08:00
except Exception as e:
2024-11-07 17:01:14 +08:00
print(f"Error occurred: {e}")
2024-10-17 15:33:58 +08:00
return DEFAULT_RESULT
2024-12-10 09:02:39 +08:00
#TODO:可以通过判断格式来看是否需要调用GPT 1.1 2.1....
2024-09-30 16:23:39 +08:00
if __name__ == "__main__":
clause_path = r'C:\Users\Administrator\Desktop\文件解析问题\文件解析问题\910f6c0c-8a71-438d-a25c-7a6b34aa689c\tmp\clause1.json'
2024-12-12 16:06:20 +08:00
merged_baseinfo_path=r"D:\flask_project\flask_app\static\output\output1\ce279982-aeeb-4f08-ab39-df6ee2732eae\ztbfile.pdf"
2024-10-22 21:02:54 +08:00
# file_path = 'D:\\flask_project\\flask_app\\static\\output\\fee18877-0c60-4c28-911f-9a5f7d1325a7\\clause1.json'
2024-09-30 16:23:39 +08:00
try:
2024-12-25 14:35:52 +08:00
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
2024-09-30 16:23:39 +08:00
res2=json.dumps(res,ensure_ascii=False,indent=4)
print(res2)
except ValueError as e:
print(e)