137 lines
5.8 KiB
Python
137 lines
5.8 KiB
Python
import json
|
||
import re
|
||
from functools import cmp_to_key
|
||
|
||
from flask_app.general.投标人须知正文提取指定内容 import process_nested_data, transform_json, get_requirements_with_gpt,concatenate_keys_values
|
||
|
||
|
||
#提取两个大标题之间的内容
|
||
def extract_between_sections(data, target_values):
|
||
target_found = False
|
||
extracted_data = {}
|
||
current_section_title = ""
|
||
section_pattern = re.compile(r'^[一二三四五六七八九十]+$') # 匹配 "一", "二", "三" 等大标题
|
||
current_block = {}
|
||
|
||
# 遍历所有键值对
|
||
for key, value in data.items():
|
||
# 只匹配形如 "一": "竞争性磋商响应文件" 的章节标题
|
||
if section_pattern.match(key): #匹配到大标题...
|
||
if target_found:
|
||
# 如果已经找到了符合的章节，并且遇到了另一个章节
|
||
# 保存当前块并重置
|
||
if current_block:
|
||
extracted_data[current_section_title] = current_block
|
||
current_block = {}
|
||
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):
|
||
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
|
||
|
||
elif target_found: # 匹配到普通序号...
|
||
current_block[key] = value
|
||
|
||
# 保存最后一个块（如果有的话）
|
||
if current_block:
|
||
extracted_data[current_section_title] = current_block
|
||
|
||
return extracted_data
|
||
|
||
|
||
# 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
|
||
|
||
|
||
"""
|
||
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'键'的列表。
|
||
"""
|
||
|
||
# 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):
|
||
"""
|
||
从公告中提取特定类型的内容。
|
||
|
||
Args:
|
||
merged_baseinfo_path (str): 合并后的基础信息路径。
|
||
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 或 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 = ""
|
||
|
||
# 映射 type 到 target_values
|
||
type_target_map = {
|
||
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", "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:
|
||
print(f"Error: Invalid type specified: {type}. Use 1, 2, 3, or 4.")
|
||
return DEFAULT_RESULT
|
||
|
||
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
|
||
print(json.dumps(extracted_data_concatenated,ensure_ascii=False,indent=4))
|
||
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
|
||
|
||
except Exception as e:
|
||
print(f"Error occurred: {e}")
|
||
return DEFAULT_RESULT
|
||
|
||
#TODO:可以通过判断格式来看是否需要调用GPT 1.1 2.1....
|
||
|
||
if __name__ == "__main__":
|
||
clause_path = r'C:\Users\Administrator\Desktop\招标文件\output4\tmp\clause1.json'
|
||
merged_baseinfo_path=r"C:\Users\Administrator\Desktop\fsdownload\a110ed59-00e8-47ec-873a-bd4579a6e628\ztbfile_merged_baseinfo.pdf"
|
||
# file_path = 'D:\\flask_project\\flask_app\\static\\output\\fee18877-0c60-4c28-911f-9a5f7d1325a7\\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
|
||
res2=json.dumps(res,ensure_ascii=False,indent=4)
|
||
print(res2)
|
||
except ValueError as e:
|
||
print(e)
|