171 lines
7.8 KiB
Python
171 lines
7.8 KiB
Python
# -*- encoding:utf-8 -*-
|
||
import json
|
||
import os.path
|
||
import re
|
||
from flask_app.general.json_utils import extract_content_from_json # 可以选择性地导入特定的函数
|
||
from flask_app.main.提取打勾符号 import read_pdf_and_judge_main
|
||
from flask_app.general.多线程提问 import multi_threading
|
||
from flask_app.general.通义千问long import qianwen_long,upload_file
|
||
#调用qianwen-ask之后，组织提示词问百炼。
|
||
|
||
def construct_judge_questions(json_data):
|
||
"""
|
||
根据提供的 JSON 数据，生成一个问题列表。每个问题针对一个值为 '未知' 的键，
|
||
要求以 JSON 格式提供该键的值，仅限于 '是', '否', '未知'。
|
||
|
||
Args:
|
||
json_data (str or dict): 输入的 JSON 数据，可以是 JSON 字符串或字典。
|
||
|
||
Returns:
|
||
list: 包含生成的问题的列表。如果没有符合条件的键，则返回空列表。
|
||
"""
|
||
# 如果 json_data 是字符串，尝试将其解析为字典
|
||
if isinstance(json_data, str):
|
||
try:
|
||
parsed_data = json.loads(json_data)
|
||
except json.JSONDecodeError:
|
||
print("输入的 json_data 不是有效的 JSON 字符串。")
|
||
return []
|
||
elif isinstance(json_data, dict):
|
||
parsed_data = json_data
|
||
else:
|
||
print("输入的 json_data 必须是 JSON 字符串或字典。")
|
||
return []
|
||
# 提取值为 '未知' 的键
|
||
question_keys = [key for key, value in parsed_data.items() if value == '未知']
|
||
# 如果没有符合条件的键，返回空列表
|
||
if not question_keys:
|
||
return []
|
||
# 生成问题列表
|
||
questions = []
|
||
for key in question_keys:
|
||
question = (
|
||
f"请你根据投标人须知前附表中的信息回答，本项目{key}，"
|
||
f"请按json格式给我提供信息，键名为'{key}'，键值仅限于'是','否','未知'，如果原文中对应内容处为'/'或者'\'，键值填'否'。"
|
||
)
|
||
questions.append(question)
|
||
return questions
|
||
|
||
|
||
def merge_json_to_list(merged):
|
||
"""Merge updates into the original data by modifying specific keys based on their value ('是' or '否'), and create a list based on these values."""
|
||
chosen_numbers = []
|
||
|
||
# 处理是否允许分包 保持'是否允许分包'键名主要是由于存在'未知'的情况。
|
||
if merged.get('是否允许分包') == '是':
|
||
chosen_numbers.append(1)
|
||
merged.pop('是否允许分包', None)
|
||
elif merged.get('是否允许分包') == '否':
|
||
merged['分包'] = '不允许'
|
||
merged.pop('是否允许分包', None)
|
||
|
||
guarantee_key = '是否递交投标保证金' if '是否递交投标保证金' in merged else '是否递交磋商保证金'
|
||
if merged.get(guarantee_key) == '是':
|
||
chosen_numbers.extend([2, 3])
|
||
merged.pop(guarantee_key, None)
|
||
elif merged.get(guarantee_key) == '否':
|
||
guarantee_type = '投标' if '投标' in guarantee_key else '磋商'
|
||
merged[f'{guarantee_type}保证金'] = '不提交'
|
||
merged[f'退还{guarantee_type}保证金'] = '/'
|
||
merged.pop(guarantee_key, None)
|
||
|
||
# 处理是否有履约保证金
|
||
if merged.get('是否提交履约保证金') == '是':
|
||
chosen_numbers.append(4)
|
||
merged.pop('是否提交履约保证金', None)
|
||
elif merged.get('是否提交履约保证金') == '否':
|
||
merged['履约保证金'] = '不提交'
|
||
merged.pop('是否提交履约保证金', None)
|
||
|
||
# 处理是否有招标代理服务费
|
||
if merged.get('是否有招标代理服务费') == '是':
|
||
chosen_numbers.append(5)
|
||
merged.pop('是否有招标代理服务费', None)
|
||
elif merged.get('是否有招标代理服务费') == '否':
|
||
merged['招标代理服务费'] = '无'
|
||
merged.pop('是否有招标代理服务费', None)
|
||
|
||
if merged.get('是否组织踏勘现场') == '是':
|
||
chosen_numbers.append(6)
|
||
merged.pop('是否组织踏勘现场',None)
|
||
elif merged.get('是否组织踏勘现场') == '否':
|
||
merged['踏勘现场']='不组织'
|
||
merged.pop('是否组织踏勘现场', None)
|
||
|
||
preparation_key = '是否召开投标预备会' if '是否召开投标预备会' in merged else '是否召开投标答疑会'
|
||
if merged.get(preparation_key) == '是':
|
||
chosen_numbers.append(7)
|
||
merged.pop(preparation_key, None)
|
||
elif merged.get(preparation_key) == '否':
|
||
meeting_type = '预备会' if '预备会' in preparation_key else '答疑会'
|
||
merged[f'投标{meeting_type}']='不召开'
|
||
merged.pop(preparation_key,None)
|
||
|
||
if merged.get('是否允许偏离') == '是':
|
||
chosen_numbers.append(8)
|
||
merged.pop('是否允许偏离',None)
|
||
elif merged.get('是否允许偏离') == '否':
|
||
merged['偏离']='不允许'
|
||
merged.pop('是否允许偏离', None)
|
||
|
||
return chosen_numbers, merged
|
||
|
||
|
||
|
||
def read_questions_from_judge(file_path, indices):
|
||
questions = []
|
||
|
||
# 读取文件内容
|
||
with open(file_path, 'r', encoding='utf-8') as file:
|
||
content = file.read()
|
||
|
||
# 正则表达式提取问题
|
||
pattern = r'(\d+)\.(.*?)#pdf提取之后的提示词|(\d+)\.(.*?)(?=\d+\.|$)'
|
||
matches = re.findall(pattern, content, re.DOTALL)
|
||
|
||
# 解析匹配到的内容并提取对应序号的问题
|
||
for match in matches:
|
||
num = match[0] or match[2]
|
||
question = match[1].strip() or match[3].strip()
|
||
if int(num) in indices:
|
||
questions.append(question)
|
||
|
||
return questions
|
||
|
||
|
||
def judge_whether_main(file_path,output_folder): #传入招标文件中‘投标人须知前附表’
|
||
user_query1 = "请你依据以上信息，回答以下问题：是否组织踏勘现场？是否召开投标预备会？是否允许偏离？是否退还投标文件？是否允许分包? 是否需要递交投标保证金？是否需要提交履约保证金（履约担保）？是否有招标代理服务费？请按json格式给我提供信息，键名分别为'是否组织踏勘现场','是否召开投标预备会','是否允许偏离','是否退还投标文件',是否允许分包','是否递交投标保证金','是否提交履约保证金','是否有招标代理服务费',键值仅限于'是','否','未知',若存在矛盾信息，请回答'未知'。"
|
||
output_txt_path = os.path.join(output_folder,'judge_exist.txt')
|
||
read_pdf_and_judge_main(file_path, output_txt_path) #提取打勾符号
|
||
file_id=upload_file(output_txt_path)
|
||
qianwen_answer=qianwen_long(file_id,user_query1)
|
||
user_querys = construct_judge_questions(extract_content_from_json(qianwen_answer)) # 提取回答为”未知“的键
|
||
# 判断user_query2是否为空
|
||
if user_querys:
|
||
file_id = upload_file(file_path)
|
||
qianwen_results = multi_threading(user_querys, "", file_id, 2) # 1代表使用百炼rag 2代表使用qianwen-long
|
||
qianwen_list = [extract_content_from_json(res) for _, res in qianwen_results] if qianwen_results else [] #整个前附表一起传问千问long
|
||
return process_judge_content(qianwen_answer, qianwen_list)
|
||
|
||
else:
|
||
print("正常现象，没有'未知',无需调用qianwen-long")
|
||
original = extract_content_from_json(qianwen_answer)
|
||
return merge_json_to_list(original)
|
||
|
||
def process_judge_content(original_json, qianwen_list): #用新的数据合并旧数据
|
||
"""Process judging content by merging updates into the original JSON data."""
|
||
original = extract_content_from_json(original_json)
|
||
for i in qianwen_list:
|
||
original.update(i)
|
||
return merge_json_to_list(original)
|
||
|
||
|
||
|
||
if __name__ == "__main__":
|
||
file_path="C:\\Users\Administrator\\Desktop\\fsdownload\\ee2d8828-bae0-465a-9171-7b2dd7453251\\ztbfile_tobidders_notice_table.pdf"
|
||
output_dir='C:\\Users\Administrator\\Desktop\\fsdownload\\ee2d8828-bae0-465a-9171-7b2dd7453251\\tmp'
|
||
chosen_numbers, merged=judge_whether_main(file_path,output_dir)
|
||
print(chosen_numbers)
|
||
print(merged)
|
||
|