246 lines
11 KiB
Python
246 lines
11 KiB
Python
import json
|
||
import docx
|
||
import re
|
||
import os
|
||
from PyPDF2 import PdfReader
|
||
from flask_app.main.截取pdf import clean_page_content,extract_common_header
|
||
|
||
def extract_text_from_docx(file_path):
|
||
doc = docx.Document(file_path)
|
||
return '\n'.join([para.text for para in doc.paragraphs])
|
||
|
||
|
||
# def extract_text_from_pdf(file_path):
|
||
# # 从PDF文件中提取文本
|
||
# common_header = extract_common_header(file_path)
|
||
# pdf_document = PdfReader(file_path)
|
||
# text = ""
|
||
# # 遍历每一页
|
||
# for page in pdf_document.pages:
|
||
# # 提取当前页面的文本
|
||
# page_text = page.extract_text() if page.extract_text() else ""
|
||
# # 清洗页面文本
|
||
# page_text = clean_page_content(page_text, common_header)
|
||
# # 将清洗后的文本添加到总文本中
|
||
# text += page_text+"\n"
|
||
# return text
|
||
def extract_text_from_pdf(file_path, start_word, end_pattern):
|
||
# 从PDF文件中提取文本
|
||
common_header = extract_common_header(file_path)
|
||
pdf_document = PdfReader(file_path)
|
||
all_pages_text = []
|
||
start_index = None
|
||
# 处理所有页面
|
||
for i, page in enumerate(pdf_document.pages):
|
||
page_text = page.extract_text() if page.extract_text() else ""
|
||
cleaned_text = clean_page_content(page_text, common_header)
|
||
|
||
# 在第一页查找开始位置
|
||
if i == 0 and start_index is None:
|
||
start_match = re.search(start_word, cleaned_text, re.MULTILINE)
|
||
if start_match:
|
||
start_index = start_match.start()
|
||
cleaned_text = cleaned_text[start_index:]
|
||
|
||
# 在最后一页查找结束位置
|
||
if i == len(pdf_document.pages) - 1:
|
||
for pattern in end_pattern:
|
||
matches = list(re.finditer(pattern, cleaned_text, re.MULTILINE))
|
||
if matches:
|
||
end_index = matches[-1].start()
|
||
cleaned_text = cleaned_text[:end_index]
|
||
break
|
||
|
||
all_pages_text.append(cleaned_text)
|
||
|
||
# 合并所有页面的文本
|
||
full_text = "\n".join(all_pages_text)
|
||
# print(full_text)
|
||
return full_text
|
||
def extract_section(text, start_pattern, end_phrases):
|
||
# 查找开始模式
|
||
start_match = re.search(start_pattern, text)
|
||
if not start_match:
|
||
return "" # 如果没有找到匹配的开始模式，返回空字符串
|
||
start_index = start_match.end() # 从匹配的结束位置开始
|
||
|
||
# 初始化结束索引为文本总长度
|
||
end_index = len(text)
|
||
|
||
# 遍历所有结束短语，查找第一个出现的结束短语
|
||
for phrase in end_phrases:
|
||
match = re.search(phrase, text[start_index:], flags=re.MULTILINE)
|
||
if match:
|
||
end_index = start_index + match.start() # 更新结束索引为匹配到的开始位置
|
||
break # 找到第一个匹配后立即停止搜索
|
||
|
||
# 提取并返回从开始模式后到结束模式前的内容
|
||
return text[start_index:end_index]
|
||
|
||
def compare_headings(current, new):
|
||
# 使用过滤来确保只处理非空且为数字的部分
|
||
current_nums = [int(num) for num in current.split('.') if num.isdigit()]
|
||
new_nums = [int(num) for num in new.split('.') if num.isdigit()]
|
||
|
||
# 比较数字序列以确定标题的层次关系
|
||
for c, n in zip(current_nums, new_nums):
|
||
if n > c:
|
||
return True
|
||
elif n < c:
|
||
return False
|
||
|
||
# 如果新标题有更多层次，认为是新的子章节
|
||
return len(new_nums) > len(current_nums)
|
||
|
||
|
||
def should_add_newline(content, keywords, max_length=20):
|
||
content_str = ''.join(content).strip()
|
||
return any(keyword in content_str for keyword in keywords) or len(content_str) <= max_length
|
||
|
||
def handle_content_append(current_content, line_content, append_newline, keywords):
|
||
if append_newline:
|
||
if should_add_newline(current_content, keywords):
|
||
current_content.append('\n') # 添加换行符
|
||
append_newline = False
|
||
current_content.append(line_content)
|
||
return append_newline
|
||
|
||
#对二级标题如x.x进行额外处理：如果当前处理内容包含keywords中的内容，则必须保留换行符/如果当前内容字数大于20，不保留换行。
|
||
def parse_text_by_heading(text):
|
||
keywords = ['包含', '以下']
|
||
data = {}
|
||
current_key = None
|
||
current_content = []
|
||
append_newline = False
|
||
lines = text.split('\n')
|
||
for i, line in enumerate(lines): #由于本身就是按行读取处理，因此保存的时候不带'\n'
|
||
line_stripped = line.strip()
|
||
# 匹配形如 '1.1'、'2.2.3' 等至少包含一个点的标题，并确保其前后没有字母或括号
|
||
match = re.match(r'^(?<![a-zA-Z（(])(\d+(?:\.\d+)+)\s*(.*)', line_stripped)
|
||
if not match:
|
||
match = re.match(r'^(\d+\.)\s*(.+)$', line_stripped)
|
||
|
||
if match:
|
||
new_key, line_content = match.groups()
|
||
line_content = line_content.lstrip('.')
|
||
# 检查是否应该更新当前键和内容
|
||
if current_key is None or (compare_headings(current_key, new_key) and (
|
||
len(current_content) == 0 or current_content[-1][-1] != '第')): #current_content： 内容占两行则是： ['本招标项目已具备招标条件，现对本项目施工监理进行招标，项目概况见本', '章附件一。']
|
||
if current_key is not None:
|
||
# 将之前的内容保存到data中，保留第一个换行符，后续的换行符转为空字符
|
||
content_string = ''.join(current_content).strip()
|
||
data[current_key] = content_string.replace(' ', '')
|
||
current_key = new_key
|
||
current_content = [line_content]
|
||
# 只有当标题级别为两级（如 1.1）时，才设置 append_newline 为 True
|
||
# append_newline = len(new_key.split('.')) == 2
|
||
append_newline = len(new_key.rstrip('.').split('.')) <= 2
|
||
else:
|
||
append_newline = handle_content_append(current_content, line_content, append_newline, keywords)
|
||
else:
|
||
if line_stripped:
|
||
append_newline = handle_content_append(current_content, line_stripped, append_newline, keywords)
|
||
|
||
if current_key is not None:
|
||
# 保存最后一部分内容
|
||
content_string = ''.join(current_content).strip()
|
||
data[current_key] = content_string.replace(' ', '')
|
||
|
||
return data
|
||
|
||
def convert_to_json(file_path, start_word, end_phrases):
|
||
if file_path.endswith('.docx'):
|
||
text = extract_text_from_docx(file_path)
|
||
elif file_path.endswith('.pdf'):
|
||
text = extract_text_from_pdf(file_path,start_word,end_phrases)
|
||
else:
|
||
raise ValueError("Unsupported file format")
|
||
# 提取从 start_word 开始到 end_phrases 结束的内容
|
||
# text = extract_section(text, start_word, end_phrases)
|
||
# print(text)
|
||
parsed_data = parse_text_by_heading(text)
|
||
return parsed_data
|
||
|
||
def convert_clause_to_json(input_path,output_folder,type=1):
|
||
if not os.path.exists(input_path):
|
||
print(f"The specified file does not exist: {input_path}")
|
||
return ""
|
||
if type==1:
|
||
start_word = "投标人须知正文"
|
||
end_phrases = [
|
||
r'^第[一二三四五六七八九十]+章\s*评标办法',
|
||
r'^评标办法前附表',
|
||
r'^附(?:录|件|表)(?:一)?[:：]'
|
||
]
|
||
else:
|
||
start_word = r'第[一二三四五六七八九十]+章\s*招标公告|第一卷|招标编号:|招标编号：'
|
||
end_phrases=[r'第[一二三四五六七八九十]+章\s*投标人须知',r'投标人须知前附表']
|
||
result = convert_to_json(input_path, start_word, end_phrases)
|
||
# 检查输出文件夹是否存在，如果不存在则创建
|
||
if not os.path.exists(output_folder):
|
||
os.makedirs(output_folder)
|
||
print(f"Created output folder: {output_folder}")
|
||
file_name = "clause1.json" if type == 1 else "clause2.json"
|
||
output_path = os.path.join(output_folder, file_name)
|
||
with open(output_path, 'w', encoding='utf-8') as f:
|
||
json.dump(result, f, indent=4, ensure_ascii=False)
|
||
# post_process_json(output_path)
|
||
print(f"投标人须知正文条款提取成json文件: The data has been processed and saved to '{output_path}'.")
|
||
return output_path
|
||
|
||
# def post_process_json(json_file_path): #处理一级标题如'5.1'过长的内容 zbtest20
|
||
# # 读取 JSON 文件
|
||
# with open(json_file_path, 'r', encoding='utf-8') as file:
|
||
# data = json.load(file)
|
||
# processed_data = {}
|
||
# for key, value in data.items():
|
||
# # 检查是否是一级标题（如 '5.'），并且其值包含 '\n'
|
||
# if re.match(r'^\d+\.\s*$', key) and '\n' in value:
|
||
# # 分割标题和正文
|
||
# title, content = value.split('\n', 1)
|
||
#
|
||
# # 添加原来的标题作为 '5.0'，其值为原来标题的内容（即 title）
|
||
# processed_data[key] = title.strip()
|
||
# sub_key = f"{key.rstrip('.')}." + "0" # 自动生成 '5.0'，与 '5.' 一致，保证点号的存在
|
||
#
|
||
# processed_data[sub_key] = title.strip()
|
||
#
|
||
# # 初始化计数器
|
||
# sub_count = 1
|
||
#
|
||
# # 根据子序号 '1.' 或 '1、' 进行分割
|
||
# sub_sections = re.split(r'(\d+[\.\、])\s*', content)
|
||
#
|
||
# current_sub_content = ""
|
||
# for i in range(1, len(sub_sections), 2):
|
||
# sub_number = sub_sections[i].strip() # 获取子序号
|
||
# sub_content = sub_sections[i + 1].strip() # 获取内容
|
||
#
|
||
# # 生成三级标题，如 '5.0.1', '5.0.2'
|
||
# sub_key_with_number = f"{sub_key}.{sub_count}"
|
||
# processed_data[sub_key_with_number] = sub_content
|
||
# sub_count += 1
|
||
#
|
||
# else:
|
||
# # 如果没有分割需求，保留原数据
|
||
# processed_data[key] = value
|
||
|
||
# # 将修改后的数据重新写入到原来的 JSON 文件中
|
||
# with open(json_file_path, 'w', encoding='utf-8') as file:
|
||
# json.dump(processed_data, file, ensure_ascii=False, indent=4)
|
||
|
||
if __name__ == "__main__":
|
||
file_path = 'C:\\Users\\Administrator\\Desktop\\招标文件\\招标test文件夹\\zbtest20_tobidders_notice.pdf'
|
||
# file_path='C:\\Users\\Administrator\\Desktop\\货物标\\output4\\招招招标文件（一中多媒体报告厅教学设备）_tobidders_notice_part1.pdf'
|
||
# start_word = "投标人须知正文"
|
||
# end_phrases = [
|
||
# r'^第[一二三四五六七八九十]+章\s+评标办法', r'^评标办法前附表', r'^附录：', r'^附录一：', r'^附件：', r'^附件一：',
|
||
# r'^附表：', r'^附表一：', r'^附录:', r'^附录一:', r'^附件:', r'^附件一:', r'^附表:', r'^附表一:',
|
||
# ]
|
||
output_folder = 'C:\\Users\\Administrator\\Desktop\\招标文件\\招标test文件夹\\tmp'
|
||
try:
|
||
output_path = convert_clause_to_json(file_path,output_folder)
|
||
print(f"Final JSON result saved to: {output_path}")
|
||
except ValueError as e:
|
||
print("Error:", e)
|