314 lines
14 KiB
Python
314 lines
14 KiB
Python
import json
|
||
import docx
|
||
import re
|
||
import os
|
||
|
||
import fitz
|
||
from PyPDF2 import PdfReader
|
||
from flask_app.货物标.货物标截取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])
|
||
|
||
#PYPDF2版本
|
||
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
|
||
|
||
#fitz库版本
|
||
# 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)
|
||
# doc = fitz.open(file_path)
|
||
# all_pages_text = []
|
||
# start_index = None
|
||
#
|
||
# # 处理所有页面
|
||
# for i in range(len(doc)):
|
||
# page = doc[i]
|
||
# page_text = page.get_text()
|
||
# cleaned_text = clean_page_content(page_text, common_header)
|
||
# print(cleaned_text)
|
||
# print("yes")
|
||
# # 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(doc) - 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)
|
||
# # 关闭文档
|
||
# doc.close()
|
||
#
|
||
# return full_text
|
||
|
||
def chinese_to_arabic(chinese_num):
|
||
cn_num = {
|
||
'一': 1, '二': 2, '三': 3, '四': 4, '五': 5,
|
||
'六': 6, '七': 7, '八': 8, '九': 9, '十': 10,
|
||
'十一': 11, '十二': 12, '十三': 13, '十四': 14, '十五': 15
|
||
}
|
||
return cn_num.get(chinese_num, 0)
|
||
|
||
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
|
||
|
||
"""
|
||
保存换行符的具体逻辑：
|
||
|
||
对于二级标题（如 1.1），如果其后的内容包含关键词或内容较短（<=20字符），会在内容前添加一个换行符。
|
||
这个换行符会被保留在 current_content 列表中。
|
||
当处理下一个标题时，之前的内容（包括可能存在的换行符）会被合并并保存到 data 字典中。
|
||
|
||
解决了'一''二'这类标题出现在正文中的情况。但是目前的逻辑是如果'一'已有了，就加入正文，否则'一'作为新的标题。
|
||
"""
|
||
#提取json主函数
|
||
def parse_text_by_heading(text):
|
||
keywords = ['包含', '以下']
|
||
data = {}
|
||
current_key = None
|
||
current_key_chinese = None
|
||
current_value_chinese = None
|
||
current_content = []
|
||
append_newline = False
|
||
skip_subheadings = False
|
||
|
||
lines = text.split('\n')
|
||
for i, line in enumerate(lines):
|
||
line_stripped = line.strip().replace('．', '.')
|
||
match = re.match(r'^(?<![a-zA-Z（(])(\d+(?:\.\d+)+)\s*(.*)', line_stripped)
|
||
if not match:
|
||
match = re.match(r'^(\d+\.)\s*(.+)$', line_stripped)
|
||
|
||
if match:
|
||
new_key, line_content = match.groups()
|
||
line_content = line_content.lstrip('.．、,')
|
||
|
||
if any(keyword in line_content for keyword in keywords):
|
||
skip_subheadings = True
|
||
else:
|
||
skip_subheadings = False
|
||
|
||
# 保存中文标题的内容
|
||
if current_key_chinese is not None:
|
||
data[current_key_chinese] = current_value_chinese
|
||
current_key_chinese = None
|
||
|
||
# 保存阿拉伯数字的标题内容
|
||
if current_key is None or (compare_headings(current_key, new_key) and (
|
||
len(current_content) == 0 or current_content[-1][-1] != '第')):
|
||
if current_key is not None:
|
||
content_string = ''.join(current_content).strip()
|
||
data[current_key] = data.get(current_key, '') + content_string.replace(' ', '')
|
||
current_key = new_key
|
||
current_content = [line_content]
|
||
append_newline = len(new_key.rstrip('.').split('.')) <= 2
|
||
else:
|
||
append_newline = handle_content_append(current_content, line_stripped, append_newline, keywords)
|
||
else:
|
||
if not skip_subheadings:
|
||
# 合并了中文标题、字母标题、阿拉伯数字标题的匹配逻辑
|
||
pattern_title = re.compile(r'^\s*(?:[（(]\s*([一二三四五六七八九十]{1,2})\s*[)）]\s*|([一二三四五六七八九十]{1,2})\s*、\s*)')
|
||
chinese_match = pattern_title.match(line_stripped)
|
||
letter_match = re.match(r'^([A-Z])\.\s*(.*)$', line_stripped)
|
||
arabic_match = re.match(r'^(\d+、)\s*(.+)$', line_stripped)
|
||
|
||
# 优化了处理逻辑，减少冗余
|
||
if chinese_match or letter_match or arabic_match:
|
||
# 保存之前的 key 的内容（无论是中文标题还是阿拉伯数字标题）
|
||
if current_key is not None:
|
||
content_string = ''.join(current_content).strip()
|
||
data[current_key] = data.get(current_key, '') + content_string.replace(' ', '')
|
||
if current_key_chinese is not None:
|
||
data[current_key_chinese] = current_value_chinese
|
||
current_key_chinese = None
|
||
|
||
# 处理中文标题
|
||
if chinese_match:
|
||
current_key_chinese = chinese_match.group(1) or chinese_match.group(2)
|
||
current_value_chinese = line_stripped[chinese_match.end():].lstrip('.．、,')
|
||
if current_key_chinese in data:
|
||
handle_content_append(current_content, '\n' + line_stripped, append_newline, keywords)
|
||
current_key_chinese = None
|
||
# 处理字母标题
|
||
elif letter_match:
|
||
letter_key, letter_value = letter_match.groups()
|
||
letter_to_chinese = {
|
||
'A': '一', 'B': '二', 'C': '三', 'D': '四', 'E': '五',
|
||
'F': '六', 'G': '七', 'H': '八', 'I': '九', 'J': '十',
|
||
'K': '十一', 'L': '十二', 'M': '十三', 'N': '十四', 'O': '十五'
|
||
}
|
||
current_key_chinese = letter_to_chinese.get(letter_key, letter_key)
|
||
current_value_chinese = letter_value.lstrip('.．、,')
|
||
if current_key_chinese in data:
|
||
handle_content_append(current_content, '\n' + line_stripped, append_newline, keywords)
|
||
current_key_chinese = None
|
||
# 处理阿拉伯数字标题
|
||
elif arabic_match:
|
||
arabic_key, arabic_value = arabic_match.groups()
|
||
current_key = arabic_key.replace('、', '.')
|
||
current_content = [arabic_value]
|
||
append_newline = True
|
||
continue
|
||
|
||
# 如果没有匹配标题，继续处理正文内容
|
||
if line_stripped:
|
||
append_newline = handle_content_append(current_content, line_stripped, append_newline, keywords)
|
||
|
||
# 最后保存最后一个 key 对应的内容
|
||
if current_key is not None:
|
||
content_string = ''.join(current_content).strip()
|
||
data[current_key] = data.get(current_key, '') + content_string.replace(' ', '')
|
||
if current_key_chinese is not None:
|
||
data[current_key_chinese] = current_value_chinese
|
||
|
||
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")
|
||
# # print(text)
|
||
# parsed_data = parse_text_by_heading(text)
|
||
# return parsed_data
|
||
|
||
def convert_clause_to_json(file_path,output_folder,type=1,suffix_counter="1.json"):
|
||
if not os.path.exists(file_path):
|
||
print(f"The specified file does not exist: {file_path}")
|
||
return ""
|
||
if type == 1:
|
||
start_word = r'^\s*[（(]?\s*[一1]\s*[)）]?\s*[、.]*\s*(说\s*明|总\s*则)'
|
||
end_pattern = [r'^第[一二三四五六七八九十百千]+(?:章|部分)\s*[\u4e00-\u9fff]+']
|
||
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_pattern=[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']
|
||
text = extract_text_from_pdf(file_path, start_word, end_pattern)
|
||
result = parse_text_by_heading(text)
|
||
# result = convert_to_json(input_path, start_word, end_pattern)
|
||
# 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"
|
||
# file_name = f"clause{suffix_counter}.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)
|
||
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 process_folder(input_folder, output_folder):
|
||
# 获取输入文件夹中的所有文件，过滤掉不以 'part2' 结尾的文件
|
||
files = [f for f in os.listdir(input_folder) if os.path.isfile(os.path.join(input_folder, f)) and f.endswith('part2.pdf')]
|
||
|
||
# 初始化后缀计数器
|
||
suffix_counter = 1
|
||
|
||
# 遍历文件并对每个文件进行处理
|
||
for file_name in files:
|
||
file_path = os.path.join(input_folder, file_name)
|
||
suffix = str(suffix_counter) # 后缀为递增的数字
|
||
try:
|
||
# 调用 convert_clause_to_json，传递文件路径、输出文件夹和递增的后缀
|
||
output_path = convert_clause_to_json(file_path, output_folder, 1, suffix)
|
||
print(f"Processed file: {file_name}, JSON saved to: {output_path}")
|
||
except ValueError as e:
|
||
print(f"Error processing {file_name}: {e}")
|
||
|
||
# 后缀递增
|
||
suffix_counter += 1
|
||
|
||
#TODO:'C:\\Users\\Administrator\\Desktop\\货物标\\output4\\广水农商行门禁控制主机及基础验证设备采购项目——磋商文件（定稿）（三次）_tobidders_notice_part2.pdf' PYPDF2库读取有遗漏
|
||
#TODO:标题'一'应该越来越大，与'1.1'的逻辑分开'；
|
||
#TODO:911904c3-4d6e-47e0-acd8-b05e582209cb文件夹运行有问题，百炼出错了
|
||
if __name__ == "__main__":
|
||
# file_path = 'D:\\flask_project\\flask_app\\static\\output\\cfd4959d-5ea9-4112-8b50-9e543803f029\\ztbfile_tobidders_notice.pdf'
|
||
file_path='C:\\Users\\Administrator\\Desktop\\货物标\\output4\\磋商文件_tobidders_notice_part2.pdf'
|
||
# file_path = 'C:\\Users\\Administrator\\Desktop\\货物标\\output4\\2-招标文件（2020年广水市中小学教师办公电脑系统及多媒体“班班通”设备采购安装项目）_tobidders_notice_part2.pdf'
|
||
output_folder = 'C:\\Users\\Administrator\\Desktop\\货物标\\output4\\tmp2'
|
||
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)
|
||
|
||
# input_folder = 'C:\\Users\\Administrator\\Desktop\\货物标\\output4'
|
||
# output_folder = 'C:\\Users\\Administrator\\Desktop\\货物标\\output4\\tmp1'
|
||
#
|
||
# # 调用 process_folder 来处理整个文件夹中的所有文件
|
||
# process_folder(input_folder, output_folder)
|