337 lines
16 KiB
Python
337 lines
16 KiB
Python
import re
|
||
|
||
from PyPDF2 import PdfReader
|
||
|
||
from flask_app.货物标.截取pdf货物标版 import clean_page_content,extract_common_header
|
||
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,in_special_section):
|
||
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)
|
||
if in_special_section:
|
||
current_content.append('\n')
|
||
return append_newline
|
||
|
||
|
||
import re
|
||
|
||
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
|
||
last_main_number = None
|
||
temp_title = None # 临时存储以点号开头但不带数字的标题
|
||
|
||
# 定义所有需要的正则表达式模式
|
||
pattern_numbered = re.compile(r'^\s*([一二三四五六七八九十]{1,2})\s*、\s*') # 一、
|
||
pattern_parentheses = re.compile(r'^\s*[（(]\s*([一二三四五六七八九十]{1,2})\s*[)）]\s*') # （一）
|
||
pattern_letter_initial = re.compile(r'^([A-Z])\.?\s*(.*)$') # 初始扫描时匹配 A内容 或 A. 内容
|
||
pattern_letter = re.compile(r'^([A-Z])\.\s*(.*)$') # 主循环中严格匹配 A. 内容
|
||
pattern_arabic = re.compile(r'^(\d+、)\s*(.+)$') # 1、内容
|
||
|
||
initial_heading_pattern = None
|
||
special_section_keywords = ['文件的组成', '文件的构成','文件包括:','文件包括：','雷同认定','包括以下内容'] # 定义特殊章节关键词
|
||
in_special_section = False # 标志是否在特殊章节中
|
||
lines = text.split('\n')
|
||
|
||
def check_year_pattern(line):
|
||
# 检查是否为年份模式，如 '2014年'
|
||
line_no_spaces = line.replace(' ', '')
|
||
return re.match(r'^\d{4}\s*年', line_no_spaces) is not None
|
||
|
||
def is_heading(line):
|
||
# 检查是否匹配任何标题模式
|
||
patterns = [
|
||
r'^(?<![a-zA-Z（(])(\d+(?:\.\d+)+)\s*(.*)', # 匹配 '12.1 内容'
|
||
r'^(\d+\.)\s*(.+)$', # 匹配 '12. 内容'
|
||
r'^[．.](\d+(?:[．.]\d+)*)\s*(.+)$', # 匹配 '.12.1 内容'
|
||
r'^(\d+)([^.\d].*)' # 匹配 '27 内容'
|
||
]
|
||
for pattern in patterns:
|
||
if re.match(pattern, line):
|
||
return True
|
||
return False
|
||
|
||
# 预先扫描前5行，检查是否有匹配任何标题模式
|
||
first_five_lines = lines[:5]
|
||
has_initial_heading_patterns = False
|
||
for line in first_five_lines:
|
||
line_stripped = line.strip().replace('．', '.')
|
||
if (pattern_numbered.match(line_stripped) or
|
||
pattern_parentheses.match(line_stripped) or
|
||
pattern_letter_initial.match(line_stripped) or
|
||
pattern_arabic.match(line_stripped)):
|
||
has_initial_heading_patterns = True
|
||
break
|
||
|
||
for i, line in enumerate(lines):
|
||
line_stripped = line.strip().replace('．', '.')
|
||
|
||
# 如果在一个章节内，跳过年份模式的行
|
||
if check_year_pattern(line_stripped) and current_key is not None:
|
||
current_content.append(line_stripped)
|
||
continue
|
||
|
||
# 匹配带数字的标题，例如 '12.1 内容'
|
||
match = re.match(r'^(?<![a-zA-Z（(])(\d+(?:\.\d+)+)\s*(.*)', line_stripped)
|
||
if not match:
|
||
match = re.match(r'^(\d+\.)\s*(.+)$', line_stripped)
|
||
|
||
# 检查是否进入或退出特殊章节
|
||
if is_heading(line_stripped):
|
||
if any(keyword in line_stripped for keyword in special_section_keywords):
|
||
in_special_section = True
|
||
elif in_special_section:
|
||
in_special_section = False
|
||
|
||
# 以下是原有的匹配逻辑
|
||
# 匹配以点号开头并带有数字的情况，例如 '.12.1 内容'
|
||
dot_match = re.match(r'^[．.](\d+(?:[．.]\d+)*)\s*(.+)$', line_stripped)
|
||
|
||
# 匹配以点号开头但不带数字的情况，例如 '.投标文件的制作和签署'
|
||
dot_text_match = re.match(r'^[．.]\s*(\D.+)$', line_stripped)
|
||
|
||
# 匹配不带点号的纯数字开头的情况，例如 '27xxxxx'
|
||
pure_number_match = re.match(r'^(\d+)([^.\d)(）].*)', line_stripped) # 不允许出现右括号
|
||
|
||
if match:
|
||
new_key, line_content = match.groups()
|
||
line_content = line_content.lstrip('.．、,').replace(' ', '')
|
||
|
||
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
|
||
|
||
# 处理临时标题与新主标题的关联
|
||
if temp_title:
|
||
# 提取主编号（第一个点前的数字）
|
||
main_number = new_key.split('.')[0]
|
||
# 关联临时标题到主编号
|
||
data[f"{main_number}."] = temp_title
|
||
temp_title = None # 重置临时标题
|
||
|
||
# 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
|
||
current_key = new_key
|
||
current_content = [line_content]
|
||
append_newline = len(new_key.rstrip('.').split('.')) <= 2
|
||
last_main_number = new_key.split('.')[0]
|
||
else:
|
||
append_newline = handle_content_append(current_content, line_stripped, append_newline, keywords, in_special_section)
|
||
|
||
elif dot_match:
|
||
# 处理以点号开头并带有数字的情况
|
||
sub_number, line_content = dot_match.groups()
|
||
sub_number = sub_number.replace('．', '.')
|
||
if last_main_number:
|
||
new_key = f"{last_main_number}.{sub_number}"
|
||
if current_key is not None:
|
||
content_string = ''.join(current_content).strip()
|
||
data[current_key] = data.get(current_key, '') + content_string
|
||
current_key = new_key
|
||
current_content = [line_content]
|
||
append_newline = True
|
||
else:
|
||
append_newline = handle_content_append(current_content, line_stripped, append_newline, keywords, in_special_section)
|
||
|
||
elif dot_text_match:
|
||
# 处理以点号开头但不带数字的情况，存储到临时变量
|
||
temp_title = dot_text_match.group(1).strip()
|
||
continue # 跳过进一步处理该行
|
||
|
||
elif pure_number_match:
|
||
# 处理不带点号的纯数字开头的情况
|
||
new_key_candidate, line_content = pure_number_match.groups()
|
||
new_key_candidate += '.' # 添加点号
|
||
line_content = line_content.lstrip('.．、,').replace(' ', '')
|
||
|
||
# 提取当前键名和新键名的数值部分
|
||
def extract_number(key):
|
||
return float(key.rstrip('.').split('.')[0])
|
||
|
||
current_key_num = extract_number(current_key) if current_key else 0
|
||
new_key_num = extract_number(new_key_candidate)
|
||
|
||
# 判断新键名是否为新的标题，条件是键名递增且差值在5以内
|
||
if (current_key is None or (new_key_num > current_key_num and new_key_num - current_key_num <= 5)):
|
||
# 处理临时标题与新主标题的关联
|
||
if temp_title:
|
||
main_number = new_key_candidate.split('.')[0]
|
||
data[f"{main_number}."] = temp_title
|
||
temp_title = None # 重置临时标题
|
||
|
||
# 开始新的标题
|
||
if current_key is not None:
|
||
content_string = ''.join(current_content).strip()
|
||
data[current_key] = data.get(current_key, '') + content_string
|
||
current_key = new_key_candidate
|
||
current_content = [line_content]
|
||
append_newline = True
|
||
last_main_number = new_key_candidate.rstrip('.')
|
||
else:
|
||
# 将当前行视为当前标题的内容
|
||
append_newline = handle_content_append(current_content, line_stripped, append_newline, keywords, in_special_section)
|
||
|
||
else:
|
||
# 根据预先设置的标志决定是否执行这部分代码
|
||
if has_initial_heading_patterns and not skip_subheadings and not in_special_section:
|
||
numbered_match = pattern_numbered.match(line_stripped)
|
||
parentheses_match = pattern_parentheses.match(line_stripped)
|
||
letter_match = pattern_letter.match(line_stripped)
|
||
arabic_match = pattern_arabic.match(line_stripped)
|
||
|
||
# 判断当前行是否匹配了任何标题模式
|
||
if numbered_match or parentheses_match or letter_match or arabic_match:
|
||
# 如果初始标题模式尚未设置，则记录当前匹配的标题模式
|
||
if initial_heading_pattern is None:
|
||
if numbered_match:
|
||
initial_heading_pattern = 'numbered'
|
||
elif parentheses_match:
|
||
initial_heading_pattern = 'parentheses'
|
||
elif letter_match:
|
||
initial_heading_pattern = 'letter'
|
||
elif arabic_match:
|
||
initial_heading_pattern = 'arabic'
|
||
|
||
# 确定当前匹配的标题模式
|
||
if numbered_match:
|
||
current_heading_pattern = 'numbered'
|
||
elif parentheses_match:
|
||
current_heading_pattern = 'parentheses'
|
||
elif letter_match:
|
||
current_heading_pattern = 'letter'
|
||
elif arabic_match:
|
||
current_heading_pattern = 'arabic'
|
||
|
||
# 如果当前标题模式与初始标题模式一致，创建新的键值对
|
||
if current_heading_pattern == initial_heading_pattern:
|
||
# 保存之前的 key 的内容
|
||
if current_key is not None:
|
||
content_string = ''.join(current_content).strip()
|
||
data[current_key] = data.get(current_key, '') + content_string
|
||
if current_key_chinese is not None:
|
||
data[current_key_chinese] = current_value_chinese
|
||
current_key_chinese = None
|
||
|
||
# 处理匹配到的标题
|
||
if current_heading_pattern == 'numbered':
|
||
current_key_chinese = numbered_match.group(1)
|
||
current_value_chinese = line_stripped[numbered_match.end():].lstrip('.．、,').replace(' ', '')
|
||
elif current_heading_pattern == 'parentheses':
|
||
current_key_chinese = parentheses_match.group(1)
|
||
current_value_chinese = line_stripped[parentheses_match.end():].lstrip('.．、,').replace(' ', '')
|
||
elif current_heading_pattern == 'letter':
|
||
# 字母标题处理
|
||
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('.．、,').replace(' ', '')
|
||
elif current_heading_pattern == 'arabic':
|
||
arabic_key, arabic_value = arabic_match.groups()
|
||
current_key = arabic_key.replace('、', '.')
|
||
current_content = [arabic_value.replace(' ', '')]
|
||
append_newline = True
|
||
last_main_number = current_key.rstrip('.')
|
||
continue
|
||
else:
|
||
# 当前标题模式与初始模式不一致，将该行视为内容
|
||
if line_stripped:
|
||
append_newline = handle_content_append(current_content, line_stripped, append_newline, keywords, in_special_section)
|
||
else:
|
||
# 未匹配到任何标题模式，将该行视为内容
|
||
if line_stripped:
|
||
append_newline = handle_content_append(current_content, line_stripped, append_newline, keywords, in_special_section)
|
||
else:
|
||
# 在特殊章节中，所有内容都作为当前标题的内容
|
||
if line_stripped:
|
||
append_newline = handle_content_append(current_content, line_stripped, append_newline, keywords, in_special_section)
|
||
|
||
# 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
|
||
if current_key_chinese is not None:
|
||
data[current_key_chinese] = current_value_chinese
|
||
if temp_title:
|
||
# 处理任何未关联的临时标题
|
||
data["未分类标题"] = temp_title # 您可以根据需要选择合适的键名
|
||
|
||
return data
|
||
|
||
|
||
|
||
|
||
|
||
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)
|
||
exclusion_pattern = re.compile(r'文件的构成|文件的组成|须对应|需对应|须按照|需按照|须根据|需根据') #仅匹配第一页和最后一页，不需要exclusion_pattern
|
||
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)
|
||
# print(cleaned_text)
|
||
# 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:
|
||
matches = list(re.finditer(end_pattern, cleaned_text, re.MULTILINE))
|
||
if matches:
|
||
end_index = matches[-1].start()
|
||
cleaned_text = cleaned_text[:end_index]
|
||
|
||
all_pages_text.append(cleaned_text)
|
||
|
||
# 合并所有页面的文本
|
||
full_text = "\n".join(all_pages_text)
|
||
return full_text
|
||
|