triple-extraxtion/pdfReader.py

import io
from pdfminer.converter import TextConverter
from pdfminer.pdfinterp import PDFPageInterpreter
from pdfminer.pdfinterp import PDFResourceManager
from pdfminer.pdfpage import PDFPage
import re
from get_result import get_ner_result, get_re_result
import requests 

# 将pdf文件中的文本提取到一个字符串中
def extract_text_from_pdf(pdf_path):
    if pdf_path.startswith('http'):  # 如果是从URL读取PDF文件
        # 从URL中读取PDF文件的二进制数据
        response = requests.get(pdf_path)
        fh = io.BytesIO(response.content)
    else:
        fh = open(pdf_path, 'rb')
    # 创建PDF资源管理器
    rsrcmgr = PDFResourceManager()
    # 创建一个字符串对象
    retstr = io.StringIO()
    # 创建一个PDF设备对象
    device = TextConverter(rsrcmgr, retstr, laparams=None)
    # 创建一个PDF解析器对象
    interpreter = PDFPageInterpreter(rsrcmgr, device)
    # 处理文档当中的每个页面
    for page in PDFPage.get_pages(fh, caching=True, check_extractable=True):
        interpreter.process_page(page)
    # 获取解析器生成的文本并关闭设备
    text = retstr.getvalue()
    device.close()
    retstr.close()
    if not fh.closed:
        fh.close()
    # 将文本字符串中的每行结尾的“\n”替换为空字符串
    text = re.sub(r'\n$', '', text, flags=re.MULTILINE)
    # 返回提取到的文本字符串
    return text

# 对文本进行分句
def split_sentences(text):
    # 可以根据实际情况修改分句的正则表达式
    sentence_pattern = re.compile(r'([。；！？\\?!])')
    sentences = sentence_pattern.split(text)
    # 将分句符与句子合并
    merged_sentences = []
    for i in range(0, len(sentences)-1, 2):
        merged_sentences.append(sentences[i] + sentences[i+1])
    if len(sentences) % 2 == 1:
        merged_sentences.append(sentences[-1])
    return merged_sentences