ICCV 2019 论文摘要提取

ICCV 1000多篇论文，用 python 提取出每篇论文的摘要。

论文改名，原文件名称是：作者_标题_ICCV_2019_paper.pdf， 去掉开始和结尾，缩短标题

import os
import glob

path = "F:\\BaiduNetdiskDownload\\ICCV2019"
files = glob.glob("F:\\BaiduNetdiskDownload\\ICCV2019\\**.pdf")

for file in files:
    base = file.split("\\")[3]
    author = base.split("_")[0]
    post = "_ICCV_2019_paper"
    main = base.replace(author+"_", "").replace(post, "").replace("_", " ")
    filename = os.path.join(path, main)
    os.rename(file, filename)

print('done')

提取摘要（少部分文章的排版不一样，未能提取到有效摘要）

# -*- coding: utf-8 -*-
'''
ICCV 2019 论文摘要提取
'''
import glob
import sys
import importlib

importlib.reload(sys)

from pdfminer.pdfparser import PDFParser, PDFDocument
from pdfminer.pdfinterp import PDFResourceManager, PDFPageInterpreter
from pdfminer.converter import PDFPageAggregator
from pdfminer.layout import *
from pdfminer.pdfinterp import PDFTextExtractionNotAllowed

# 以下 PDF 处理代码是从某篇 优快云 博文中抄来的，没记住原文链接
def parse(pdffile):
    fp = open(pdffile, 'rb')
    parser = PDFParser(fp)
    doc = PDFDocument()
    parser.set_document(doc)
    doc.set_parser(parser)
    doc.initialize()

    # 检测文档是否提供txt转换，不提供就忽略
    if not doc.is_extractable:
        raise PDFTextExtractionNotAllowed
    else:
        # 创建 PDf 资源管理器 来管理共享资源
        rsrcmgr = PDFResourceManager()
        # 创建一个PDF设备对象
        laparams = LAParams()
        device = PDFPageAggregator(rsrcmgr, laparams=laparams)
        # 创建一个PDF解释器对象
        interpreter = PDFPageInterpreter(rsrcmgr, device)

        for page in doc.get_pages():  # doc.get_pages() 获取page列表
            found = False
            interpreter.process_page(page)
            # 接受该页面的LTPage对象
            layout = device.get_result()
            for x in layout:
                if isinstance(x, LTTextBoxHorizontal):  # 获取文本内容
                    results = x.get_text()
                    abstract = str(results).replace("-\n", "").replace("\n", " ")
                    if found:
                        return abstract
                    else:
                        if abstract.startswith("Abstract"):
                            found = True

            # 只处理第一页上的 abstract
            return "-------- error --------"


if __name__ == '__main__':
    path = "F:\\BaiduNetdiskDownload\\ICCV2019"
    files = glob.glob("F:\\BaiduNetdiskDownload\\ICCV2019\\**.pdf")

    with open("abstract_ICCV2019.txt", "w", encoding="utf-8") as fout:
        for pdf in files:
            abstract = parse(pdf)
            basefile = pdf.split("\\")[-1]
            print(basefile)
            fout.write("Title: " + basefile + "\n")
            fout.write("Abstract: " + abstract + "\n\n")

    print("done!")