python_invoices/test.py

import json
import pdfplumber
import os


class FapiaoShell():
    # 适用电子发票
    def __init__(self, path):
        self.do_load(path)

    def do_load(self, arg):
        """ 加载发票文件夹"""
        if not os.path.isdir(arg):
            return
        os.chdir(os.path.dirname(arg))
        pdfs = []
        for root, _, files in os.walk(arg):
            for fn in files:
                ext = os.path.splitext(fn)[1].lower()
                if ext != '.pdf':
                    continue
                fpth = os.path.join(root, fn)
                fpth = os.path.relpath(fpth)
                pdfs.append(fpth)

        print(f'pdf文件: {pdfs}')
        pdf_ctxs = self._parse_pdfs(pdfs)
        # 导出json文件
        with open('test.json', 'w', encoding='utf-8') as F:
            json.dump(pdf_ctxs, F, ensure_ascii=False)
        print('完成!')

    def _parse_pdfs(self, pdfs):
        """ 分析 """
        result = []
        for fpth in pdfs:
            with pdfplumber.open(fpth) as pdf:
                page = pdf.pages[0]
                # 空格替换，转成list
                list = ''.join(page.extract_text()).replace(
                    ' ', '').split('\n')
                result.append(list)

        return result


if __name__ == '__main__':
    FapiaoShell('./pd')