電子發票太多，想統計下總額異常困難，網上工具不好用，花了2個小時實現一份，測試過中石油、京東開具的電子發票還行，部分發票名稱失敗有問題不影響統計，有需要的小伙伴自己拿去改吧。
import cmd
import sys
import json
import pdfplumber
import os
from pprint import pprint
class FapiaoShell(cmd.Cmd):
""" 發票 """
intro = '歡迎使用發票提取工具，輸入?(help)獲取幫助消息和命令列表，CTRL+C退出程序。\n'
prompt = '\n輸入命令: '
doc_header = "詳細文檔 (輸入 help <命令>):"
misc_header = "友情提示:"
undoc_header = "沒有幫助文檔:"
nohelp = "*** 沒有命令(%s)的幫助信息 "
def __init__(self):
super().__init__()
def do_load(self, arg):
""" 加載發票 例如：load D:\ """
if not os.path.isdir(arg):
print('參數必須是目錄!')
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)
print(f'發現pdf文件: {fpth}')
pdfs.append(fpth)
pdf_ctxs = self._parse_pdfs(pdfs)
total = {
'內容': pdf_ctxs,
'發票數': len(pdf_ctxs),
'總計': 0,
}
for fpth, info in pdf_ctxs:
total['總計'] += float(info['總計'])
print('\n保存到 結果.json...')
with open("結果.json", 'w', encoding='utf-8') as json_file:
json.dump(total,
json_file,
ensure_ascii=False,
sort_keys=True,
indent=4,
separators=(', ', ': '))
print('完成!')
def _parse_pdfs(self, pdfs):
""" 分析 """
result = []
for fpth in pdfs:
info = {}
with pdfplumber.open(fpth) as pdf:
page = pdf.pages[0]
if '增值稅電子普通發票' not in ''.join(page.extract_text()):
result.append((fpth, {}))
inf = self._extrace_from_words(page.extract_words())
info.update(inf)
inf = self._extrace_from_table(page.extract_tables()[0])
info.update(inf)
result.append((fpth, info))
return result
def _extrace_from_words(self, words):
""" 從單詞中提取 """
info = {}
lines = {}
for word in words:
top = int(word['top'])
bottom = int(word['bottom'])
pos = (top + bottom) // 2
text = word['text']
if pos not in lines:
lines[pos] = [text]
else:
lines[pos].append(text)
lines_pack = []
last_pos = None
for pos in sorted(lines):
arr = lines[pos]
if len(lines_pack) > 0 and pos - last_pos <= 10:
lines_pack[-1] += arr
continue
lines_pack.append(arr)
last_pos = pos
continue
for pack in lines_pack:
for idx, line in enumerate(pack):
if '電子普通發票' in line:
info['標題'] = line
continue
if '發票代碼:' in line:
info['發票代碼'] = line.split(':')[1]
continue
if '發票號碼:' in line:
info['發票號碼'] = line.split(':')[1]
continue
if '開票日期:' in line:
year = line.split(':')[1]
month = [ln for ln in pack if ln.isdigit()][0]
day = [ln[:2] for ln in pack if '日' in ln][0]
info['開票日期'] = f'{year}-{month}-{day}'
continue
if '機器編號:' in line:
info['機器編號'] = [ln for ln in pack if ln.isdigit()
and len(ln) > 10][0]
continue
if '碼:' in line:
c1 = pack[idx].split(':')[1]
c2 = pack[idx+1]
c3 = pack[idx+2]
c4 = pack[idx+3]
info['校驗碼'] = f'{c1} {c2} {c3} {c4}'
continue
if '收款人:' in line:
info['收款人'] = line.split(':')[1]
continue
if '開票人:' in line:
info['開票人'] = line.split(':')[1]
continue
return info
def _extrace_from_table(self, table):
""" 從表中提取 """
info = {}
if len(table) != 4:
return None
# 購買方
for cell in table[0]:
if not cell:
continue
lines = cell.splitlines()
for line in lines:
if '名 稱:' in line:
info['購買方名稱'] = line.split(':')[1]
continue
if len(line) == 18 and line.isalnum():
info['購買方稅號'] = line
continue
if len(line) == 27:
if '密碼' not in info:
info['密碼'] = []
info['密碼'].append(line)
continue
# 詳細
for cell in table[1]:
if not cell:
continue
lines = cell.splitlines()
for line in lines:
if '貨物或應稅勞務、服務名稱' in line:
info['商品'] = lines[1:-1]
break
if '金 額' in line:
info['總金額'] = lines[-1][1:]
break
if '稅 額' in line:
info['總稅額'] = lines[-1][1:]
break
# 合計
for cell in table[2]:
if not cell:
continue
lines = cell.splitlines()
for line in lines:
if '¥' in line:
info['總計'] = line[1:]
# 銷售方
for cell in table[3]:
if not cell:
continue
lines = cell.splitlines()
for line in lines:
if '名 稱:' in line:
info['銷售方名稱'] = line.split(':')[1]
continue
if len(line) == 18 and line.isalnum():
info['銷售方稅號'] = line
continue
return info
if __name__ == '__main__':
try:
FapiaoShell().cmdloop()
except KeyboardInterrupt:
print('\n\n再見！')