本人小公司企业主一枚，经常接受电子发票，然后有空发给会计，默认的发票基本都是发票号，看不出是哪个公司开的，哪个公司收的。
经过多次研究使用pymupdf读取pdf格式，但是不同省份和城市开出的电子发票细节格式并不相同，最终是按区域先读取再细分信息。
读取多数信息了，至于改名规则可以根据喜好修改。如果有相同的发票会在重复的文件前面加上下划线
_
便于删除
目前广东省，厦门市，浙江省开出的发票测试无误，本程序有一定的容错机制其他省份的大概率是没问题，除非格式特殊
ps：不支持图片及完全曲线化的pdf电子发票（无字体pdf）
python 3.7 及fitz pymupdf ver =1.18.14下运行
# _*_ coding:utf-8 _*_
# fitz pymupdf ver =1.18.14
# Python 3.7
# 2021-11-17
# by gingeer @ CSDN
# 按位置获取发票信息并改名
# pymupdf doc see https://pymupdf.readthedocs.io/en/latest/
## make a new pdf with same size of old one.
##doc = fitz.open("some.file")
##page = doc[0]
##paths = page.get_drawings() # extract existing drawings
### this is a list of "paths", which can directly be drawn again using Shape
### -------------------------------------------------------------------------
###
### define some output page with the same dimensions
##outpdf = fitz.open()
##outpage = outpdf.new_page(width=page.rect.width, height=page.rect.height)
##shape = outpage.new_shape() # make a drawing canvas for the output page
import sys
import fitz
import re
import os, math
import datetime
# pip install pyMuPDF not fitz!!!!!!!!!!!!!
def make_text(words):
"""Return textstring output of get_text("words").
Word items are sorted for reading sequence left to right,
top to bottom.
"""
line_dict = {} # key: vertical coordinate, value: list of words
words.sort(key=lambda w: w[0]) # sort by horizontal coordinate
for w in words: # fill the line dictionary
y1 = round(w[3], 1) # bottom of a word: don't be too picky!
word = w[4] # the text of the word
line = line_dict.get(y1, []) # read current line content
line.append(word) # append new word
line_dict[y1] = line # write back to dict
lines = list(line_dict.items())
lines.sort() # sort vertically
lines_2=[" ".join(line[1]) for line in lines]
return lines_2
## return "\n".join(lines_2) #原方法
pdfPath='./'
pagesfiles=[]
newnames=[]
files = os.listdir(pdfPath)
pdffiles = [f for f in files if f.lower().endswith('.pdf')]
##print(pdffiles)
#大区域按百分比划分
invoice_rect_per = [0.69,0.02,0.978,0.2]
buyer_rect_per = [0.16,0.214, 0.57,0.369]
seller_rect_per = [0.16,0.744, 0.57,0.884]
total_amounts_per =[0.666,0.692, 0.95,0.749]
for pdffile in pdffiles:
print(pdffile)
doc = fitz.open(pdfPath+pdffile)
page = doc[0]
## print(page)
x_end=page.rect[2]
y_end=page.rect[3]
shape=page.new_shape()
## print(x_end,y_end,page.rect)
shape.draw_rect(page.rect)
words = page.get_text("words") # list of words on page
##===debug===
## for word in words:
## print(word[0]/x_end,word[1]/y_end,word[2]/x_end,word[3]/y_end,word[4])
## print(word)
## text_point=fitz.Point(word[0],word[3])
## shape.draw_circle(text_point,1)
## shape.insertText(text_point,word[4],color=(1,0,0),fontsize=8)
#===debug end===
invoive_rect = fitz.Rect(invoice_rect_per[0] * x_end,invoice_rect_per[1] * y_end, invoice_rect_per[2] * x_end,invoice_rect_per[3] * y_end)
buyer_rect = fitz.Rect(buyer_rect_per[0] * x_end,buyer_rect_per[1] * y_end,buyer_rect_per[2] * x_end,buyer_rect_per[3] * y_end)
seller_rect = fitz.Rect(seller_rect_per[0] * x_end,seller_rect_per[1] * y_end,seller_rect_per[2] * x_end,seller_rect_per[3] * y_end)
total_amounts_rect = fitz.Rect(total_amounts_per[0] * x_end,total_amounts_per[1] * y_end,total_amounts_per[2] * x_end,total_amounts_per[3] * y_end)
shape.draw_rect(total_amounts_rect) ## check positions
#-----获取重要信息-------
#----发票信息
mywords_1 = [w for w in words if fitz.Rect(w[:4]) in invoive_rect]
mywords_2 = [w for w in words if fitz.Rect(w[:4]).intersects(invoive_rect)]
invoice_infos = make_text(mywords_1)
check_invoice=False
for info in invoice_infos:
if re.search('\d{12}',info) != None:
invoice_cata = info
check_invoice=True
elif re.search('\d{8}',info) != None:
invoice_num = info
check_invoice=True
elif re.search('(\d{4})\D+(\d{2})\D+(\d{2})',info) !=None:
date_re_result=re.search('(\d{4})\D+(\d{2})\D+(\d{2})',info)
invoice_date='-'.join([date_re_result.group(1),date_re_result.group(2),date_re_result.group(3)])
check_invoice=True
if not check_invoice:
print('不是发票文件',invoice_infos)
# ===debug===
## shape.finish(width=0.5,color=(1,0,0))
## shape.commit()
## doc.save('xjj.pdf')
## doc.close()
## dpi=200
## zoom_x = dpi/72
## zoom_y = dpi/72
## mat = fitz.Matrix(zoom_x, zoom_y).preRotate(0)
## pix = page.getPixmap(matrix=mat, alpha=False)
## pix.set_dpi(dpi,dpi)
## pix.writePNG(pdffile + '.png')
# ===debug end===
continue
else:
invoice_num=invoice_num.replace('发票号码','')
invoice_num=invoice_num.replace('：','')
invoice_num=invoice_num.replace(':','')
invoice_num=invoice_num.strip()
#----购方
buyer_words = [w for w in words if fitz.Rect(w[:4]) in buyer_rect]
buyer_infos = make_text(buyer_words)
for info in buyer_infos:
if info.find('公司') >= 0:
buyer=info
break
elif info =='个人':
buyer='个人'
#----卖方
seller_words = [w for w in words if fitz.Rect(w[:4]) in seller_rect]
seller_infos = make_text(seller_words)
for info in seller_infos:
if len(info) >3:
seller=info
break
#----合计金额
total_amount_words = [w for w in words if fitz.Rect(w[:4]) in total_amounts_rect]
total_amount_infos = make_text(total_amount_words)
total_amount=total_amount_infos[-1]
total_amount=total_amount.replace('¥','')
total_amount=total_amount.replace('￥','')
#----新名字
connector='_'
newname_elements=[invoice_date,invoice_num,buyer+'(购)',seller,'￥'+total_amount]
newname=connector.join(newname_elements) + '.pdf'
## doc.save(pdffile)
if newname == pdffile:
print('发票名已经正规化！')
newnames.append(newname)
doc.close()
print('------------------------')
continue
if newname not in newnames and newname not in pdffiles:
print(f'原名：{pdffile}\n新名：{newname}')
newnames.append(newname)
doc.close()
os.rename(pdffile,newname)
else:
while newname in newnames:
newname = '_'+newname
newnames.append(newname)
doc.close()
os.rename(pdffile,'_'+newname)
print('重复文件名!已加前缀')
print('------------------------')
# ===debug===
## shape.finish(width=0.5,color=(1,0,0))
## shape.commit()
## dpi=200
## zoom_x = dpi/72
## zoom_y = dpi/72
## mat = fitz.Matrix(zoom_x, zoom_y).preRotate(0)
## pix = page.getPixmap(matrix=mat, alpha=False)
## pix.set_dpi(dpi,dpi)
## pix.writePNG(pdffile + '.png')
# ===debug end===
# print(newname)
版权声明：本文为gingeers原创文章，遵循 CC 4.0 BY-SA 版权协议，转载请附上原文出处链接和本声明。