python: 用百度API读取增值税发票信息

 

# encoding: utf-8
# 版权所有 2023 涂聚文有限公司
# 许可信息查看：
# 描述：
# Author    : geovindu,Geovin Du 涂聚文.
# IDE       : PyCharm 2023.1 python 311
# Datetime  : 2023/9/30 6:56
# User      : geovindu
# Product   : PyCharm
# Project   : pythonTkinterDemo
# File      : BaiduOCRAPI.py
# explain   : 学习

'''
{'words_result_num': 46, 'words_result': {'InvoiceNumDigit': '', 'CommodityUnit': [{'row': '1', 'word': '个'}], 'PurchaserAddress': '', 'SheetNum': '', 'CommodityType': [], 'TotalAmount': '100.83', 'Checker': '李思', 'PurchaserBank': '', 'Agent': '否', 'Password': '*71<09/<5*61/*67/+5-5>0>876-4794<3/2*802209-<->*7/973<52466>1-7<74+86*/*>1882>1-1>87<86588<6>0<803719>+895-*', 'InvoiceTypeOrg': '海南增值税电子普通发票', 'InvoiceCodeConfirm': '046002200111', 'TotalTax': '9.07', 'ServiceType': '日用品食品', 'CommodityTaxRate': [{'row': '1', 'word': '9%'}], 'CommodityTax': [{'row': '1', 'word': '9.07'}], 'SellerBank': '中国工商银行股份有限公司三亚解放支行2201030109200160703', 'Remarks': '订单号:278170028513', 'SellerAddress': '海南省三亚市崖州区崖州湾科技城标准厂房二期三楼C274区22165500-1602/1699', 'NoteDrawer': '王梅', 'InvoiceTag': '其他', 'InvoiceNumConfirm': '82098742', 'OnlinePay': '', 'Payee': '王陆', 'CommodityName': [{'row': '1', 'word': '*其他食品*食品'}], 'CommodityVehicleType': [], 'InvoiceCode': '046002200111', 'AmountInWords': '壹佰零玖圆玖角', 'AmountInFiguers': '109.90', 'City': '', 'InvoiceType': '电子普通发票', 'CommodityEndDate': [], 'PurchaserName': '重庆海开科技有限公司', 'InvoiceDate': '2023年07月31日', 'CommodityNum': [{'row': '1', 'word': '1'}], 'PurchaserRegisterNum': '915001075828145135', 'MachineCode': '661719672092', 'CommodityPlateNum': [], 'CheckCode': '47587549392161874692', 'SellerRegisterNum': '91460200MA5T41103U', 'CommodityPrice': [{'row': '1', 'word': '100.83'}], 'CommodityStartDate': [], 'SellerName': '三亚京东佳禹贸易有限公司', 'CommodityAmount': [{'row': '1', 'word': '100.83'}], 'Province': '海南省', 'InvoiceNum': '82098742'}, 'pdf_file_size': 1, 'log_id': 1707917072629791560}
Json {'words_result_num': 46, 'words_result': {'InvoiceNumDigit': '', 'CommodityUnit': [{'row': '1', 'word': '个'}], 'PurchaserAddress': '', 'SheetNum': '', 'CommodityType': [], 'TotalAmount': '100.83', 'Checker': '李思', 'PurchaserBank': '', 'Agent': '否', 'Password': '*71<09/<5*61/*67/+5-5>0>876-4794<3/2*802209-<->*7/973<52466>1-7<74+86*/*>1882>1-1>87<86588<6>0<803719>+895-*', 'InvoiceTypeOrg': '海南增值税电子普通发票', 'InvoiceCodeConfirm': '046002200111', 'TotalTax': '9.07', 'ServiceType': '日用品食品', 'CommodityTaxRate': [{'row': '1', 'word': '9%'}], 'CommodityTax': [{'row': '1', 'word': '9.07'}], 'SellerBank': '中国工商银行股份有限公司三亚解放支行2201030109200160703', 'Remarks': '订单号:278170028513', 'SellerAddress': '海南省三亚市崖州区崖州湾科技城标准厂房二期三楼C274区22165500-1602/1699', 'NoteDrawer': '王梅', 'InvoiceTag': '其他', 'InvoiceNumConfirm': '82098742', 'OnlinePay': '', 'Payee': '王陆', 'CommodityName': [{'row': '1', 'word': '*其他食品*食品'}], 'CommodityVehicleType': [], 'InvoiceCode': '046002200111', 'AmountInWords': '壹佰零玖圆玖角', 'AmountInFiguers': '109.90', 'City': '', 'InvoiceType': '电子普通发票', 'CommodityEndDate': [], 'PurchaserName': '重庆海开科技有限公司', 'InvoiceDate': '2023年07月31日', 'CommodityNum': [{'row': '1', 'word': '1'}], 'PurchaserRegisterNum': '915001075828145135', 'MachineCode': '661719672092', 'CommodityPlateNum': [], 'CheckCode': '47587549392161874692', 'SellerRegisterNum': '91460200MA5T41103U', 'CommodityPrice': [{'row': '1', 'word': '100.83'}], 'CommodityStartDate': [], 'SellerName': '三亚京东佳禹贸易有限公司', 'CommodityAmount': [{'row': '1', 'word': '100.83'}], 'Province': '海南省', 'InvoiceNum': '82098742'}, 'pdf_file_size': 1, 'log_id': 1707917072629791560}

'''

import os
import base64
import requests
import pandas as pd
import json



class BaiduOCR(object):
    """
    利用百度API读取发票信息（pdf,image文件）
    """

    AppID="40226401"
    APIKey="QuXMNizc80gTmUznKDRqQX3D"
    SecretKey="h6aHaGLssw51CYGtR3dvX1wGg6BBm0zi"

    def __init__(self):
        """

        """
        self.AppID="40226401"
        self.APIKey="SeP54f3RO7GqifYpX8DPQGQE"  #QuXMNizc80gTmUznKDRqQX3D
        self.SecretKey="c2zXHOWM2hlxeECEgwqG6UWlLTN1kQRs"  #h6aHaGLssw51CYGtR3dvX1wGg6BBm0zi



    def getAccessToken(self):
        """

        :param APIKey:
        :param SecretKey:
        :return:
        """
        '''
        host = f"https://aip.baidubce.com/oauth/2.0/token?client_secret={self.SecretKey}&grant_type=client_credentials&client_id={self.APIKey}"
        response = requests.get(host)
        return response.json()['access_token']
        '''
        url = "https://aip.baidubce.com/oauth/2.0/token"
        params = {"grant_type": "client_credentials", "client_id": self.APIKey, "client_secret": self.SecretKey}
        return str(requests.post(url, params=params).json().get("access_token"))

    def getContent(self,accessToken, pdfFile):
        """

        :param accessToken
        :param pdfFile:
        :return:
        """
        #headers = {'content-type': 'application/x-www-form-urlencoded'}
        #request_url = f"https://aip.baidubce.com/rest/2.0/ocr/v1/vat_invoice?access_token={accessToken}"
        f = open(pdfFile, 'rb')
        pdf = base64.b64encode(f.read())
        print(pdf)
        print(accessToken)
        request_url = f"https://aip.baidubce.com/rest/2.0/ocr/v1/vat_invoice"
        params = {"pdf_file": pdf}
        access_token =accessToken # '[调用鉴权接口获取的token]'
        request_url = request_url + "?access_token=" + access_token
        headers = {'content-type': 'application/x-www-form-urlencoded'}
        response = requests.post(request_url, data=params, headers=headers)
        if response:
            print(response.json())

        #print(pdf)
        #params = {"pdf_file": pdf}
        #response = requests.post(request_url, data=params, headers=headers)
        #print(response.json())
        return response.json()


    def getContentPng(self,accessToken, pngFile):
        """

        :param accessToken
        :param pngFile:
        :return:
        """
        #headers = {'content-type': 'application/x-www-form-urlencoded'}
        #request_url = f"https://aip.baidubce.com/rest/2.0/ocr/v1/vat_invoice?access_token={accessToken}"
        f = open(pngFile, 'rb')
        pdf = base64.b64encode(f.read())
        print(pdf)
        print(accessToken)
        request_url = f"https://aip.baidubce.com/rest/2.0/ocr/v1/vat_invoice"
        params = {"image": pdf}
        access_token =accessToken # '[调用鉴权接口获取的token]'
        request_url = request_url + "?access_token=" + access_token
        headers = {'content-type': 'application/x-www-form-urlencoded'}
        response = requests.post(request_url, data=params, headers=headers)
        if response:
            print(response.json())

        #print(pdf)
        #params = {"pdf_file": pdf}
        #response = requests.post(request_url, data=params, headers=headers)
        #print(response.json())
        return response.json()

    def getUsefulInfo(self,content, pdfName):
        """

        :param content
        :param pdfName:
        :return:
        """
        jsonstr = content
        print("Json",jsonstr)
        words_result = jsonstr['words_result']
        info = {'发票文件名': pdfName,
                '发票号码': str(words_result['InvoiceNum']),
                '开票日期': words_result['InvoiceDate'],
                '货物名称': words_result['CommodityName'][0]['word'],
                '未税金额': words_result['CommodityAmount'][0]['word'],
                '货物税率': words_result['CommodityTaxRate'][0]['word'],
                '货物税额': words_result['CommodityTax'][0]['word'],
                '合计金额': words_result['TotalAmount'],
                '合计税额': words_result['TotalTax'],
                '价税合计（小写）': words_result['AmountInFiguers'],
                '价税合计（大写）': words_result['AmountInWords'],
                '销售方名称': words_result['SellerName'],
                '销售方纳税人识别号': words_result['SellerRegisterNum'],
                '销售方银行及账户': words_result['SellerBank'],
                '销售方地址及电话': words_result['SellerAddress'],
                '购买方名称':words_result['PurchaserName'],
                '购买方纳税人识别号':words_result['PurchaserRegisterNum'],
                '机器编号':words_result['MachineCode']
                }
        return info

　　

 

调用：用京东多张发票测试成功

 

    ocr=Common.BaiduOCRAPI.BaiduOCR()
    pdfFilelist = os.listdir("invoice/")
    infolist = []
    for pdfFile in pdfFilelist:
        if pdfFile.split(".")[-1] == 'pdf':
            pdfName = pdfFile.split(".")[:-1]
            print(pdfFile)
            access_token =ocr.getAccessToken()
            content = ocr.getContent(access_token, "invoice/" + pdfFile)
            info = ocr.getUsefulInfo(content, pdfName)
            infolist.append(info)

    df = pd.DataFrame(infolist)
    print(df)
    #df.to_excel('增值税发票信息统计.xlsx', sheet_name="geovindu",index=False)
    with pd.ExcelWriter('geovindu.xlsx') as writer:  #, mode='a' 附加
        df.to_excel(writer, sheet_name='geovindu', index=False)