百度API批量识别表格与文字（记录）

1.安装百度识别python包。pip install baidu-aip

2.申请ID、KEY等。打开百度，查看全部产品，找到AI.

注册、登录百度智能云。

找到文字识别，创建应用，申请到ID及两个KEY：

3.简单代码实现。

from aip import AipOcr

#

""" 你的 APPID AK SK """

APP_ID = '1530xxxx'

API_KEY = 'fXp9dxxxxxxxxxxxxxxxxxxx'

SECRET_KEY = 'Ltvxxxxxxxxxxxxxxxxxxxxxxxxxxxxx'

client = AipOcr(APP_ID, API_KEY, SECRET_KEY)

#

""" 读取图片 """

def get_file_content(filePath):

    with open(filePath, 'rb') as fp:

        return fp.read()

 

path='C:/Users/Administrator/Desktop/ocr'#设定目录集中存放

img = get_file_content(path+'/'+u'eng2.png')

#

#文本识别

res=client.general(img)#结果为：字典》列表》字典三级结构，分行识别

strs=''

for item in res['words_result']:#提取结果

    strs+=item['words']

print(strs)

#

4.批量识别表格与文字。

# -*- coding: UTF-8 -*-

from aip import AipOcr

import time

import os

import requests

 

""" 你的 APPID AK SK """

APP_ID = '1544xxxx'

API_KEY = 'VAg0XiGxxxxxxxxxxxxxmoy4'

SECRET_KEY = '3GuciPY1Zxxxxxxxxxxxxxxxxxxxxxxo'

client = AipOcr(APP_ID, API_KEY, SECRET_KEY)

 

""" 读取图片 """

def get_file_content(filePath):

    with open(filePath, 'rb') as fp:

        return fp.read()

 

def write_txt(x_fname,x_fstr):#字符串写入文件

    #time1=time.strftime("%Y-%m-%d")

    wt=open(x_fname+'.txt','w')

    wt.write(x_fstr.encode('utf8'))

    print u'保存文件成功：'+x_fname+'.txt'

    wt.close()

    return

 

def file_download(url, wf_name):

    r = requests.get(url)

    print u'已获取网络对象'

    #print type(r)#调试点

    #print wf_name#调试点

    with open(wf_name+'.xls', 'wb') as f:

        f.write(r.content)

    print u'下载文件成功：'+wf_name+'.xls'

    return

 

def ocr(img,wf_name):

    res=client.tableRecognitionAsync(img)

    print res

    try:

        req_id=res['result'][0]['request_id']#表格图片分析后的ID，获取

        print u'ID已获取：',req_id

        while True:

            print u'暂停3秒...读取res2'

            time.sleep(3)

            res2=client.getTableRecognitionResult(req_id)#第二次：通过ID获取表格文件XLS地址

            time.sleep(1)

            print u'res2读取成功！',res2

            if res2.has_key('error_code'):#i不是表格,作为普通文本识别

                """ 调用通用文字识别, 图片为本地图片 """

                print u'表格识别错误，重新获取文本res'

                res=client.general(img)

                time.sleep(1)

                print u'重新获取文本res成功！'

                print res

                strs=''

                for item in res['words_result']:

                    strs+=item['words']

                write_txt(wf_name,strs)

                break

            else:

                try:

                    print u'是否完成msg'

                    msg=res2['result']['ret_msg']

                    if msg == u'已完成':

                        url = res2['result']['result_data']

                        print u'已获取下载地址：',url

                        print u'文件名：',wf_name

                        file_download(url, wf_name)

                        print u'下载完成！'

                        break

                    else:

                        time.sleep(1)

                except Exception as e:

                    print u'错误'

                    break

                else:

                    print u'等一哈...',

    except Exception as e:

        print u'出错了....'

 

def main():

    path='C:/Users/Administrator/Desktop/ocr'#设定目录集中存放

    isExists=os.path.exists(path) # 判断结果

    if not isExists:# 如果不存在则创建目录

         os.makedirs(path) # 创建目录操作函数

    print u'桌面OCR目录OK！'

    while True:

        inp=raw_input(u'请将文件拷贝到桌面OCR目录下，输入文件名：\n如果有多个文件，请在ocr目录下新建目录，文件放在该目录下，输入目录名即可(ctrl+c跳出)：')

        if len(inp)!=0 and os.path.isfile(path+'/'+inp):

            print u'找到一个文件！识别开始...'

            img = get_file_content(path+'/'+inp)

            print u'获取长度：'+str(len(img))

            wf_name =path+'/'+inp.split('.')[0]

            ocr(img,wf_name)

            break

        if len(inp)!=0 and os.path.isdir(path+'/'+inp):

            print u'找到一个目录！'

            path=path+'/'+inp

            list1=os.listdir(path)

            if len(list1)>0:

                print u'找到'+str(len(list1))+u'个文件！识别开始....'

                n=0

                for name in list1:

                     img = get_file_content(path+'/'+name)

                     wf_name=path+'/'+name.split('.')[0]

                     ocr(img,wf_name)

                     n+=1

                print u'共写'+str(n)+u'个文件！'

                break

        print u'文件不存在！'  

 

 

if __name__ == '__main__':

    main()