Python 破解文字验证码登录

# -*- coding: utf-8 -*-
import re
from selenium.webdriver import ActionChains
from selenium import webdriver
from PIL import Image
import json
import time
from BaiDu_SDK.aip.ocr import AipOcr
from PIL import Image
import pytesseract
from urllib.request import urlretrieve


class crack_text_captcha(object):
    def __init__(self, driver):
        super(crack_text_captcha, self).__init__()
        # selenium驱动
        self.driver = driver
        # 百度文字识别appid
        self.APP_ID = '17809029'
        # 百度文字识别appkey
        self.APP_KEY = 'XyoghidG1NXFuGu9jlkG9YmC'
        # 百度文字识别secretkey
        self.SECRET_KEY = '1RcuwsLprEKiYBkuGL9ef127MZc4jpeX'
        # 百度文字识别sdk客户端
        self.CLIENT = AipOcr(self.APP_ID, self.APP_KEY, self.SECRET_KEY)

    # 访问网站填写表单信息
    def fill_form(self,):
        # 填写登录表单信息并保存验证码图片
        user_element = self.driver.find_element_by_id('username1')
        user_element.clear()
        user_element.send_keys('13104966565')
        pass_element = self.driver.find_element_by_id('password1')
        pass_element.clear()
        pass_element.send_keys('@Ahjm456')
        valid_button = self.driver.find_element_by_id('selectyzm_text')
        valid_button.click()
        time.sleep(1)
        self.driver.save_screenshot('web_image.png')

    # 图片二值化便于识别
    def image_to_binary(self,):
        try:
            left = 1079
            top = 344
            right = 1540
            bottom = 668
            # 裁剪图
            img = Image.open('./web_image.png')
            img = img.crop((left, top, right, bottom))
            # 1.二值化转灰色
            pic = img.convert("L")
            # 2.二值化
            # pic = img.point(lambda p: p * 4)
            pic.save("grey_code.png")
            time.sleep(3)
        except Exception as e:
            print('图片二值化失败：%s' % str(e))

    # 读取本地图片
    def get_file(self,):
        file_path = "./grey_code.png"
        with open(file_path, 'rb') as pic:
            return pic.read()

    # 获取小图文字
    def target_words(self,):
        try:
            text_element = self.driver.find_element_by_id('selectyzm_text')
            match = re.compile('"(.*)"')
            result = match.findall(text_element.text)
            # print(result)
            return result
        except Exception as e:
            print("获取小图文字失败%s" % str(e))

    # 获取大图的文字和位置信息
    def position_info(self,text):
        try:
            op = {'language_type': 'CHN_ENG', 'recognize_granularity': 'small'}
            # 高精度
            # res = self.CLIENT.accurate(self.get_file(), options=op)
            time.sleep(20)
            # 通用
            res = self.CLIENT.general(self.get_file(), options=op)
            print("接口返回的位置信息:%s" % res)
            # 所有文字的位置信息
            all_pos_info = []
            # 要点击的文字位置信息
            target_pos_info = []
            for item in res['words_result']:
                all_pos_info.extend(item['chars'])
            # 按次序筛选出要点击的文字位置信息
            words = self.target_words()
            for word in words:
                for item in all_pos_info:
                    if word == item['char']:
                        all_pos_info.append(item)
            # return target_pos_info
            return all_pos_info
        except Exception as e:
            print("获取文字位置信息失败%s" % str(e))

    # 按顺序点击文字
    def click_words(self,location_info):
        # 获取到大图的element
        img_element = self.driver.find_element_by_xpath(
            "/html/body/div[2]/div[11]/div[2]/div[1]/form/div[1]/div[2]/img")
        # 根据小图文字在大图中的顺序依次点击
        for info in location_info:
            ActionChains(self.driver).move_to_element_with_offset(
                to_element=img_element, xoffset=info['location']['left'] + 20,
                yoffset=info['location']['top'] + 20).click().perform()
            time.sleep(1)

    # 下载数字验证码
    def pic_download(self,):
        IMAGE_URL = "http://cpquery.sipo.gov.cn/freeze.main?txn-code=createImgServlet&freshStept=1"
        urlretrieve(IMAGE_URL, './verify_code.png')

    # 两个数字加减验证码
    def crack_image_code(self,):
        # 下载验证码图片
        self.pic_download()
        # 图片转字符串
        result = pytesseract.image_to_string(Image.open('./verify_code.png'))
        str_result = result.encode('utf-8')
        try:
            # one_num=str_result[0]
            # two_num=str_result[2]
            # symbol=str_result[1]
            print(str_result)
            print(str_result[0])
            print(str_result[1])
            print(str_result[2])

            symbol = str_result[1]
            if symbol == '+':
                return (int(str_result[0]) + int(str_result[2]))
            else:
                return (int(str_result[0]) - int(str_result[2]))
        except Exception as e:
            print(str(e))
            return -99

    # 核心逻辑
    def main(self,):
        self.fill_form()
        self.image_to_binary()
        time.sleep(3)
        text = self.target_words()
        print(type(text))
        print("目标文字为：%s" % text)
        words_location = self.position_info(text)
        print("目标文字位置信息：%s" % words_location)
        # 由于识别不准确 字数不全 需要重新识别
        # while len(words_location) != 3:
        #     self.image_to_binary()
        #     time.sleep(3)
        #     self.target_words()
        #     self.position_info()
        # time.sleep(3)
        print("文字识别成功")
        self.click_words(words_location)
        # 点击登录按钮
        # log_in_button = self.driver.find_element_by_id('publiclogin')
        # log_in_button.click()
        # print("登录成功")
        # 同意声明
        # agree = self.driver.find_element_by_id('agreeid')
        # agree.click()
        # # 点击继续
        # go_on = self.driver.find_element_by_id('goBtn')
        # go_on.click()
        # # 输入专利号
        # input_patent_num = self.driver.find_element_by_id('select-key:shenqingh')
        # input_patent_num.clear()
        # # 填写专利号
        # input_patent_num.send_keys('2016800900500')
        # # 数字加减验证码的结果
        #
        # # 填写验证码
        # verify_code = self.driver.find_element_by_id('very-code')
        # verify_code.clear()
        # verify_code.send_keys()
        # # 点击查询按钮
        # query_btn = self.driver.find_element_by_id('query')
        # query_btn.click()

        # 保存查询结果

        # 调用接口传出数据


def unlock():
    driver = webdriver.Chrome()
    # 打开Chrome浏览器，需要将Chrome的驱动放在当前文件夹
    driver.maximize_window()
    driver.get('http://cpquery.sipo.gov.cn/')
    # 开始破解
    crack = crack_text_captcha(driver)
    crack.main()


if __name__ == '__main__':
    unlock()