#OCR 识别图形验证码

import pytesseract
from PIL import Image

def tesseract_basic(image_path: str) -> str:
    """直接用 Tesseract 识别纯基础款验证码"""
    # 打开图像
    img = Image.open(image_path)
    # 配置 Tesseract（限定字母数字组合、单块识别）
    config = r"--psm 7 --oem 3 -c tessedit_char_whitelist=ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789"
    # 识别并去除首尾空白
    return pytesseract.image_to_string(img, config=config).strip()

# 测试
if __name__ == "__main__":
    print(tesseract_basic("simple_captcha.png"))

import cv2
import numpy as np

def opencv_preprocess(image_path: str) -> np.ndarray:
    """轻量增强款验证码通用预处理"""
    # 1. 读取图像
    img = cv2.imread(image_path)
    # 2. 灰度化（减少计算量，突出明暗对比）
    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    # 3. 自适应阈值二值化（比固定阈值更适配背景渐变）
    thresh = cv2.adaptiveThreshold(
        gray, 255,
        cv2.ADAPTIVE_THRESH_GAUSSIAN_C,  # 高斯加权局部阈值
        cv2.THRESH_BINARY_INV,  # 黑白反转，让文字变成白色（255）
        11,  # 邻域块大小（必须是奇数，11是通用经验值）
        2    # 阈值偏移量（通用经验值）
    )
    # 4. 中值滤波去噪（适合去除椒盐噪点/干扰线断点）
    denoised = cv2.medianBlur(thresh, 3)
    # 5. 形态学闭操作（填补文字小缺口，可选但建议加）
    kernel = np.ones((2, 2), np.uint8)
    processed = cv2.morphologyEx(denoised, cv2.MORPH_CLOSE, kernel)
    
    # 调试时可以保存/显示预处理结果
    # cv2.imwrite("processed_captcha.png", processed)
    # cv2.imshow("Processed", processed)
    # cv2.waitKey(0)
    # cv2.destroyAllWindows()
    
    return processed

from paddleocr import PaddleOCR
import cv2

# 全局初始化一次 PaddleOCR（首次运行会自动下载中英文轻量模型，约10MB）
# use_angle_cls=True：识别旋转180°的文字
# lang="en"：优先英文识别；lang="ch"：中英文混合识别
ocr = PaddleOCR(use_angle_cls=True, lang="en", show_log=False)  # show_log=False 关闭冗余日志

def paddle_ocr_processed(processed_img: np.ndarray) -> str:
    """识别预处理后的验证码"""
    # 调用 OCR 接口，result 格式是 [[[[x1,y1],[x2,y2],[x3,y3],[x4,y4]], ('text', confidence)]]
    result = ocr.ocr(processed_img, cls=True)
    # 只取第一块文字（验证码通常是单块连续的）
    if result and result[0]:
        text, confidence = result[0][0][1]
        # 可选：过滤低置信度结果
        # if confidence < 0.8:
        #     return ""
        return text.strip()
    return ""

# 组合预处理+识别
def full_paddle_ocr(image_path: str) -> str:
    processed = opencv_preprocess(image_path)
    return paddle_ocr_processed(processed)

# 测试
if __name__ == "__main__":
    print(full_paddle_ocr("enhanced_captcha.png"))

from selenium import webdriver
from selenium.webdriver.common.by import By
from selenium.webdriver.support.ui import WebDriverWait
from selenium.webdriver.support import expected_conditions as EC
import time
import random
from full_paddle_ocr import full_paddle_ocr  # 假设上面的代码存为 full_paddle_ocr.py

def auto_login_scrape_center():
    """scrape.center 模拟登录全流程"""
    # 1. 初始化 Selenium 4 浏览器（推荐用 Service 统一管理驱动）
    options = webdriver.ChromeOptions()
    # 可选：伪装浏览器指纹（防简单的反爬）
    options.add_argument("user-agent=Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/125.0.0.0 Safari/537.36")
    options.add_experimental_option("excludeSwitches", ["enable-automation"])  # 隐藏自动化标识
    options.add_experimental_option("useAutomationExtension", False)
    
    service = webdriver.ChromeService()
    driver = webdriver.Chrome(service=service, options=options)
    
    try:
        # 2. 访问目标网站
        driver.get("https://captcha7.scrape.center/")
        time.sleep(random.uniform(1, 2))  # 随机延迟，模拟真人操作
        
        # 3. 输入用户名密码
        username_input = WebDriverWait(driver, 10).until(
            EC.presence_of_element_located((By.NAME, "username"))
        )
        username_input.send_keys("admin")
        time.sleep(random.uniform(0.5, 1))
        
        password_input = driver.find_element(By.NAME, "password")
        password_input.send_keys("admin")
        time.sleep(random.uniform(0.5, 1))
        
        # 4. 获取并识别验证码
        captcha_img = driver.find_element(By.CSS_SELECTOR, ".captcha img")
        # 直接用 element.screenshot() 截图，避免整页截图后裁剪
        captcha_img.screenshot("current_captcha.png")
        code = full_paddle_ocr("current_captcha.png")
        print(f"识别到的验证码：{code}")
        
        # 5. 输入验证码并提交
        captcha_input = driver.find_element(By.NAME, "captcha")
        captcha_input.send_keys(code)
        time.sleep(random.uniform(0.5, 1))
        
        submit_btn = driver.find_element(By.CSS_SELECTOR, "button[type='submit']")
        submit_btn.click()
        
        # 6. 验证登录成功
        WebDriverWait(driver, 10).until(
            EC.text_to_be_present_in_element((By.TAG_NAME, "h2"), "登录成功")
        )
        print("🎉 登录成功！")
        
    finally:
        # 7. 关闭浏览器
        time.sleep(2)
        driver.quit()

if __name__ == "__main__":
    auto_login_scrape_center()