adb工具截取手机屏幕，发送截图到电脑，PHP截取文字部分，调用Google的开源的OCR识别工具Tesseract-OCR，转为文字，调用百度搜索题目和答案，根据题目加选项的结果数量和首页出现的次数，综合判断给出答案。
浏览器输出结果
<?php
class SearchAnswer
{
const WORK_DIR = __DIR__;
//tesseract 路径
private static $tesseract_path = '"'. WORK_DIR.'"\tool\Tesseract-OCR\tesseract" ';
//语言包目录和参数
private static $tessdata = ' --tessdata-dir "D:\dev\Tesseract-OCR\tessdata" ';
private static $lang = ' -l chi_sim --psm 6 ';
//截图文字区域
//private static $question_pos = array(80, 320, 1000, 1200); //百万英雄
private static $question_pos = array(120, 320, 1000, 1200); //百万英雄
//$question_pos = array(80, 400, 1000, 1300); //腾讯
//$question_pos = array(80, 320, 1000, 1100); //百度 好看视频
const BROWSER = '"C:\Program Files (x86)\Google\Chrome\Application\chrome.exe"';
private static $negative_words_arr = ['不是', '没有', '不属于', '不在', '不会', '不需要', '不包含', '不准确', '不可能', '不途经', '不包括', '不正确', '不作为', '错误', '不适合', '不应该', '不相当', '无关', '不相符'];
/**
* 手机截图并推送到pc
* @return string
*/
/*因为PC端adb默认开启USB调试的，故这里必须先杀掉adb服务，如下命令：
adb kill-server
然后，通过命令连接设备ip和端口，如上图所示：
adb connect 192.168.1.4:5555
最后，如果要断开连接，则
adb disconnect*/
static function screencap()
{
$new_file = self::WORK_DIR . '/question/screen.png';
shell_exec('adb shell screencap -p | perl -pe "s/\x0D\x0A/\x0A/g" > ' . $new_file); //cmder 下
return $new_file;
}
/**
* 截取手机截图的问题和选项区域，并生成图片
* @param $src_img
* @param string $file_name
* @return string
*/
static function getImg($src_img, $file_name = 'question')
{
$dst_w = self::$question_pos[2] - self::$question_pos[0];
$dst_h = self::$question_pos[3] - self::$question_pos[1];
//只处理png截图 剪裁
$src_im = imagecreatefrompng($src_img);
$dst_im = imagecreate($dst_w, $dst_h);
//$dst_im = imagecreatetruecolor($dst_w, $dst_h);
imagecopy($dst_im, $src_im, 0, 0, self::$question_pos[0], self::$question_pos[1], $dst_w, $dst_h);
//$file_name = dirname($src_img) . "/$file_name.png";
$file_name2 = self::WORK_DIR . '/question/' . $file_name . '.png';
imagepng($dst_im, $file_name2);
imagedestroy($dst_im);
return $file_name2;
}
/**
* 获取图片的题目和选项
* @param $ocr_img
* @param string $ocr_text
* @return array 题目和选项
*/
static function getOcrText($ocr_img, $ocr_text = 'question')
{
$output = self::WORK_DIR . "/question/$ocr_text";
$tessdata_cmd = self::$tesseract_path . $ocr_img . ' ' . $output . self::$tessdata . self::$lang . ' >log.log 2>log2.log';
shell_exec($tessdata_cmd);
$text = file_get_contents($output . '.txt');
$text = trim($text, "\n\f");
//一 识别为_
$text = str_replace([' ', '_'], ['', '一'], $text);
$text_arr = array_filter(explode("?\n", $text));
$question = preg_replace('/^([0-9oO]+\.)/', '', $text_arr[0]);
//$choices_arr = explode("\n", $text_arr[1]);
$choices_arr = array_filter(explode("\n", $text_arr[1]), function ($var) {
if (!empty($var)) {
return $var;
} else {
return false;
}
});
$choices_arr = array_values($choices_arr);
return array('question' => str_replace(["\n"], [''], $question), 'choices' => $choices_arr);
}
// function array_empty($var)
// {
// if (!empty($var)) {
// return $var;
// } else {
// return false;
// }
// }
static function search_result($content, $info)
{
//$content = "百度为您找到相关结果约41,900,000个";
//$content = getCurlSearch($url);
//$content = curl_get_https($url);
$pattern = "/百度为您找到相关结果约([\d\,]*)个/";
preg_match($pattern, $content, $match);
$search_amount['choices'] = $info;
$search_amount['count'] = str_replace(',', '', $match[1]);
$search_amount['count2'] = substr_count($content, $info);
return $search_amount;
}
/**
* 打开浏览器搜索
* @param $info
*/
static function methodOne($info)
{
$baidu = ' http://www.baidu.com/s?wd="' . $info['question'] . '"';
system(self::BROWSER . $baidu);
}
static function methodTwo($info)
{
//问题去掉否定词
$negative_words_arr2 = self::negativeWords(self::$negative_words_arr);
$question = str_replace(self::$negative_words_arr, [' '], $info['question'], $negative_words_count);
$title = '<h2 class="round"> </h2>';
$question2 = $info['question'];
if ($negative_words_count > 0) {
$title = "<h4 style='color: red;' class='title'>否定词</h4>";
$question2 = str_replace(self::$negative_words_arr, $negative_words_arr2, $info['question']);
}
//$title .= '<h3 class="title">' . $info['question'] . '</h3>';
$title .= '<h3 class="title">' . $question2 . '</h3>';
$options = '';
$url_arr = [];
foreach ($info['choices'] as $key => $val) {
//$param = $info['question'] . ' +' . $val;
$param = $question . ' +' . $val;
$param = urlencode($param);
$url_arr[] = 'https://www.baidu.com/s?wd=' . $param;
}
$result = self::multi_get($url_arr, $info);
$count = [];
$count = array_column($result, 'count');
$max = max($count);
$min = min($count);
$count2 = array_column($result, 'count2');
$max2 = max($count2);
$min2 = min($count2);
foreach ($result as $k => $v) {
$this_answer = '';
$answer_one = '';
$answer_two = '';
if ($negative_words_count == 0 && $v['count2'] == $max2) {
$answer_two = 'class="two"';
$this_answer = 'class="this_answer"';
} elseif ($negative_words_count > 0 && $v['count2'] == $min2) {
$answer_two = 'class="two"';
$this_answer = 'class="this_answer"';
}
if ($negative_words_count == 0 && $v['count'] == $max) {
$answer_one = 'class="one"';
} elseif ($negative_words_count > 0 && $v['count'] == $min) {
$answer_one = 'class="one"';
}
$options .= "<p class='answer'><span $this_answer>{$v['choices']} </span> <span class='score'><span $answer_one>{$v['count']} </span> 首页：<span $answer_two>{$v['count2']} </span></span></p>";
}
echo $title . $options;
}
/**
* @param $url
* @return mixed|string
*/
static function getCurlSearch($url)
{
$content = "错误信息";
if (function_exists('curl_init')) {
$ch = curl_init();
curl_setopt($ch, CURLOPT_URL, $url);
curl_setopt($ch, CURLOPT_TIMEOUT, 5);
//curl_setopt($ch, CURLOPT_USERAGENT, _USERAGENT_);
curl_setopt($ch, CURLOPT_REFERER, "http://www.baidu.com/");
curl_setopt($ch, CURLOPT_RETURNTRANSFER, 1);
//curl_setopt($ch, CURLOPT_COOKIESESSION, true);
curl_setopt($ch, CURLOPT_USERAGENT, "Mozilla/5.0 (Windows NT 10.0; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/55.0.2883.87 Safari/537.36");
$content = curl_exec($ch);
curl_close($ch);
}
return $content;
}
/**
* @param $url
* @return mixed
*/
static function curl_get_https($url)
{
$curl = curl_init(); // 启动一个CURL会话
curl_setopt($curl, CURLOPT_URL, $url);
curl_setopt($curl, CURLOPT_HEADER, 0);
curl_setopt($curl, CURLOPT_RETURNTRANSFER, 1);
curl_setopt($curl, CURLOPT_SSL_VERIFYPEER, false); // 跳过证书检查
//curl_setopt($curl, CURLOPT_SSL_VERIFYHOST, true); // 从证书中检查SSL加密算法是否存在
curl_setopt($curl, CURLOPT_USERAGENT, "Mozilla/5.0 (Windows NT 10.0; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/55.0.2883.87 Safari/537.36");
$tmpInfo = curl_exec($curl); //返回api的json对象
//关闭URL请求
curl_close($curl);
return $tmpInfo; //返回json对象
}
/**
* @param $url
* @param $info
* @return array
*/
static function multi_get($url, $info)
{
$chs = array();
$result = array();
$mh = curl_multi_init();
foreach ($url as $val) {
$ch = curl_init();
$chs[] = $ch;
//$this->prepare($ch);
curl_setopt($ch, CURLOPT_URL, $val);
//curl_setopt($ch, CURLOPT_POST, 1);
curl_setopt($ch, CURLOPT_HEADER, false);
curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, false);
//curl_setopt($ch, CURLOPT_HTTPHEADER, $headers);
curl_setopt($ch, CURLOPT_USERAGENT, "Mozilla/5.0 (Windows NT 10.0; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/55.0.2883.87 Safari/537.36");
curl_multi_add_handle($mh, $ch);
}
$running = null;
do {
curl_multi_exec($mh, $running);
usleep(100);
} while ($running);
foreach ($chs as $k => $ch) {
$content = curl_multi_getcontent($ch);
//$code = curl_getinfo($ch, CURLINFO_HTTP_CODE);
//echo $content;
//file_put_contents(time().rand(1000,9999).'.html',$content);
$result[] = self::search_result($content, $info['choices'][$k]);
curl_multi_remove_handle($mh, $ch);
}
curl_multi_close($mh);
return $result;
}
/**
* 处理否定词
* @return array
*/
static function negativeWords()
{
$negative_words_arr2 = array_map(function ($val) {
return "<span style='color: red'>$val</span>";
}, self::$negative_words_arr);
return $negative_words_arr2;
}
/**
* 运行
*/
static function run()
{
$screencap = self::screencap();
//$screencap = 'F:\dati\question\screen.png';
$question_img = self::getImg($screencap);
$question = self::getOcrText($question_img);
self::methodTwo($question);
file_put_contents(self::WORK_DIR . '/question/options', json_encode($question, JSON_UNESCAPED_UNICODE));
}
}
