This commit is contained in:
杨志
2026-01-20 18:30:18 +08:00
parent 4f0c2112e7
commit c36f73caa2
2 changed files with 196 additions and 215 deletions

View File

@@ -39,29 +39,24 @@ class Crawler extends BaseController
]);
}
if (empty($cookiesParam)) {
$service = new CrawlerService();
try {
$cookies = $service->parseCookies($cookiesParam);
} catch (\Exception $e) {
return json([
'code' => 0,
'msg' => '请填写Cookie数据',
'msg' => $e->getMessage(),
]);
}
// 解析JSON格式的cookies
$cookies = is_string($cookiesParam) ? json_decode($cookiesParam, true) : $cookiesParam;
if (json_last_error() !== JSON_ERROR_NONE || empty($cookies)) {
return json([
'code' => 0,
'msg' => 'Cookie数据格式错误请检查JSON格式',
]);
}
$cookieString = $this->buildCookieString($cookies);
$cookieString = $service->buildCookieString($cookies);
// 构建URL获取HTML - 使用GET请求
$url = "http://gzrsks.oumakspt.com:62/tyzpwb/stuchooseexam/selectPosition.htm?examstupid=1015&userid={$userid}&bmid={$bmid}&examid={$examid}&aa={$aa}";
$baseUrl = $service->getBaseUrl();
$url = "{$baseUrl}/stuchooseexam/selectPosition.htm?examstupid=1015&userid={$userid}&bmid={$bmid}&examid={$examid}&aa={$aa}";
// 构建Referer URL与浏览器实际访问一致
$refererUrl = "http://gzrsks.oumakspt.com:62/tyzpwb/stuchooseexam/input.htm";
$refererUrl = "{$baseUrl}/stuchooseexam/input.htm";
$ch = curl_init();
curl_setopt($ch, CURLOPT_URL, $url);
@@ -72,17 +67,7 @@ class Crawler extends BaseController
curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, false);
curl_setopt($ch, CURLOPT_SSL_VERIFYHOST, false);
curl_setopt($ch, CURLOPT_ENCODING, ''); // 自动处理gzip编码
curl_setopt($ch, CURLOPT_HTTPHEADER, [
'Accept: text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,image/apng,*/*;q=0.8,application/signed-exchange;v=b3;q=0.7',
'Accept-Language: zh-CN,zh;q=0.9',
'Accept-Encoding: gzip, deflate',
'Cache-Control: no-cache',
'Connection: keep-alive',
'Cookie: ' . $cookieString,
'User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/143.0.0.0 Safari/537.36',
'Referer: ' . $refererUrl,
'Upgrade-Insecure-Requests: 1',
]);
curl_setopt($ch, CURLOPT_HTTPHEADER, $service->buildHtmlHeaders($cookieString, $refererUrl));
$html = curl_exec($ch);
$httpCode = curl_getinfo($ch, CURLINFO_HTTP_CODE);
@@ -103,7 +88,6 @@ class Crawler extends BaseController
]);
}
$service = new CrawlerService();
$options = $service->extractDsdmOptions($html);
return json([
@@ -120,53 +104,6 @@ class Crawler extends BaseController
}
}
/**
* 构建Cookie字符串
* @param array $cookies Cookie数组
* @return string
*/
private function buildCookieString(array|string $cookies): string
{
// 如果直接传入原始Cookie字符串优先使用
if (is_string($cookies)) {
return trim($cookies);
}
// 只保留用户填写的Cookie支持同名键多值如双JSESSIONID
$cookieData = $cookies['请求 Cookie'] ?? $cookies;
$normalized = [];
foreach ($cookieData as $key => $value) {
$k = trim((string)$key);
if ($k === '') {
continue;
}
$vals = is_array($value) ? $value : [$value];
foreach ($vals as $v) {
$v = trim((string)$v);
if ($v === '') {
continue;
}
$normalized[$k][] = $v;
}
}
// 将 JSESSIONID 放在最前(如果存在),其余按键名顺序,多值全部保留
$parts = [];
if (isset($normalized['JSESSIONID'])) {
foreach ($normalized['JSESSIONID'] as $v) {
$parts[] = 'JSESSIONID=' . $v;
}
unset($normalized['JSESSIONID']);
}
foreach ($normalized as $k => $vArr) {
foreach ($vArr as $v) {
$parts[] = $k . '=' . $v;
}
}
return implode('; ', $parts);
}
/**
* 获取职位代码列表
@@ -202,23 +139,16 @@ class Crawler extends BaseController
]);
}
if (empty($cookiesParam)) {
return json([
'code' => 0,
'msg' => '请填写Cookie数据',
]);
}
// 解析JSON格式的cookies
$cookies = is_string($cookiesParam) ? json_decode($cookiesParam, true) : $cookiesParam;
if (json_last_error() !== JSON_ERROR_NONE || empty($cookies)) {
return json([
'code' => 0,
'msg' => 'Cookie数据格式错误请检查JSON格式',
]);
}
$service = new CrawlerService();
try {
$cookies = $service->parseCookies($cookiesParam);
} catch (\Exception $e) {
return json([
'code' => 0,
'msg' => $e->getMessage(),
]);
}
$treeData = $service->getPositionTree($dsdm, $examid, $bmid, $userid, (string)$aa, $cookies);
// 提取所有CODE作为zwdm
@@ -268,22 +198,16 @@ class Crawler extends BaseController
]);
}
if (empty($cookiesParam)) {
return json([
'code' => 0,
'msg' => '请填写Cookie数据',
]);
}
$cookies = is_string($cookiesParam) ? json_decode($cookiesParam, true) : $cookiesParam;
if (json_last_error() !== JSON_ERROR_NONE || empty($cookies)) {
return json([
'code' => 0,
'msg' => 'Cookie数据格式错误请检查JSON格式',
]);
}
$service = new CrawlerService();
try {
$cookies = $service->parseCookies($cookiesParam);
} catch (\Exception $e) {
return json([
'code' => 0,
'msg' => $e->getMessage(),
]);
}
$treeData = $service->getPositionTree($dsdm, $examid, $bmid, $userid, (string)$aa, $cookies);
$zwdmList = [];
@@ -339,48 +263,20 @@ class Crawler extends BaseController
]);
}
if (empty($cookiesParam)) {
return json([
'code' => 0,
'msg' => '请填写Cookie数据',
]);
}
// 解析JSON格式的cookies
$cookies = is_string($cookiesParam) ? json_decode($cookiesParam, true) : $cookiesParam;
if (json_last_error() !== JSON_ERROR_NONE || empty($cookies)) {
return json([
'code' => 0,
'msg' => 'Cookie数据格式错误请检查JSON格式',
]);
}
$service = new CrawlerService();
try {
$cookies = $service->parseCookies($cookiesParam);
} catch (\Exception $e) {
return json([
'code' => 0,
'msg' => $e->getMessage(),
]);
}
$info = $service->getPositionInfo($zwdm, $examid, $cookies);
if (!empty($info)) {
// 处理单条数据或数组数据
if (isset($info[0])) {
$item = $info[0];
} else {
$item = $info;
}
// 计算竞争比(格式:招聘人数:审核通过人数)
$zprs = isset($item['zprs']) ? intval($item['zprs']) : 0;
$bkrs = isset($item['bkrs']) ? intval($item['bkrs']) : 0;
$competitionRatio = $zprs > 0 && $bkrs > 0 ? $zprs . ':' . $bkrs : ($zprs > 0 ? $zprs . ':0' : '0:0');
$result = [
'sbmc' => $item['sbmc'] ?? '', // 省份
'dsmc' => $item['dsmc'] ?? '', // 地区
'zpdwmc' => $item['zpdwmc'] ?? '', // 招聘单位/用人司局
'zwmc' => $item['zwmc'] ?? '', // 职位名称
'zwdm' => $item['zwdm'] ?? $zwdm, // 职位代码
'zprs' => $zprs, // 招聘人数
'bkrs' => $bkrs, // 审核通过人数
'competition_ratio' => $competitionRatio, // 竞争比(格式:招聘人数:审核通过人数)
];
$result = $service->formatPositionInfo($info, $zwdm);
return json([
'code' => 1,
@@ -428,23 +324,16 @@ class Crawler extends BaseController
]);
}
if (empty($cookiesParam)) {
return json([
'code' => 0,
'msg' => '请填写Cookie数据',
]);
}
// 解析JSON格式的cookies
$cookies = is_string($cookiesParam) ? json_decode($cookiesParam, true) : $cookiesParam;
if (json_last_error() !== JSON_ERROR_NONE || empty($cookies)) {
return json([
'code' => 0,
'msg' => 'Cookie数据格式错误请检查JSON格式',
]);
}
$service = new CrawlerService();
try {
$cookies = $service->parseCookies($cookiesParam);
} catch (\Exception $e) {
return json([
'code' => 0,
'msg' => $e->getMessage(),
]);
}
$results = $service->batchGetPositionInfo($zwdmList, $examid, $cookies);
return json([