job修改

This commit is contained in:
chengxl
2025-08-22 16:24:59 +08:00
parent 131a7c8014
commit c2ed27c0df

View File

@@ -140,7 +140,6 @@ class Aireview extends Base
$sContent = $oHelperFunction->filterAllTags($sContent); $sContent = $oHelperFunction->filterAllTags($sContent);
//将文章内容拆分参考文献 //将文章内容拆分参考文献
$aDealContent = $this->dealContent($sContent); $aDealContent = $this->dealContent($sContent);
$sBefore= empty($aDealContent['before']) ? '' : $aDealContent['before']; $sBefore= empty($aDealContent['before']) ? '' : $aDealContent['before'];
$sReference = empty($aDealContent['after']) ? '' : $aDealContent['after']; $sReference = empty($aDealContent['after']) ? '' : $aDealContent['after'];
if(in_array($sQuestionFields, ['attribute'])){//科学性和创新性 if(in_array($sQuestionFields, ['attribute'])){//科学性和创新性
@@ -193,50 +192,72 @@ class Aireview extends Base
* @param sContent 文章内容 * @param sContent 文章内容
*/ */
private function dealContent($sContent = '',$regex = null){ private function dealContent($sContent = '',$regex = null){
if(empty($sContent)){ if (empty($sContent)) {
return ['before' => '', 'after' => '']; return ['before' => '', 'after' => ''];
} }
// 1. 限制匹配范围末尾30% // 1. 优化字符串长度计算
$contentLength = strlen($sContent); $contentLength = mb_strlen($sContent);
$searchStart = $contentLength > 5000 ? (int)($contentLength * 0.7) : 0; $searchStart = $contentLength > 5000 ? (int)($contentLength * 0.6) : 0;
$searchContent = substr($sContent, $searchStart);
// 2. 正则模式优化 // 2. 截取搜索区域
$searchContent = $searchStart > 0
? mb_substr($sContent, $searchStart, null, 'UTF-8')
: $sContent;
// 3. 正则模式优化 - 重点处理拼写错误和特殊格式
if ($regex === null) { if ($regex === null) {
// 关键词列表增加容错处理
$keywords = [ $keywords = [
'references?', 'bibliograph(?:y|ies)', 'r?reference[s]?', // 允许关键词前多一个r处理Rreferences这类拼写错误
'works? cited', 'citation(?:s)?' 'bibliograph(?:y|ies)',
'works? cited',
'citation[s]?',
'literature cited',
'reference list'
]; ];
// 正则模式优化:
// 1. 允许关键词与前面内容直接连接(解决"article.Rreferences"这种没有空格的情况)
// 2. 增强对冒号和方括号的支持(匹配"References:[1]"这种格式)
$pattern = sprintf( $pattern = sprintf(
'/(?:^|\s)\s*#*\s*(%s)\s*[:\-]?\s*(?:$|\s)/i', '/(?:^|\s|[(\[{<"\']|[\p{P}])?\s*\#*[*_]*\s*(%s)\s*[*_]*\s*[:\-=.]?\s*(?:$|\s|[)\]}>"\'.|[\d{[])*/i',
implode('|', $keywords) implode('|', $keywords)
); );
$regex = $pattern; $regex = $pattern;
} }
// 3. 匹配并处理结果 // 4. 高效匹配 - 从字符串末尾开始搜索最后一个匹配项
if (preg_match_all($regex, $searchContent, $matches, PREG_OFFSET_CAPTURE)) { // 使用带偏移量的循环匹配找到最后一个符合条件的关键词
$lastMatch = end($matches[0]); $lastPos = 0;
$refPosition = $searchStart + $lastMatch[1]; $matchFound = false;
$offset = 0;
$searchLen = mb_strlen($searchContent);
// 4. 合并字符串处理 // 从字符串末尾向前搜索,提高长文本效率
$before = substr($sContent, 0, $refPosition); while ($offset < $searchLen && preg_match($regex, $searchContent, $match, PREG_OFFSET_CAPTURE, $offset)) {
$after = substr($sContent, $refPosition); $lastPos = $match[0][1];
$matchFound = true;
// 移动偏移量继续查找下一个匹配
$offset = $lastPos + mb_strlen($match[0][0]);
}
if ($matchFound) {
$refPosition = $searchStart + $lastPos;
// 一次性处理空白和换行 $process = static function($str) {
$process = function($str) { return trim(str_replace("\n", '', $str));
return str_replace("\n", '', trim($str));
}; };
return [ return [
'before' => $process($before), 'before' => $process(mb_substr($sContent, 0, $refPosition, 'UTF-8')),
'after' => $process($after) 'after' => $process(mb_substr($sContent, $refPosition, null, 'UTF-8'))
]; ];
} }
// 未匹配时处理 // 未匹配时处理
return [ return [
'before' => str_replace("\n", '', trim($sContent)), 'before' => trim(str_replace("\n", '', $sContent)),
'after' => '' 'after' => ''
]; ];
// $lastPos = strrpos($sContent, 'Reference'); // $lastPos = strrpos($sContent, 'Reference');