slake-school-service/app/Services/Crawl/CrawlKeywordParser.php

<?php

namespace App\Services\Crawl;

class CrawlKeywordParser
{
    /**
     * 解析多个关键词：支持空格、中英文逗号、分号、换行分隔。
     *
     * @return list<string>
     */
    public static function parse(?string $raw): array
    {
        if ($raw === null || trim($raw) === '') {
            return [];
        }

        $parts = preg_split('/[\s,，;；\n\r]+/u', trim($raw), -1, PREG_SPLIT_NO_EMPTY) ?: [];

        $keywords = [];
        foreach ($parts as $part) {
            $kw = trim($part);
            if ($kw !== '' && ! in_array($kw, $keywords, true)) {
                $keywords[] = $kw;
            }
        }

        return $keywords;
    }

    /**
     * 标题或摘要是否命中任一关键词。
     *
     * @param  list<string>  $keywords
     */
    public static function matchesAny(string $title, ?string $summary, array $keywords): bool
    {
        if ($keywords === []) {
            return true;
        }

        $haystack = $title.($summary ?? '');

        foreach ($keywords as $keyword) {
            if ($keyword !== '' && mb_stripos($haystack, $keyword) !== false) {
                return true;
            }
        }

        return false;
    }

    /**
     * 按逗号/分号/换行拆成多个检索短语；无分隔符时整段视为一个短语（短语内空格在 arXiv 查询中 AND）。
     *
     * @return list<string>
     */
    public static function parsePhrases(?string $raw): array
    {
        $raw = trim((string) $raw);
        if ($raw === '') {
            return [];
        }

        if (preg_match('/[,，;；\n\r]/u', $raw)) {
            $parts = preg_split('/[,，;；\n\r]+/u', $raw, -1, PREG_SPLIT_NO_EMPTY) ?: [];
        } else {
            $parts = [$raw];
        }

        $phrases = [];
        foreach ($parts as $part) {
            $phrase = trim($part);
            if ($phrase !== '' && ! in_array($phrase, $phrases, true)) {
                $phrases[] = $phrase;
            }
        }

        return $phrases;
    }

    /**
     * 构建 arXiv search_query：短语内空格 AND，多短语之间 OR。
     */
    public static function buildArxivSearchQuery(?string $raw): string
    {
        $phrases = self::parsePhrases($raw);
        if ($phrases === []) {
            return 'all:*';
        }

        $clauseParts = [];
        foreach ($phrases as $phrase) {
            $terms = preg_split('/\s+/u', $phrase, -1, PREG_SPLIT_NO_EMPTY) ?: [];
            $ands = [];
            foreach ($terms as $term) {
                $clean = preg_replace('/[^\p{L}\p{N}\s\-_]/u', '', $term);
                if ($clean !== '') {
                    $ands[] = 'all:'.$clean;
                }
            }
            if ($ands === []) {
                continue;
            }
            $clauseParts[] = count($ands) === 1 ? $ands[0] : '('.implode(' AND ', $ands).')';
        }

        if ($clauseParts === []) {
            return 'all:*';
        }

        return count($clauseParts) === 1
            ? $clauseParts[0]
            : '('.implode(' OR ', $clauseParts).')';
    }
}