math_cms/app/Services/PaperPartExtractorService.php

<?php

namespace App\Services;

use App\Models\PaperPart;
use App\Models\SourcePaper;
use Illuminate\Support\Collection;
use Illuminate\Support\Facades\DB;
use Illuminate\Support\Str;

class PaperPartExtractorService
{
    /**
     * 基于卷子 Markdown 拆分题型区块。
     */
    public function extract(SourcePaper $paper): Collection
    {
        $parts = $this->splitIntoParts($paper->raw_markdown);

        return DB::transaction(function () use ($paper, $parts) {
            $paper->parts()->delete();

            $result = collect();
            foreach ($parts as $idx => $part) {
                $result->push(PaperPart::create([
                    'source_paper_id' => $paper->id,
                    'order' => $idx + 1,
                    'title' => $part['title'] ?? null,
                    'type' => $part['type'] ?? null,
                    'raw_markdown' => $part['raw'],
                    'question_count' => $part['question_count'] ?? null,
                    'detected_features' => $part['detected_features'] ?? [],
                ]));
            }

            return $result;
        });
    }

    public function splitIntoParts(string $markdown): array
    {
        $lines = preg_split('/\r\n|\r|\n/', $markdown);
        $segments = [];
        $current = ['title' => null, 'buffer' => []];

        $partPattern = '/^(#{2,3})\s*(第? ?[一二三四五六七八九十0-9IVX]+[部分卷]|选择题|填空题|解答题|综合题|计算题|应用题)/u';
        $commentPattern = '/<!--\s*part:\s*(.+?)\s*-->/i';

        foreach ($lines as $line) {
            $trimmed = trim($line);

            // 支持隐藏的区块标记
            if (preg_match($commentPattern, $trimmed, $cm)) {
                if (!empty($current['buffer'])) {
                    $segments[] = $this->finalizeSegment($current);
                }
                $current = [
                    'title' => trim($cm[1]),
                    'buffer' => [$line],
                ];
                continue;
            }

            if (preg_match($partPattern, $line, $m)) {
                if (!empty($current['buffer'])) {
                    $segments[] = $this->finalizeSegment($current);
                }
                $current = [
                    'title' => trim($m[0], "# \t"),
                    'buffer' => [$line],
                ];
            } else {
                $current['buffer'][] = $line;
            }
        }

        if (!empty($current['buffer'])) {
            $segments[] = $this->finalizeSegment($current);
        }

        if (empty($segments)) {
            return [[
                'title' => null,
                'type' => 'mixed',
                'raw' => trim($markdown),
                'detected_features' => [],
            ]];
        }

        return $segments;
    }

    protected function finalizeSegment(array $segment): array
    {
        $raw = trim(implode("\n", $segment['buffer']));
        $title = $segment['title'];

        return [
            'title' => $title,
            'type' => $this->detectType($title),
            'raw' => $raw,
            'detected_features' => [
                'title' => $title,
            ],
        ];
    }

    protected function detectType(?string $title): ?string
    {
        if (!$title) {
            return null;
        }

        return match (true) {
            Str::contains($title, '选择') => 'choice',
            Str::contains($title, '填空') => 'fill',
            Str::contains($title, ['解答', '简答', '分析']) => 'answer',
            Str::contains($title, ['计算', '推导']) => 'calc',
            default => 'mixed',
        };
    }
}