math_cms/scripts/audit_question_stem_quality.php

<?php

declare(strict_types=1);

require __DIR__.'/../vendor/autoload.php';
$app = require __DIR__.'/../bootstrap/app.php';
$kernel = $app->make(Illuminate\Contracts\Console\Kernel::class);
$kernel->bootstrap();

use Illuminate\Support\Facades\DB;

$limit = isset($argv[1]) ? max(1, (int) $argv[1]) : 50000;
$outDir = isset($argv[2]) ? rtrim($argv[2], '/'): '/tmp';
$table = isset($argv[3]) ? trim((string) $argv[3]) : 'questions';
@mkdir($outDir, 0777, true);

$rows = DB::connection('remote_mysql')
    ->table($table)
    ->select('id', 'question_type', 'stem')
    ->whereNotNull('stem')
    ->orderByDesc('id')
    ->limit($limit)
    ->get();

$issues = [
    'unbalanced_dollar' => [],
    'suspicious_latex_env' => [],
    'mixed_placeholder_inside_math' => [],
    'compare_blank_between_math_tokens' => [],
    'tail_backslash_dollar_marker' => [],
];

$push = static function(array &$bucket, object $row, string $reason): void {
    if (count($bucket) >= 2000) {
        return;
    }
    $bucket[] = [
        'id' => (int) $row->id,
        'question_type' => (string) $row->question_type,
        'reason' => $reason,
        'stem_preview' => mb_substr((string) $row->stem, 0, 220),
    ];
};

foreach ($rows as $row) {
    $stem = (string) $row->stem;

    $dollarCount = substr_count($stem, '$');
    if (($dollarCount % 2) !== 0) {
        $push($issues['unbalanced_dollar'], $row, 'odd number of $ delimiters');
    }

    if (preg_match('/\\\\begin\{[^}]*$/u', $stem) || preg_match('/\\\\end\{[^}]*$/u', $stem)) {
        $push($issues['suspicious_latex_env'], $row, 'truncated \\begin/\\end block');
    }

    if (preg_match('/\$(?:[^$]|\\\\.)*(?:\\\\underline\{[^}]*\}|_{2,}|[（(](?:\s|&nbsp;|&#160;|　)*[）)])(?:[^$]|\\\\.)*\$/u', $stem)) {
        $push($issues['mixed_placeholder_inside_math'], $row, 'placeholder token appears inside $...$');
    }

    if (preg_match('/\$[^$]*\$\s*_{2,}\s*\$[^$]*\$/u', $stem) || preg_match('/\$[^$]*\$\s*[（(](?:\s|&nbsp;|&#160;|　)*[）)]\s*\$[^$]*\$/u', $stem)) {
        $push($issues['compare_blank_between_math_tokens'], $row, 'blank token inserted between two math segments');
    }

    if (preg_match('/\\\\+\$(?=\s*(?:<[^>]+>\s*)*$)/u', $stem)) {
        $push($issues['tail_backslash_dollar_marker'], $row, 'tail backslash-dollar marker used as blank placeholder');
    }
}

$summary = [
    'table' => $table,
    'scan_limit' => $limit,
    'scanned_rows' => count($rows),
    'counts' => array_map('count', $issues),
    'generated_at' => date('c'),
];

$stamp = date('Ymd_His');
$summaryPath = "$outDir/question_stem_quality_summary_$stamp.json";
$detailPath = "$outDir/question_stem_quality_details_$stamp.json";
file_put_contents($summaryPath, json_encode($summary, JSON_UNESCAPED_UNICODE|JSON_PRETTY_PRINT));
file_put_contents($detailPath, json_encode($issues, JSON_UNESCAPED_UNICODE|JSON_PRETTY_PRINT));

echo json_encode([
    'summary_path' => $summaryPath,
    'detail_path' => $detailPath,
    'summary' => $summary,
], JSON_UNESCAPED_UNICODE|JSON_PRETTY_PRINT), "\n";