math_cms/scripts/generate_sample_placeholder_audit_pdf.php

<?php

/**
 * 从占位符审计明细（ndjson）中按类型轮询抽样 N 道题，生成「题目质检」PDF（与 POST /api/questions/pdf 同源逻辑）。
 *
 * 用法：
 *   php scripts/generate_sample_placeholder_audit_pdf.php [--count 30]
 *       [--detail storage/app/audit_placeholder/rendered_placeholder_audit_details_*.ndjson]
 *       [--connection mysql] [--table questions]
 *
 * 若不指定 --detail，则自动选用 storage/app/audit_placeholder 下最新的 rendered_placeholder_audit_details_*.ndjson。
 *
 * 抽样优先级（轮询）：
 *   internal_placeholder_token_leak → broken_left_right_split → span_then_dollar_before_han
 *   → blank_between_math_segments → unbalanced_dollar_after_render → math_ends_with_operator_before_blank
 */

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 App\Services\ExamPdfExportService;
use Illuminate\Support\Facades\DB;

$options = getopt('', [
    'count::',
    'detail::',
    'connection::',
    'table::',
]);

$count = isset($options['count']) ? max(1, min(100, (int) $options['count'])) : 30;
$connection = isset($options['connection']) ? trim((string) $options['connection']) : config('database.default');
$table = isset($options['table']) ? trim((string) $options['table']) : 'questions';

$detailPath = isset($options['detail']) ? trim((string) $options['detail']) : '';
if ($detailPath === '') {
    $auditDir = dirname(__DIR__).'/storage/app/audit_placeholder';
    $glob = glob($auditDir.'/rendered_placeholder_audit_details_*.ndjson') ?: [];
    if ($glob === []) {
        fwrite(STDERR, "No detail ndjson found under {$auditDir}. Run scripts/audit_rendered_placeholder_integrity.php first.\n");
        exit(1);
    }
    usort($glob, static fn(string $a, string $b): int => strcmp($b, $a));
    $detailPath = $glob[0];
}

if (! is_readable($detailPath)) {
    fwrite(STDERR, "Cannot read detail file: {$detailPath}\n");
    exit(1);
}

$priorityTypes = [
    'internal_placeholder_token_leak',
    'broken_left_right_split',
    'span_then_dollar_before_han',
    'blank_between_math_segments',
    'unbalanced_dollar_after_render',
    'math_ends_with_operator_before_blank',
];

$buckets = array_fill_keys($priorityTypes, []);

$fh = fopen($detailPath, 'rb');
if ($fh === false) {
    fwrite(STDERR, "Failed to open {$detailPath}\n");
    exit(1);
}
while (($line = fgets($fh)) !== false) {
    $line = trim($line);
    if ($line === '') {
        continue;
    }
    $row = json_decode($line, true);
    if (! is_array($row) || ! isset($row['issue'], $row['id'])) {
        continue;
    }
    $issue = (string) $row['issue'];
    $id = (int) $row['id'];
    if ($id <= 0 || ! isset($buckets[$issue])) {
        continue;
    }
    $buckets[$issue][] = $id;
}
fclose($fh);

foreach ($buckets as $issue => &$ids) {
    $seen = [];
    $unique = [];
    foreach ($ids as $id) {
        if (! isset($seen[$id])) {
            $seen[$id] = true;
            $unique[] = $id;
        }
    }
    $ids = $unique;
}
unset($ids);

$selectedOrder = [];
$selectedSet = [];

while (count($selectedOrder) < $count) {
    $progress = false;
    foreach ($priorityTypes as $type) {
        if (count($selectedOrder) >= $count) {
            break 2;
        }
        while ($buckets[$type] !== []) {
            $id = array_shift($buckets[$type]);
            if (! isset($selectedSet[$id])) {
                $selectedSet[$id] = true;
                $selectedOrder[] = $id;
                $progress = true;
                break;
            }
        }
    }
    if (! $progress) {
        break;
    }
}

if ($selectedOrder === []) {
    fwrite(STDERR, "No issue rows parsed from {$detailPath}; nothing to sample.\n");
    exit(1);
}

if (count($selectedOrder) < $count) {
    fwrite(STDERR, 'warning: only '.count($selectedOrder)." unique ids available (requested {$count}).\n");
}

$questions = DB::connection($connection)
    ->table($table)
    ->whereIn('id', $selectedOrder)
    ->get();

$questionMap = [];
foreach ($questions as $q) {
    $questionMap[(int) $q->id] = $q;
}

$missing = array_values(array_filter($selectedOrder, static fn(int $id): bool => ! isset($questionMap[$id])));
if ($missing !== []) {
    fwrite(STDERR, 'warning: ids not found in '.$table.' ('.$connection.'): '.implode(',', $missing)."\n");
}

$groupedQuestions = groupQuestionsByType($questionMap, $selectedOrder);
$paper = buildVirtualPaper('占位符抽样质检_'.count($selectedOrder).'题', 'placeholder_audit_sample', $groupedQuestions);

/** @var ExamPdfExportService $pdf */
$pdf = $app->make(ExamPdfExportService::class);
$result = $pdf->generateQuestionCheckPdf(
    $paper,
    $groupedQuestions,
    ['name' => '质检抽样', 'grade' => '________'],
    ['name' => '________']
);

$manifestPath = dirname(__DIR__).'/storage/app/audit_placeholder/sample_pdf_manifest_'.date('Ymd_His').'.json';
$manifest = [
    'generated_at' => date('c'),
    'detail_source' => $detailPath,
    'connection' => $connection,
    'table' => $table,
    'requested_count' => $count,
    'sampled_question_ids' => $selectedOrder,
    'pdf_result' => $result,
];

@mkdir(dirname($manifestPath), 0777, true);
file_put_contents($manifestPath, json_encode($manifest, JSON_UNESCAPED_UNICODE | JSON_PRETTY_PRINT));

echo json_encode([
    'detail_source' => $detailPath,
    'manifest_path' => $manifestPath,
    'sampled_ids' => $selectedOrder,
    'pdf' => $result,
], JSON_UNESCAPED_UNICODE | JSON_PRETTY_PRINT)."\n";

/**
 * @param  array<int, object>  $questionMap
 */
function groupQuestionsByType(array $questionMap, array $originalOrder): array
{
    $grouped = [
        'choice' => [],
        'fill' => [],
        'answer' => [],
    ];

    $questionNumber = 1;

    foreach ($originalOrder as $id) {
        if (! isset($questionMap[$id])) {
            continue;
        }

        $q = $questionMap[$id];
        $type = normalizeQuestionType($q->question_type ?? null);

        $questionObj = (object) [
            'id' => $q->id,
            'question_number' => $questionNumber++,
            'content' => $q->stem,
            'options' => is_string($q->options) ? json_decode($q->options, true) : ($q->options ?? []),
            'answer' => $q->answer,
            'solution' => $q->solution,
            'score' => getDefaultScore($type),
            'difficulty' => $q->difficulty,
            'kp_code' => $q->kp_code,
        ];

        $grouped[$type][] = $questionObj;
    }

    return $grouped;
}

function normalizeQuestionType(?string $type): string
{
    if (! $type) {
        return 'answer';
    }

    $type = strtolower(trim($type));

    $typeMap = [
        'choice' => 'choice',
        '选择题' => 'choice',
        'single_choice' => 'choice',
        'multiple_choice' => 'choice',
        'fill' => 'fill',
        '填空题' => 'fill',
        'blank' => 'fill',
        'answer' => 'answer',
        '解答题' => 'answer',
        'subjective' => 'answer',
        'calculation' => 'answer',
        'proof' => 'answer',
    ];

    return $typeMap[$type] ?? 'answer';
}

function getDefaultScore(string $type): int
{
    return match ($type) {
        'choice' => 5,
        'fill' => 5,
        'answer' => 10,
        default => 5,
    };
}

/**
 * @param  array<string, array<int, object>>  $groupedQuestions
 */
function buildVirtualPaper(string $paperName, string $studentId, array $groupedQuestions): object
{
    $totalScore = 0;
    $totalQuestions = 0;

    foreach ($groupedQuestions as $questions) {
        foreach ($questions as $q) {
            $totalScore += $q->score;
            $totalQuestions++;
        }
    }

    $paperId = $studentId.'_'.time().'_'.uniqid();

    return (object) [
        'paper_id' => $paperId,
        'paper_name' => $paperName,
        'total_score' => $totalScore,
        'total_questions' => $totalQuestions,
        'created_at' => now()->toDateTimeString(),
    ];
}