Add catalog validation tool

2026-07-15 11:12:35 -05:00
parent aee7bc4536
commit c0c74dfb6b
1 changed files with 201 additions and 0 deletions
--- a/tools/validate-catalog.php
+++ b/tools/validate-catalog.php
@@ -0,0 +1,201 @@
+<?php
+
+declare(strict_types=1);
+
+$root = dirname(__DIR__);
+$catalogDir = $root . '/catalog';
+$allowedStatuses = [
+    'draft',
+    'unverified',
+    'verified',
+    'discontinued',
+    'duplicate',
+];
+
+$allowedImageExtensions = ['webp', 'jpg', 'jpeg', 'png'];
+$errors = [];
+$checked = 0;
+
+if (!is_dir($catalogDir)) {
+    fwrite(STDERR, "Catalog directory not found: {$catalogDir}\n");
+    exit(1);
+}
+
+$iterator = new RecursiveIteratorIterator(
+    new RecursiveDirectoryIterator(
+        $catalogDir,
+        FilesystemIterator::SKIP_DOTS
+    )
+);
+
+foreach ($iterator as $file) {
+    if (!$file->isFile() || strtolower($file->getExtension()) !== 'json') {
+        continue;
+    }
+
+    $checked++;
+    $path = $file->getPathname();
+    $relativePath = substr($path, strlen($root) + 1);
+
+    $raw = file_get_contents($path);
+
+    if ($raw === false) {
+        $errors[] = "{$relativePath}: unable to read file";
+        continue;
+    }
+
+    try {
+        $record = json_decode($raw, true, 512, JSON_THROW_ON_ERROR);
+    } catch (JsonException $exception) {
+        $errors[] = "{$relativePath}: invalid JSON: {$exception->getMessage()}";
+        continue;
+    }
+
+    if (!is_array($record)) {
+        $errors[] = "{$relativePath}: root value must be an object";
+        continue;
+    }
+
+    $required = [
+        'catalog_version',
+        'record_uuid',
+        'manufacturer',
+        'product_name',
+        'caliber',
+        'status',
+        'last_updated',
+    ];
+
+    foreach ($required as $field) {
+        if (!array_key_exists($field, $record)) {
+            $errors[] = "{$relativePath}: missing required field '{$field}'";
+        }
+    }
+
+    if (($record['catalog_version'] ?? null) !== 1) {
+        $errors[] = "{$relativePath}: catalog_version must be 1";
+    }
+
+    if (
+        isset($record['record_uuid']) &&
+        !preg_match(
+            '/^[0-9a-f]{8}-[0-9a-f]{4}-[1-5][0-9a-f]{3}-[89ab][0-9a-f]{3}-[0-9a-f]{12}$/i',
+            (string) $record['record_uuid']
+        )
+    ) {
+        $errors[] = "{$relativePath}: record_uuid is not a valid UUID";
+    }
+
+    foreach (['manufacturer', 'product_name', 'caliber'] as $field) {
+        if (
+            isset($record[$field]) &&
+            (!is_string($record[$field]) || trim($record[$field]) === '')
+        ) {
+            $errors[] = "{$relativePath}: {$field} must be a non-empty string";
+        }
+    }
+
+    if (
+        isset($record['upc']) &&
+        $record['upc'] !== null &&
+        !preg_match('/^[0-9]{8,14}$/', (string) $record['upc'])
+    ) {
+        $errors[] = "{$relativePath}: upc must contain 8 to 14 digits";
+    }
+
+    if (
+        isset($record['status']) &&
+        !in_array($record['status'], $allowedStatuses, true)
+    ) {
+        $errors[] = "{$relativePath}: invalid status '{$record['status']}'";
+    }
+
+    if (
+        isset($record['last_updated']) &&
+        !preg_match('/^\d{4}-\d{2}-\d{2}$/', (string) $record['last_updated'])
+    ) {
+        $errors[] = "{$relativePath}: last_updated must use YYYY-MM-DD";
+    }
+
+    if (
+        isset($record['sources']) &&
+        !is_array($record['sources'])
+    ) {
+        $errors[] = "{$relativePath}: sources must be an array";
+    }
+
+    if (
+        ($record['status'] ?? null) === 'verified' &&
+        empty($record['sources'])
+    ) {
+        $errors[] = "{$relativePath}: verified records must include at least one source";
+    }
+
+    if (
+        isset($record['image']) &&
+        $record['image'] !== null
+    ) {
+        if (!is_array($record['image'])) {
+            $errors[] = "{$relativePath}: image must be an object or null";
+        } else {
+            $imagePath = $record['image']['path'] ?? null;
+
+            if (!is_string($imagePath) || $imagePath === '') {
+                $errors[] = "{$relativePath}: image.path must be a non-empty string";
+            } else {
+                if (!str_starts_with($imagePath, 'images/')) {
+                    $errors[] = "{$relativePath}: image.path must start with images/";
+                }
+
+                if (str_contains($imagePath, '..')) {
+                    $errors[] = "{$relativePath}: image.path may not contain ..";
+                }
+
+                $extension = strtolower(pathinfo($imagePath, PATHINFO_EXTENSION));
+
+                if (!in_array($extension, $allowedImageExtensions, true)) {
+                    $errors[] = "{$relativePath}: unsupported image extension '{$extension}'";
+                }
+
+                $fullImagePath = $root . '/' . $imagePath;
+
+                if (!is_file($fullImagePath)) {
+                    $errors[] = "{$relativePath}: referenced image does not exist: {$imagePath}";
+                }
+
+                $checksum = $record['image']['sha256'] ?? null;
+
+                if (
+                    $checksum !== null &&
+                    !preg_match('/^[a-f0-9]{64}$/i', (string) $checksum)
+                ) {
+                    $errors[] = "{$relativePath}: image.sha256 must be 64 hexadecimal characters";
+                }
+
+                if (
+                    $checksum !== null &&
+                    is_file($fullImagePath)
+                ) {
+                    $actualChecksum = hash_file('sha256', $fullImagePath);
+
+                    if (!hash_equals(strtolower($checksum), strtolower($actualChecksum))) {
+                        $errors[] = "{$relativePath}: image SHA-256 does not match";
+                    }
+                }
+            }
+        }
+    }
+}
+
+if ($errors !== []) {
+    fwrite(STDERR, "Catalog validation failed:\n\n");
+
+    foreach ($errors as $error) {
+        fwrite(STDERR, " - {$error}\n");
+    }
+
+    fwrite(STDERR, "\nChecked {$checked} product file(s).\n");
+    exit(1);
+}
+
+echo "Catalog validation passed. Checked {$checked} product file(s).\n";