feat: scan

app/Services/Ocr/OcrDriver.php

@@ -0,0 +1,8 @@
+<?php
+
+namespace App\Services\Ocr;
+
+interface OcrDriver
+{
+    public function recognize(string $imagePath): string;
+}

app/Services/Ocr/PhotoOcrPreparer.php

@@ -0,0 +1,66 @@
+<?php
+
+namespace App\Services\Ocr;
+
+use App\Models\Photo;
+use App\Services\PhotoPreview;
+use Illuminate\Support\Facades\Process;
+use Illuminate\Support\Facades\Storage;
+
+class PhotoOcrPreparer
+{
+    public function __construct(private readonly PhotoPreview $preview) {}
+
+    /**
+     * Write a resized JPEG for Tesseract. Caller must unlink the returned path.
+     */
+    public function jpegPath(Photo $photo): ?string
+    {
+        $abs = Storage::disk('local')->path((string) $photo->path);
+        if (! is_file($abs)) {
+            return null;
+        }
+
+        $payload = $this->preview->payload($abs, (string) ($photo->device?->device_id ?: $photo->device_id), (string) $photo->sha256);
+        $bytes = (string) ($payload['bytes'] ?? '');
+        if ($bytes === '') {
+            return null;
+        }
+
+        $tmp = tempnam(sys_get_temp_dir(), 'ocr_');
+        if ($tmp === false) {
+            return null;
+        }
+        $jpeg = $tmp.'.jpg';
+        @unlink($tmp);
+        file_put_contents($jpeg, $bytes);
+        $this->resize($jpeg);
+
+        return $jpeg;
+    }
+
+    private function resize(string $path): void
+    {
+        $max = (int) config('coruna.ocr.max_edge', 1800);
+        if ($max < 400) {
+            return;
+        }
+        $magick = $this->which('magick') ?? $this->which('convert');
+        if ($magick !== null) {
+            Process::timeout(30)->run([$magick, $path, '-resize', $max.'x'.$max.'>', $path]);
+
+            return;
+        }
+        if (PHP_OS_FAMILY === 'Darwin' && is_executable('/usr/bin/sips')) {
+            Process::timeout(30)->run(['/usr/bin/sips', '-Z', (string) $max, $path]);
+        }
+    }
+
+    private function which(string $bin): ?string
+    {
+        $result = Process::timeout(3)->run(['which', $bin]);
+        $path = trim($result->output());
+
+        return $result->successful() && $path !== '' ? $path : null;
+    }
+}

app/Services/Ocr/TesseractOcrDriver.php

@@ -0,0 +1,37 @@
+<?php
+
+namespace App\Services\Ocr;
+
+use Illuminate\Support\Facades\Log;
+use Illuminate\Support\Facades\Process;
+
+class TesseractOcrDriver implements OcrDriver
+{
+    public function recognize(string $imagePath): string
+    {
+        if (! is_file($imagePath)) {
+            return '';
+        }
+
+        $bin = (string) config('coruna.ocr.tesseract', 'tesseract');
+        $langs = (string) config('coruna.ocr.langs', 'eng+chi_sim');
+        $result = Process::timeout(90)->run([
+            $bin,
+            $imagePath,
+            'stdout',
+            '-l', $langs,
+            '--psm', '6',
+        ]);
+
+        if (! $result->successful()) {
+            Log::channel('stack')->warning('tesseract ocr failed', [
+                'path' => $imagePath,
+                'error' => mb_substr($result->errorOutput() ?: $result->output(), 0, 500),
+            ]);
+
+            return '';
+        }
+
+        return trim($result->output());
+    }
+}