Génération PDF asynchrone : jobs avec polling + progression

- Nouveau système de jobs : POST /jobs (async), GET /jobs/{id} (status), GET /jobs/{id}/result (PDF), DELETE /jobs/{id}
- worker.php spawné en arrière-plan via nohup, lit la sortie pagedjs-cli ligne par ligne via proc_open et écrit la progression dans tmp/job_{id}.status.json
- Migration de pagedjs-cli en install local (node_modules) pour persister le patch protocolTimeout via patch-package
- CI : déploie package.json, worker.php, patches/, et lance npm install (qui réapplique les patches via postinstall)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

src/Controllers/JobsController.php

@@ -0,0 +1,120 @@
+<?php
+
+namespace Web2Print\Controllers;
+
+use Web2Print\Services\JobManager;
+
+class JobsController
+{
+    private JobManager $jobs;
+    private array $config;
+
+    public function __construct(JobManager $jobs, array $config)
+    {
+        $this->jobs = $jobs;
+        $this->config = $config;
+    }
+
+    public function create(): void
+    {
+        if ($_SERVER['REQUEST_METHOD'] !== 'POST') {
+            $this->sendError(405, 'Method not allowed');
+            return;
+        }
+
+        $input = file_get_contents('php://input');
+        if (strlen($input) > $this->config['max_html_size']) {
+            $this->sendError(413, 'Request too large');
+            return;
+        }
+
+        $data = json_decode($input, true);
+        if (json_last_error() !== JSON_ERROR_NONE) {
+            $this->sendError(400, 'Invalid JSON');
+            return;
+        }
+
+        if (empty($data['url'])) {
+            $this->sendError(400, 'url required (only URL mode is supported for async jobs)');
+            return;
+        }
+
+        $jobId = $this->jobs->create([
+            'url'      => $data['url'],
+            'filename' => !empty($data['filename']) ? basename($data['filename']) : 'document.pdf',
+            'options'  => $data['options'] ?? [],
+        ]);
+
+        http_response_code(202);
+        header('Content-Type: application/json');
+        echo json_encode(['job_id' => $jobId]);
+    }
+
+    public function status(string $jobId): void
+    {
+        if ($_SERVER['REQUEST_METHOD'] !== 'GET') {
+            $this->sendError(405, 'Method not allowed');
+            return;
+        }
+
+        $status = $this->jobs->getStatus($jobId);
+        if ($status === null) {
+            $this->sendError(404, 'Job not found');
+            return;
+        }
+
+        header('Content-Type: application/json');
+        echo json_encode($status);
+    }
+
+    public function result(string $jobId): void
+    {
+        if ($_SERVER['REQUEST_METHOD'] !== 'GET') {
+            $this->sendError(405, 'Method not allowed');
+            return;
+        }
+
+        $status = $this->jobs->getStatus($jobId);
+        if ($status === null) {
+            $this->sendError(404, 'Job not found');
+            return;
+        }
+
+        if (($status['status'] ?? null) !== 'done') {
+            $this->sendError(409, 'Job not ready (status: ' . ($status['status'] ?? 'unknown') . ')');
+            return;
+        }
+
+        $path = $this->jobs->getResultPath($jobId);
+        if ($path === null) {
+            $this->sendError(500, 'Result file missing');
+            return;
+        }
+
+        $request = $this->jobs->getRequest($jobId);
+        $filename = $request['filename'] ?? 'document.pdf';
+
+        header('Content-Type: application/pdf');
+        header('Content-Disposition: attachment; filename="' . $filename . '"');
+        header('Content-Length: ' . filesize($path));
+        readfile($path);
+    }
+
+    public function delete(string $jobId): void
+    {
+        if ($_SERVER['REQUEST_METHOD'] !== 'DELETE') {
+            $this->sendError(405, 'Method not allowed');
+            return;
+        }
+
+        $this->jobs->delete($jobId);
+        http_response_code(204);
+    }
+
+    private function sendError(int $code, string $message): void
+    {
+        http_response_code($code);
+        header('Content-Type: application/json');
+        echo json_encode(['error' => $message]);
+    }
+}

src/Services/JobManager.php

@@ -0,0 +1,104 @@
+<?php
+
+namespace Web2Print\Services;
+
+class JobManager
+{
+    private array $config;
+
+    public function __construct(array $config)
+    {
+        $this->config = $config;
+    }
+
+    public function create(array $request): string
+    {
+        $jobId = bin2hex(random_bytes(16));
+
+        $requestPath = $this->path($jobId, 'request.json');
+        $statusPath  = $this->path($jobId, 'status.json');
+
+        file_put_contents($requestPath, json_encode($request, JSON_UNESCAPED_SLASHES));
+        $this->writeStatus($jobId, ['status' => 'pending', 'createdAt' => time()]);
+
+        $cmd = sprintf(
+            'nohup %s %s %s > /dev/null 2>&1 &',
+            escapeshellcmd($this->config['php_bin']),
+            escapeshellarg(__DIR__ . '/../../worker.php'),
+            escapeshellarg($jobId)
+        );
+        exec($cmd);
+
+        return $jobId;
+    }
+
+    public function getStatus(string $jobId): ?array
+    {
+        $path = $this->path($jobId, 'status.json');
+        if (!file_exists($path)) {
+            return null;
+        }
+        $raw = @file_get_contents($path);
+        $data = json_decode($raw, true);
+        return is_array($data) ? $data : null;
+    }
+
+    public function writeStatus(string $jobId, array $status): void
+    {
+        $status['updatedAt'] = time();
+        file_put_contents(
+            $this->path($jobId, 'status.json'),
+            json_encode($status, JSON_UNESCAPED_SLASHES)
+        );
+    }
+
+    public function getRequest(string $jobId): ?array
+    {
+        $path = $this->path($jobId, 'request.json');
+        if (!file_exists($path)) {
+            return null;
+        }
+        $data = json_decode(@file_get_contents($path), true);
+        return is_array($data) ? $data : null;
+    }
+
+    public function getResultPath(string $jobId): ?string
+    {
+        $path = $this->path($jobId, 'pdf');
+        return file_exists($path) ? $path : null;
+    }
+
+    public function pdfPath(string $jobId): string
+    {
+        return $this->path($jobId, 'pdf');
+    }
+
+    public function delete(string $jobId): void
+    {
+        foreach (glob($this->config['tmp_dir'] . '/job_' . $jobId . '.*') ?: [] as $file) {
+            @unlink($file);
+        }
+    }
+
+    public function cleanOrphans(): int
+    {
+        $maxAge = $this->config['job_max_age'] ?? 3600;
+        $cutoff = time() - $maxAge;
+        $count = 0;
+
+        foreach (glob($this->config['tmp_dir'] . '/job_*.status.json') ?: [] as $statusFile) {
+            if (filemtime($statusFile) < $cutoff) {
+                if (preg_match('/job_([a-f0-9]+)\.status\.json$/', $statusFile, $m)) {
+                    $this->delete($m[1]);
+                    $count++;
+                }
+            }
+        }
+        return $count;
+    }
+
+    private function path(string $jobId, string $suffix): string
+    {
+        return $this->config['tmp_dir'] . '/job_' . $jobId . '.' . $suffix;
+    }
+}

src/Services/PdfGenerator.php

@@ -44,6 +44,7 @@ class PdfGenerator
             // Nettoyer les fichiers temporaires
             @unlink($htmlFile);
             @unlink($pdfFile);
+            $this->cleanChromiumTmp();
         }
     }
 
@@ -80,6 +81,98 @@ class PdfGenerator
 
         } finally {
             @unlink($pdfFile);
+            $this->cleanChromiumTmp();
+        }
+    }
+
+    public function generateFromUrlToFile(string $url, string $pdfFile, array $options = [], ?callable $onProgress = null): void
+    {
+        $cmd = 'TMPDIR=' . escapeshellarg($this->config['tmp_dir']) . ' ';
+        $cmd .= escapeshellcmd($this->config['pagedjs_bin']);
+        $cmd .= ' ' . escapeshellarg($url);
+        $cmd .= ' -o ' . escapeshellarg($pdfFile);
+
+        if (!empty($options['timeout'])) {
+            $cmd .= ' --timeout ' . (int)$options['timeout'];
+        } else {
+            $cmd .= ' --timeout ' . ($this->config['pagedjs_timeout'] * 1000);
+        }
+
+        // proc_open pour streamer stdout/stderr ligne par ligne
+        $descriptors = [
+            0 => ['pipe', 'r'],
+            1 => ['pipe', 'w'],
+            2 => ['pipe', 'w'],
+        ];
+        $proc = proc_open($cmd, $descriptors, $pipes);
+        if (!is_resource($proc)) {
+            throw new \Exception('Failed to spawn pagedjs-cli');
+        }
+        fclose($pipes[0]);
+
+        stream_set_blocking($pipes[1], false);
+        stream_set_blocking($pipes[2], false);
+
+        $stderrBuffer = '';
+        $partial = ['', ''];
+
+        try {
+            while (true) {
+                $status = proc_get_status($proc);
+                $read = [$pipes[1], $pipes[2]];
+                $w = null; $e = null;
+                $changed = @stream_select($read, $w, $e, 1);
+
+                if ($changed > 0) {
+                    foreach ($read as $stream) {
+                        $idx = ($stream === $pipes[1]) ? 0 : 1;
+                        $chunk = fread($stream, 4096);
+                        if ($chunk === '' || $chunk === false) continue;
+                        if ($idx === 1) $stderrBuffer .= $chunk;
+                        $partial[$idx] .= $chunk;
+                        // Découpe par ligne (\n ou \r — Ora utilise \r pour les spinners)
+                        while (preg_match('/^([^\r\n]*)[\r\n](.*)$/s', $partial[$idx], $m)) {
+                            $line = $m[1];
+                            $partial[$idx] = $m[2];
+                            if ($line !== '' && $onProgress !== null) {
+                                $onProgress($line);
+                            }
+                        }
+                    }
+                }
+
+                if (!$status['running']) {
+                    // Vider les buffers restants
+                    foreach ([$pipes[1], $pipes[2]] as $i => $stream) {
+                        $idx = ($stream === $pipes[1]) ? 0 : 1;
+                        while (($chunk = fread($stream, 4096)) !== false && $chunk !== '') {
+                            if ($idx === 1) $stderrBuffer .= $chunk;
+                            $partial[$idx] .= $chunk;
+                        }
+                        if ($partial[$idx] !== '' && $onProgress !== null) {
+                            foreach (preg_split('/[\r\n]+/', $partial[$idx]) as $line) {
+                                if ($line !== '') $onProgress($line);
+                            }
+                        }
+                    }
+                    break;
+                }
+            }
+
+            fclose($pipes[1]);
+            fclose($pipes[2]);
+            $exitCode = proc_close($proc);
+
+            if ($exitCode !== 0) {
+                $this->log('Paged.js CLI error (URL streaming): ' . $stderrBuffer);
+                throw new \Exception('PDF generation failed: ' . trim($stderrBuffer));
+            }
+
+            if (!file_exists($pdfFile)) {
+                throw new \Exception('PDF file not created');
+            }
+        } finally {
+            $this->cleanChromiumTmp();
         }
     }
 
@@ -130,6 +223,30 @@ class PdfGenerator
         return $cmd;
     }
 
+    private function cleanChromiumTmp(): void
+    {
+        $tmpDir = $this->config['tmp_dir'];
+        foreach (glob($tmpDir . '/org.chromium.Chromium.*') ?: [] as $dir) {
+            $this->rmdirRecursive($dir);
+        }
+        foreach (glob($tmpDir . '/puppeteer_dev_chrome_profile-*') ?: [] as $dir) {
+            $this->rmdirRecursive($dir);
+        }
+    }
+
+    private function rmdirRecursive(string $path): void
+    {
+        if (!is_dir($path)) {
+            @unlink($path);
+            return;
+        }
+        foreach (scandir($path) as $item) {
+            if ($item === '.' || $item === '..') continue;
+            $this->rmdirRecursive($path . '/' . $item);
+        }
+        @rmdir($path);
+    }
+
     private function log(string $message): void
     {
         $timestamp = date('Y-m-d H:i:s');