KnowledgeBase/app/Services/DocumentPdfPreviewService.php

<?php

namespace App\Services;

use App\Models\Document;
use Illuminate\Support\Facades\Storage;
use Symfony\Component\Process\Exception\ProcessFailedException;
use Symfony\Component\Process\Process;

class DocumentPdfPreviewService
{
    public function canPreview(Document $document): bool
    {
        return $document->conversion_status === 'completed'
            && ! empty($document->file_path)
            && Storage::disk('local')->exists($document->file_path);
    }

    public function getPreviewPath(Document $document): string
    {
        if (! $this->canPreview($document)) {
            throw new \RuntimeException('文档尚未完成转换或原文件不存在');
        }

        if ($this->isPdf($document)) {
            return Storage::disk('local')->path($document->file_path);
        }

        $previewPath = $this->cachedPreviewPath($document);

        if (! $this->cachedPreviewIsFresh($document, $previewPath)) {
            $this->generatePdfPreview($document, $previewPath);
        }

        return Storage::disk('previews')->path($previewPath);
    }

    public function previewUrl(Document $document): string
    {
        return route('documents.preview-pdf', $document);
    }

    public function clearCachedPreview(Document $document): void
    {
        Storage::disk('previews')->deleteDirectory((string) $document->getKey());
    }

    protected function isPdf(Document $document): bool
    {
        $extension = strtolower(pathinfo($document->display_file_name ?: $document->file_path, PATHINFO_EXTENSION));

        return $document->mime_type === 'application/pdf' || $extension === 'pdf';
    }

    protected function cachedPreviewPath(Document $document): string
    {
        return $document->getKey() . '/preview-libreoffice.pdf';
    }

    protected function cachedPreviewIsFresh(Document $document, string $previewPath): bool
    {
        if (! Storage::disk('previews')->exists($previewPath)) {
            return false;
        }

        $sourceMtime = Storage::disk('local')->lastModified($document->file_path);
        $previewMtime = Storage::disk('previews')->lastModified($previewPath);

        return $previewMtime >= $sourceMtime;
    }

    protected function generatePdfPreview(Document $document, string $previewPath): void
    {
        $sourcePath = Storage::disk('local')->path($document->file_path);
        $absolutePreviewPath = Storage::disk('previews')->path($previewPath);
        $previewDirectory = dirname($absolutePreviewPath);

        if (! is_dir($previewDirectory) && ! mkdir($previewDirectory, 0775, true) && ! is_dir($previewDirectory)) {
            throw new \RuntimeException('无法创建 PDF 预览目录');
        }

        $this->convertWithLibreOffice($sourcePath, $absolutePreviewPath, $previewDirectory);

        if (! file_exists($absolutePreviewPath) || filesize($absolutePreviewPath) === 0) {
            throw new \RuntimeException('PDF 预览生成失败');
        }
    }

    protected function convertWithLibreOffice(string $sourcePath, string $targetPath, string $workingDirectory): void
    {
        $binary = $this->resolveLibreOfficeBinary();

        if ($binary === null) {
            throw new \RuntimeException('无法生成准确的 PDF 预览：服务器未安装 LibreOffice/soffice。请安装 LibreOffice 和中文字体后重试。');
        }

        $tmpDirectory = $workingDirectory . '/tmp-' . bin2hex(random_bytes(8));
        $profileDirectory = $tmpDirectory . '/profile';

        if (! mkdir($profileDirectory, 0775, true) && ! is_dir($profileDirectory)) {
            throw new \RuntimeException('无法创建 LibreOffice 临时目录');
        }

        $process = new Process([
            $binary,
            '--headless',
            '--nologo',
            '--nofirststartwizard',
            '--norestore',
            '-env:UserInstallation=file://' . $profileDirectory,
            '--convert-to',
            'pdf',
            '--outdir',
            $tmpDirectory,
            $sourcePath,
        ]);
        $process->setTimeout((int) config('documents.conversion.timeout', 300));

        try {
            $process->mustRun();
            $convertedPath = $this->findConvertedPdf($tmpDirectory, $sourcePath);

            if ($convertedPath === null) {
                throw new \RuntimeException(trim($process->getOutput() . "\n" . $process->getErrorOutput()) ?: 'LibreOffice 未输出 PDF 文件');
            }

            if (! rename($convertedPath, $targetPath)) {
                throw new \RuntimeException('无法保存 LibreOffice 生成的 PDF 预览');
            }
        } catch (ProcessFailedException $e) {
            throw new \RuntimeException('LibreOffice 转换 PDF 失败：' . trim($e->getProcess()->getErrorOutput() ?: $e->getProcess()->getOutput()), 0, $e);
        } finally {
            $this->deleteDirectory($tmpDirectory);
        }
    }

    protected function resolveLibreOfficeBinary(): ?string
    {
        $configured = env('LIBREOFFICE_BINARY');
        $candidates = array_filter([
            is_string($configured) && $configured !== '' ? $configured : null,
            '/opt/homebrew/bin/soffice',
            '/opt/homebrew/bin/libreoffice',
            '/usr/bin/libreoffice',
            '/usr/bin/soffice',
            '/usr/local/bin/libreoffice',
            '/usr/local/bin/soffice',
            '/Applications/LibreOffice.app/Contents/MacOS/soffice',
        ]);

        foreach ($candidates as $candidate) {
            if (is_executable($candidate)) {
                return $candidate;
            }
        }

        return null;
    }

    protected function findConvertedPdf(string $directory, string $sourcePath): ?string
    {
        $expectedPath = $directory . '/' . pathinfo($sourcePath, PATHINFO_FILENAME) . '.pdf';

        if (is_file($expectedPath)) {
            return $expectedPath;
        }

        $pdfFiles = glob($directory . '/*.pdf') ?: [];

        return $pdfFiles[0] ?? null;
    }

    protected function deleteDirectory(string $directory): void
    {
        if (! is_dir($directory)) {
            return;
        }

        $items = new \RecursiveIteratorIterator(
            new \RecursiveDirectoryIterator($directory, \FilesystemIterator::SKIP_DOTS),
            \RecursiveIteratorIterator::CHILD_FIRST
        );

        foreach ($items as $item) {
            $item->isDir() ? rmdir($item->getPathname()) : unlink($item->getPathname());
        }

        rmdir($directory);
    }
}