-<?php namespace BookStack\Entities\Tools;
+<?php
+
+namespace BookStack\Entities\Tools;
use BookStack\Entities\Models\Page;
+use BookStack\Entities\Tools\Markdown\CustomListItemRenderer;
use BookStack\Entities\Tools\Markdown\CustomStrikeThroughExtension;
use BookStack\Exceptions\ImageUploadException;
use BookStack\Facades\Theme;
use BookStack\Theming\ThemeEvents;
-use BookStack\Util\HtmlContentFilter;
use BookStack\Uploads\ImageRepo;
+use BookStack\Uploads\ImageService;
+use BookStack\Util\HtmlContentFilter;
use DOMDocument;
use DOMNodeList;
use DOMXPath;
use Illuminate\Support\Str;
+use League\CommonMark\Block\Element\ListItem;
use League\CommonMark\CommonMarkConverter;
use League\CommonMark\Environment;
use League\CommonMark\Extension\Table\TableExtension;
class PageContent
{
-
protected $page;
/**
*/
public function setNewHTML(string $html)
{
- $html = $this->extractBase64Images($this->page, $html);
+ $html = $this->extractBase64ImagesFromHtml($html);
$this->page->html = $this->formatHtml($html);
$this->page->text = $this->toPlainText();
$this->page->markdown = '';
*/
public function setNewMarkdown(string $markdown)
{
+ $markdown = $this->extractBase64ImagesFromMarkdown($markdown);
$this->page->markdown = $markdown;
$html = $this->markdownToHtml($markdown);
$this->page->html = $this->formatHtml($html);
$environment->addExtension(new CustomStrikeThroughExtension());
$environment = Theme::dispatch(ThemeEvents::COMMONMARK_ENVIRONMENT_CONFIGURE, $environment) ?? $environment;
$converter = new CommonMarkConverter([], $environment);
+
+ $environment->addBlockRenderer(ListItem::class, new CustomListItemRenderer(), 10);
+
return $converter->convertToHtml($markdown);
}
/**
- * Convert all base64 image data to saved images
+ * Convert all base64 image data to saved images.
*/
- public function extractBase64Images(Page $page, string $htmlText): string
+ protected function extractBase64ImagesFromHtml(string $htmlText): string
{
if (empty($htmlText) || strpos($htmlText, 'data:image') === false) {
return $htmlText;
$body = $container->childNodes->item(0);
$childNodes = $body->childNodes;
$xPath = new DOMXPath($doc);
- $imageRepo = app()->make(ImageRepo::class);
- $allowedExtensions = ['jpg', 'jpeg', 'png', 'gif', 'webp'];
// Get all img elements with image data blobs
$imageNodes = $xPath->query('//img[contains(@src, \'data:image\')]');
foreach ($imageNodes as $imageNode) {
$imageSrc = $imageNode->getAttribute('src');
- [$dataDefinition, $base64ImageData] = explode(',', $imageSrc, 2);
- $extension = strtolower(preg_split('/[\/;]/', $dataDefinition)[1] ?? 'png');
-
- // Validate extension
- if (!in_array($extension, $allowedExtensions)) {
- $imageNode->setAttribute('src', '');
- continue;
- }
-
- // Save image from data with a random name
- $imageName = 'embedded-image-' . Str::random(8) . '.' . $extension;
- try {
- $image = $imageRepo->saveNewFromData($imageName, base64_decode($base64ImageData), 'gallery', $page->id);
- $imageNode->setAttribute('src', $image->url);
- } catch (ImageUploadException $exception) {
- $imageNode->setAttribute('src', '');
- }
+ $newUrl = $this->base64ImageUriToUploadedImageUrl($imageSrc);
+ $imageNode->setAttribute('src', $newUrl);
}
// Generate inner html as a string
return $html;
}
+ /**
+ * Convert all inline base64 content to uploaded image files.
+ */
+ protected function extractBase64ImagesFromMarkdown(string $markdown)
+ {
+ $matches = [];
+ preg_match_all('/!\[.*?]\(.*?(data:image\/.*?)[)"\s]/', $markdown, $matches);
+
+ foreach ($matches[1] as $base64Match) {
+ $newUrl = $this->base64ImageUriToUploadedImageUrl($base64Match);
+ $markdown = str_replace($base64Match, $newUrl, $markdown);
+ }
+
+ return $markdown;
+ }
+
+ /**
+ * Parse the given base64 image URI and return the URL to the created image instance.
+ * Returns an empty string if the parsed URI is invalid or causes an error upon upload.
+ */
+ protected function base64ImageUriToUploadedImageUrl(string $uri): string
+ {
+ $imageRepo = app()->make(ImageRepo::class);
+ $imageInfo = $this->parseBase64ImageUri($uri);
+
+ // Validate extension and content
+ if (empty($imageInfo['data']) || !ImageService::isExtensionSupported($imageInfo['extension'])) {
+ return '';
+ }
+
+ // Validate that the content is not over our upload limit
+ $uploadLimitBytes = (config('app.upload_limit') * 1000000);
+ if (strlen($imageInfo['data']) > $uploadLimitBytes) {
+ return '';
+ }
+
+ // Save image from data with a random name
+ $imageName = 'embedded-image-' . Str::random(8) . '.' . $imageInfo['extension'];
+
+ try {
+ $image = $imageRepo->saveNewFromData($imageName, $imageInfo['data'], 'gallery', $this->page->id);
+ } catch (ImageUploadException $exception) {
+ return '';
+ }
+
+ return $image->url;
+ }
+
+ /**
+ * Parse a base64 image URI into the data and extension.
+ *
+ * @return array{extension: array, data: string}
+ */
+ protected function parseBase64ImageUri(string $uri): array
+ {
+ [$dataDefinition, $base64ImageData] = explode(',', $uri, 2);
+ $extension = strtolower(preg_split('/[\/;]/', $dataDefinition)[1] ?? '');
+
+ return [
+ 'extension' => $extension,
+ 'data' => base64_decode($base64ImageData) ?: '',
+ ];
+ }
+
/**
* Formats a page's html to be tagged correctly within the system.
*/
}
}
+ // Set ids on nested header nodes
+ $nestedHeaders = $xPath->query('//body//*//h1|//body//*//h2|//body//*//h3|//body//*//h4|//body//*//h5|//body//*//h6');
+ foreach ($nestedHeaders as $nestedHeader) {
+ [$oldId, $newId] = $this->setUniqueId($nestedHeader, $idMap);
+ if ($newId && $newId !== $oldId) {
+ $this->updateLinks($xPath, '#' . $oldId, '#' . $newId);
+ }
+ }
+
// Ensure no duplicate ids within child items
$idElems = $xPath->query('//body//*//*[@id]');
foreach ($idElems as $domElem) {
/**
* Set a unique id on the given DOMElement.
* A map for existing ID's should be passed in to check for current existence.
- * Returns a pair of strings in the format [old_id, new_id]
+ * Returns a pair of strings in the format [old_id, new_id].
*/
protected function setUniqueId(\DOMNode $element, array &$idMap): array
{
$existingId = $element->getAttribute('id');
if (strpos($existingId, 'bkmrk') === 0 && !isset($idMap[$existingId])) {
$idMap[$existingId] = true;
+
return [$existingId, $existingId];
}
$element->setAttribute('id', $newId);
$idMap[$newId] = true;
+
return [$existingId, $newId];
}
protected function toPlainText(): string
{
$html = $this->render(true);
+
return html_entity_decode(strip_tags($html));
}
/**
- * Render the page for viewing
+ * Render the page for viewing.
*/
public function render(bool $blankIncludes = false): string
{
- $content = $this->page->html;
+ $content = $this->page->html ?? '';
if (!config('app.allow_content_scripts')) {
$content = HtmlContentFilter::removeScripts($content);
}
/**
- * Parse the headers on the page to get a navigation menu
+ * Parse the headers on the page to get a navigation menu.
*/
public function getNavigation(string $htmlContent): array
{
$doc = $this->loadDocumentFromHtml($htmlContent);
$xPath = new DOMXPath($doc);
- $headers = $xPath->query("//h1|//h2|//h3|//h4|//h5|//h6");
+ $headers = $xPath->query('//h1|//h2|//h3|//h4|//h5|//h6');
return $headers ? $this->headerNodesToLevelList($headers) : [];
}
return [
'nodeName' => strtolower($header->nodeName),
- 'level' => intval(str_replace('h', '', $header->nodeName)),
- 'link' => '#' . $header->getAttribute('id'),
- 'text' => $text,
+ 'level' => intval(str_replace('h', '', $header->nodeName)),
+ 'link' => '#' . $header->getAttribute('id'),
+ 'text' => $text,
];
})->filter(function ($header) {
return mb_strlen($header['text']) > 0;
$levelChange = ($tree->pluck('level')->min() - 1);
$tree = $tree->map(function ($header) use ($levelChange) {
$header['level'] -= ($levelChange);
+
return $header;
});
}
// Find page and skip this if page not found
+ /** @var ?Page $matchedPage */
$matchedPage = Page::visible()->find($pageId);
if ($matchedPage === null) {
$html = str_replace($fullMatch, '', $html);
return $html;
}
-
/**
* Fetch the content from a specific section of the given page.
*/
protected function fetchSectionOfPage(Page $page, string $sectionId): string
{
- $topLevelTags = ['table', 'ul', 'ol'];
- $doc = $this->loadDocumentFromHtml('<body>' . $page->html . '</body>');
+ $topLevelTags = ['table', 'ul', 'ol', 'pre'];
+ $doc = $this->loadDocumentFromHtml($page->html);
// Search included content for the id given and blank out if not exists.
$matchingElem = $doc->getElementById($sectionId);
{
libxml_use_internal_errors(true);
$doc = new DOMDocument();
+ $html = '<body>' . $html . '</body>';
$doc->loadHTML(mb_convert_encoding($html, 'HTML-ENTITIES', 'UTF-8'));
+
return $doc;
}
}
projects / bookstack / blobdiff