3 namespace BookStack\Entities\Tools;
5 use BookStack\Entities\Models\Page;
6 use BookStack\Entities\Tools\Markdown\MarkdownToHtml;
7 use BookStack\Exceptions\ImageUploadException;
8 use BookStack\Uploads\ImageRepo;
9 use BookStack\Uploads\ImageService;
10 use BookStack\Util\HtmlContentFilter;
16 use Illuminate\Support\Str;
23 * PageContent constructor.
25 public function __construct(Page $page)
31 * Update the content of the page with new provided HTML.
33 public function setNewHTML(string $html)
35 $html = $this->extractBase64ImagesFromHtml($html);
36 $this->page->html = $this->formatHtml($html);
37 $this->page->text = $this->toPlainText();
38 $this->page->markdown = '';
42 * Update the content of the page with new provided Markdown content.
44 public function setNewMarkdown(string $markdown)
46 $markdown = $this->extractBase64ImagesFromMarkdown($markdown);
47 $this->page->markdown = $markdown;
48 $html = (new MarkdownToHtml($markdown))->convert();
49 $this->page->html = $this->formatHtml($html);
50 $this->page->text = $this->toPlainText();
54 * Convert all base64 image data to saved images.
56 protected function extractBase64ImagesFromHtml(string $htmlText): string
58 if (empty($htmlText) || strpos($htmlText, 'data:image') === false) {
62 $doc = $this->loadDocumentFromHtml($htmlText);
63 $container = $doc->documentElement;
64 $body = $container->childNodes->item(0);
65 $childNodes = $body->childNodes;
66 $xPath = new DOMXPath($doc);
68 // Get all img elements with image data blobs
69 $imageNodes = $xPath->query('//img[contains(@src, \'data:image\')]');
70 foreach ($imageNodes as $imageNode) {
71 $imageSrc = $imageNode->getAttribute('src');
72 $newUrl = $this->base64ImageUriToUploadedImageUrl($imageSrc);
73 $imageNode->setAttribute('src', $newUrl);
76 // Generate inner html as a string
78 foreach ($childNodes as $childNode) {
79 $html .= $doc->saveHTML($childNode);
86 * Convert all inline base64 content to uploaded image files.
87 * Regex is used to locate the start of data-uri definitions then
88 * manual looping over content is done to parse the whole data uri.
89 * Attempting to capture the whole data uri using regex can cause PHP
90 * PCRE limits to be hit with larger, multi-MB, files.
92 protected function extractBase64ImagesFromMarkdown(string $markdown)
95 $contentLength = strlen($markdown);
97 preg_match_all('/!\[.*?]\(.*?(data:image\/.{1,6};base64,)/', $markdown, $matches, PREG_OFFSET_CAPTURE);
99 foreach ($matches[1] as $base64MatchPair) {
100 [$dataUri, $index] = $base64MatchPair;
102 for ($i = strlen($dataUri) + $index; $i < $contentLength; $i++) {
103 $char = $markdown[$i];
104 if ($char === ')' || $char === ' ' || $char === "\n" || $char === '"') {
110 $newUrl = $this->base64ImageUriToUploadedImageUrl($dataUri);
111 $replacements[] = [$dataUri, $newUrl];
114 foreach ($replacements as [$dataUri, $newUrl]) {
115 $markdown = str_replace($dataUri, $newUrl, $markdown);
122 * Parse the given base64 image URI and return the URL to the created image instance.
123 * Returns an empty string if the parsed URI is invalid or causes an error upon upload.
125 protected function base64ImageUriToUploadedImageUrl(string $uri): string
127 $imageRepo = app()->make(ImageRepo::class);
128 $imageInfo = $this->parseBase64ImageUri($uri);
130 // Validate extension and content
131 if (empty($imageInfo['data']) || !ImageService::isExtensionSupported($imageInfo['extension'])) {
135 // Validate that the content is not over our upload limit
136 $uploadLimitBytes = (config('app.upload_limit') * 1000000);
137 if (strlen($imageInfo['data']) > $uploadLimitBytes) {
141 // Save image from data with a random name
142 $imageName = 'embedded-image-' . Str::random(8) . '.' . $imageInfo['extension'];
145 $image = $imageRepo->saveNewFromData($imageName, $imageInfo['data'], 'gallery', $this->page->id);
146 } catch (ImageUploadException $exception) {
154 * Parse a base64 image URI into the data and extension.
156 * @return array{extension: string, data: string}
158 protected function parseBase64ImageUri(string $uri): array
160 [$dataDefinition, $base64ImageData] = explode(',', $uri, 2);
161 $extension = strtolower(preg_split('/[\/;]/', $dataDefinition)[1] ?? '');
164 'extension' => $extension,
165 'data' => base64_decode($base64ImageData) ?: '',
170 * Formats a page's html to be tagged correctly within the system.
172 protected function formatHtml(string $htmlText): string
174 if (empty($htmlText)) {
178 $doc = $this->loadDocumentFromHtml($htmlText);
179 $container = $doc->documentElement;
180 $body = $container->childNodes->item(0);
181 $childNodes = $body->childNodes;
182 $xPath = new DOMXPath($doc);
184 // Set ids on top-level nodes
186 foreach ($childNodes as $index => $childNode) {
187 [$oldId, $newId] = $this->setUniqueId($childNode, $idMap);
188 if ($newId && $newId !== $oldId) {
189 $this->updateLinks($xPath, '#' . $oldId, '#' . $newId);
193 // Set ids on nested header nodes
194 $nestedHeaders = $xPath->query('//body//*//h1|//body//*//h2|//body//*//h3|//body//*//h4|//body//*//h5|//body//*//h6');
195 foreach ($nestedHeaders as $nestedHeader) {
196 [$oldId, $newId] = $this->setUniqueId($nestedHeader, $idMap);
197 if ($newId && $newId !== $oldId) {
198 $this->updateLinks($xPath, '#' . $oldId, '#' . $newId);
202 // Ensure no duplicate ids within child items
203 $idElems = $xPath->query('//body//*//*[@id]');
204 foreach ($idElems as $domElem) {
205 [$oldId, $newId] = $this->setUniqueId($domElem, $idMap);
206 if ($newId && $newId !== $oldId) {
207 $this->updateLinks($xPath, '#' . $oldId, '#' . $newId);
211 // Generate inner html as a string
213 foreach ($childNodes as $childNode) {
214 $html .= $doc->saveHTML($childNode);
217 // Perform required string-level tweaks
218 $html = str_replace(' ', ' ', $html);
224 * Update the all links to the $old location to instead point to $new.
226 protected function updateLinks(DOMXPath $xpath, string $old, string $new)
228 $old = str_replace('"', '', $old);
229 $matchingLinks = $xpath->query('//body//*//*[@href="' . $old . '"]');
230 foreach ($matchingLinks as $domElem) {
231 $domElem->setAttribute('href', $new);
236 * Set a unique id on the given DOMElement.
237 * A map for existing ID's should be passed in to check for current existence.
238 * Returns a pair of strings in the format [old_id, new_id].
240 protected function setUniqueId(DOMNode $element, array &$idMap): array
242 if (!$element instanceof DOMElement) {
246 // Stop if there's an existing valid id that has not already been used.
247 $existingId = $element->getAttribute('id');
248 if (strpos($existingId, 'bkmrk') === 0 && !isset($idMap[$existingId])) {
249 $idMap[$existingId] = true;
251 return [$existingId, $existingId];
254 // Create a unique id for the element
255 // Uses the content as a basis to ensure output is the same every time
256 // the same content is passed through.
257 $contentId = 'bkmrk-' . mb_substr(strtolower(preg_replace('/\s+/', '-', trim($element->nodeValue))), 0, 20);
258 $newId = urlencode($contentId);
261 while (isset($idMap[$newId])) {
262 $newId = urlencode($contentId . '-' . $loopIndex);
266 $element->setAttribute('id', $newId);
267 $idMap[$newId] = true;
269 return [$existingId, $newId];
273 * Get a plain-text visualisation of this page.
275 protected function toPlainText(): string
277 $html = $this->render(true);
279 return html_entity_decode(strip_tags($html));
283 * Render the page for viewing.
285 public function render(bool $blankIncludes = false): string
287 $content = $this->page->html ?? '';
289 if (!config('app.allow_content_scripts')) {
290 $content = HtmlContentFilter::removeScripts($content);
293 if ($blankIncludes) {
294 $content = $this->blankPageIncludes($content);
296 $content = $this->parsePageIncludes($content);
303 * Parse the headers on the page to get a navigation menu.
305 public function getNavigation(string $htmlContent): array
307 if (empty($htmlContent)) {
311 $doc = $this->loadDocumentFromHtml($htmlContent);
312 $xPath = new DOMXPath($doc);
313 $headers = $xPath->query('//h1|//h2|//h3|//h4|//h5|//h6');
315 return $headers ? $this->headerNodesToLevelList($headers) : [];
319 * Convert a DOMNodeList into an array of readable header attributes
320 * with levels normalised to the lower header level.
322 protected function headerNodesToLevelList(DOMNodeList $nodeList): array
324 $tree = collect($nodeList)->map(function (DOMElement $header) {
325 $text = trim(str_replace("\xc2\xa0", '', $header->nodeValue));
326 $text = mb_substr($text, 0, 100);
329 'nodeName' => strtolower($header->nodeName),
330 'level' => intval(str_replace('h', '', $header->nodeName)),
331 'link' => '#' . $header->getAttribute('id'),
334 })->filter(function ($header) {
335 return mb_strlen($header['text']) > 0;
338 // Shift headers if only smaller headers have been used
339 $levelChange = ($tree->pluck('level')->min() - 1);
340 $tree = $tree->map(function ($header) use ($levelChange) {
341 $header['level'] -= ($levelChange);
346 return $tree->toArray();
350 * Remove any page include tags within the given HTML.
352 protected function blankPageIncludes(string $html): string
354 return preg_replace("/{{@\s?([0-9].*?)}}/", '', $html);
358 * Parse any include tags "{{@<page_id>#section}}" to be part of the page.
360 protected function parsePageIncludes(string $html): string
363 preg_match_all("/{{@\s?([0-9].*?)}}/", $html, $matches);
365 foreach ($matches[1] as $index => $includeId) {
366 $fullMatch = $matches[0][$index];
367 $splitInclude = explode('#', $includeId, 2);
369 // Get page id from reference
370 $pageId = intval($splitInclude[0]);
371 if (is_nan($pageId)) {
375 // Find page and skip this if page not found
376 /** @var ?Page $matchedPage */
377 $matchedPage = Page::visible()->find($pageId);
378 if ($matchedPage === null) {
379 $html = str_replace($fullMatch, '', $html);
383 // If we only have page id, just insert all page html and continue.
384 if (count($splitInclude) === 1) {
385 $html = str_replace($fullMatch, $matchedPage->html, $html);
389 // Create and load HTML into a document
390 $innerContent = $this->fetchSectionOfPage($matchedPage, $splitInclude[1]);
391 $html = str_replace($fullMatch, trim($innerContent), $html);
398 * Fetch the content from a specific section of the given page.
400 protected function fetchSectionOfPage(Page $page, string $sectionId): string
402 $topLevelTags = ['table', 'ul', 'ol', 'pre'];
403 $doc = $this->loadDocumentFromHtml($page->html);
405 // Search included content for the id given and blank out if not exists.
406 $matchingElem = $doc->getElementById($sectionId);
407 if ($matchingElem === null) {
411 // Otherwise replace the content with the found content
412 // Checks if the top-level wrapper should be included by matching on tag types
414 $isTopLevel = in_array(strtolower($matchingElem->nodeName), $topLevelTags);
416 $innerContent .= $doc->saveHTML($matchingElem);
418 foreach ($matchingElem->childNodes as $childNode) {
419 $innerContent .= $doc->saveHTML($childNode);
422 libxml_clear_errors();
424 return $innerContent;
428 * Create and load a DOMDocument from the given html content.
430 protected function loadDocumentFromHtml(string $html): DOMDocument
432 libxml_use_internal_errors(true);
433 $doc = new DOMDocument();
434 $html = '<body>' . $html . '</body>';
435 $doc->loadHTML(mb_convert_encoding($html, 'HTML-ENTITIES', 'UTF-8'));