jason/ThunderAI
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Add htmlBodyToPlainText function to sanitize HTML content. see #470

Browse source
This commit is contained in:
mic 2025-08-19 16:24:17 +02:00
parent 8878612f20
commit 8223f4ef4f
1 changed files with 13 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

13
js/mzta-utils.js
View file

@ -199,6 +199,19 @@ export function stripHtmlKeepLines(htmlString) {
.trim(); // removes leading/trailing whitespace
}
export function htmlBodyToPlainText(html) {
return html
.replace(/<head[\s\S]*?<\/head>/gi, '') // remove <head> and its content
.replace(/<br\s*\/?>/gi, '\n') // replace <br> with newline
.replace(/<\/p\s*>/gi, '\n') // replace </p> with newline
.replace(/<p\s*>/gi, '') // remove <p> tag
.replace(/<[^>]*>/g, ' ') // remove any other HTML tags
.replace(/[ \t]+\n/g, '\n') // remove spaces before newline
.replace(/\n{2,}/g, '\n') // compress multiple newlines into one
.replace(/\s+/g, ' ') // normalize spaces inside lines
.trim(); // trim spaces at start and end
}
export function convertNewlinesToBr(text) {
return text.replace(/\n/g, '<br>');
}