diff --git a/updateFeeds.js b/updateFeeds.js index b68021b..28fd387 100644 --- a/updateFeeds.js +++ b/updateFeeds.js @@ -95,7 +95,7 @@ function transformContent(item) { } let $ = cheerio.load(item.content); - // look through all the links + // look through all the links to find images let links = $('a'); let urls = []; //console.log('links', links.length); @@ -107,9 +107,10 @@ function transformContent(item) { urls.push(url); } }); + + // look through all the images let images = $('img'); images.each((i,e) => { - console.log(i,e); let url = $(e).attr('src'); // if there's an image, add it as a media attachment if (url && url.match(/(http)?s?:?(\/\/[^"']*\.(?:png|jpg|jpeg|gif|png|svg))/)) { @@ -122,8 +123,17 @@ function transformContent(item) { item.urls = urls; + // find iframe embeds and turn them into links + let iframes = $('iframe'); + iframes.each((i,e) => { + console.log('iframe',i,e); + let url = $(e).attr('src'); + $(e).replaceWith($(`[embedded content]`)); + }); + + // remove multiple line breaks - $('br').remove(); + //$('br').remove(); $('p').each((i, el) => { if($(el).html().replace(/\s| /g, '').length === 0) {$(el).remove();} }); @@ -131,6 +141,8 @@ function transformContent(item) { // couple of hacky regexes to make sure we clean up everything item.content = $('body').html().replace(/^(\n|\r)/,'').replace(/>\r+\r+\s