PR#57: Add a static file for magazine - fedora-web/websites

fedora-web / websites

#57 Add a static file for magazine

Merged 4 years ago by pfrields. Opened 4 years ago by misc.

fedora-web/ misc/websites add_static into master

Pregenerate a json file instead of doing API calls for image

Michael Scherer • 4 years ago

ea42648

Add a static file for magazine

Michael Scherer • 4 years ago

06aa2ab

.gitignore

file modified

		`@@ -6,3 +6,4 @@`
		`*.po`
		`*.pot`
		`translations`
		`+ sites/getfedora.org/static/magazine.json`

build-prod.sh

file modified

		`@@ -38,6 +38,11 @@`
		`then`
		`./scripts/pull-translations.sh`
		`fi`
		`+ if [ -x ./scripts/pull-static.sh ];`
		`+ then`
		`+ ./scripts/pull-static.sh`
		`+ fi`
		`+`
		`${PYBINARY} main.py`
		`# This intermediate step is to make sure the final mv is atomic.`
		`# This means that syncs can happen at any point in time and they have a larger chance to be fine.`

sites/getfedora.org/main.py

file modified

		`@@ -223,6 +223,11 @@`
		`def releases_json():`
		`return send_from_directory('static', 'releases.json')`

		`+`
		`+ @app.route('/magazine.json')`
		`+ def magazine_json():`
		`+ return send_from_directory('static', 'magazine.json')`
		`+`
		`@app.route('/static/fedora.gpg')`
		`def gpgkey():`
		`return send_from_directory('static', 'fedora.gpg')`

sites/getfedora.org/scripts/pull-magazine.py

file added

+23

		`@@ -0,0 +1,23 @@`
		`+ #!/usr/bin/env python`
		`+ import json`
		`+ import requests`
		`+ # do not remove the line, it avoid triggering the IDS from wpengine`
		`+ headers = {'user-agent': 'getfedora-builder/0.0.1'}`
		`+ params = {'per_page': '3'}`
		`+`
		`+ r = requests.get('https://fedoramagazine.org/wp-json/wp/v2/posts', params=params, headers=headers)`
		`+ f = open('static/magazine.json', 'w')`
		`+ posts = []`
		`+ for i in r.json()[0:3]:`
		`+ p = {}`
		`+ p['link'] = i['link']`
		`+ p['title'] = i['title']['rendered']`
		`+ p['date'] = i['date']`
		`+`
		`+ #['wp:featuredmedia'][0].href:`
		`+ image_url = i['_links']['wp:featuredmedia'][0]['href']`
		`+ r2 = requests.get(image_url,headers=headers)`
		`+ p['image_url'] = r2.json()['media_details']['sizes']['medium_large']['source_url']`
		`+ posts.append(p)`
		`+`
		`+ f.write(json.dumps(posts))`

sites/getfedora.org/scripts/pull-static.sh

file added

		`@@ -0,0 +1,2 @@`
		`+ #!/bin/bash`
		`+ python $(dirname $0)/pull-magazine.py`

sites/static/js/magazine.js

file modified

+5 -6

		`@@ -3,17 +3,16 @@`
		`if(!$('#magazineposts').length)`
		`return;`

		`- $.get( "https://fedoramagazine.org/wp-json/wp/v2/posts", { per_page: 3 } )`
		`+ $.get( "/magazine.json")`
		`.done(function( data ) {`
		`$.each(data, function(i, item){`
		`const postlink = item.link;`
		`- const posttitle = item.title.rendered;`
		`+ const posttitle = item.title;`
		`const date = new Date(item.date);`
		`const month = date.toLocaleString('en-us', { month: 'long' });`
		`const postdate = ( month + ' ' + date.getDate()+', ' +date.getFullYear());`
		`- $.get(item._links['wp:featuredmedia'][0].href).done(function(imagedata){`
		`- $("#magazineposts").append("<div class='card'><a href='"+postlink+"'><img src='"+imagedata.media_details.sizes.medium_large.source_url+"' class='card-img-top'></a><div class='card-body font-weight-bold '><a class='text-dark' href='"+postlink+"'>"+posttitle+"</a><div><small>"+postdate+"</small></div></div></div>")`
		`- });`
		`- })`
		`+ const image_url = item.image_url;`
		`+ $("#magazineposts").append("<div class='card'><a href='"+postlink+"'><img src='"+image_url+"' class='card-img-top'></a><div class='card-body font-weight-bold '><a class='text-dark' href='"+postlink+"'>"+posttitle+"</a><div><small>"+postdate+"</small></div></div></div>");`
		`+ });`
		`});`
		`});`

misc commented 4 years ago

Rather than doing a request every time for Fedora magazine,
we are just going to get a snapshot from the website. Since the
website is rebuilt every hour, this shouldn't cause problem.

Fix https://pagure.io/fedora-web/websites/issue/56

misc commented 4 years ago

I didn't test yet, but going to work now the code is out there.

rebased onto 06aa2ab

4 years ago

misc commented 4 years ago

Ok, so I tested, it seems to work. It only solve the issue for the main API call, so this should remove ~180 000 API calls. I need to do something for the 2nd one, likely the same trick.

1 new commit added

Pregenerate a json file instead of doing API calls for image

4 years ago

misc commented 4 years ago

So, now, the website is using 1 json call with all information. I kept the javascript for the widget, due to date conversion and because it was easier than redo everything, but that's likely something to see too.

I think the documentation should also be fixed, but after that PR is merged