Python for Entrepreneurs Transcripts
Chapter: Web design foundations
Lecture: Concept: Cache busting

Login or purchase this course to watch this video and the rest of the course contents.
0:01 Let's quickly review this concept of cache busting.
0:03 Recall we'd started out with the white site with black text
0:07 and it looked like this, more or less and then we made a change,
0:12 we wanted to say "switch it", so we had light text and a dark background,
0:16 without the cache busting in place well, remember that basically had no effect,
0:20 at least under certain circumstances, not the least of which is in production.
0:25 So we added this cache id query string parameter to our CSS files,
0:31 so "?cacheId=whatever", right, you can put anything after the query string question mark there,
0:37 and we wrote a function called build_cache_id
0:40 and it looks at the contents of the file and it creates an md5 hash
0:44 which has the really nice property of - it's pretty short and it changes
0:48 if and only if the contents of that file change.
0:51 So basically, every time we change a file, we get a new URL
0:54 that also gets aggressively cached but it doesn't matter to anybody
0:58 because it's always caching the new one.
1:00 So now with this in place, we no longer have that problem,
1:04 when we deploy a new version or just continue to use the site
1:07 it automatically adapts to the new CSS, regardless of what the cache settings are
1:12 and they should be extremely aggressive in production on that static folder.
1:16 Here is the code that we wrote in sort of simplistic,
1:19 "help you understand what is going on" version,
1:21 we used the dunder name to create the full path, and then we took a relative file URL,
1:27 this is basically a web relative path,
1:30 and we used out full path and a little bit of os path juggling
1:34 to come up with a local file name, verify that that worked
1:37 and then we just computed the hash with our md5 hash there and returned hash digest.
1:42 Now I said in production, we really should use a dictionary
1:45 and recompute this only if we'd never seen the file before for that run of the process,
1:51 so I've included in the demo code the file called,
1:56 which is a much more production-friendly version.
1:59 It's not identical to the one we'll have in the final product,
2:02 it's not identical to the one I am using on my websites, but it's really close.