For an app I have been working on recently, I wanted a way to let the user know if the data they were being shown was likely to be a cached version of the data or a non cached version. Because the API I was using for the app doesn’t return a timestamp of when the API call was made, I needed a different way to “guess” when the cache expired. I thought of taking a hash of the data returned from the last call and the current API call and comparing them but that would only let me know when the API returns something different and not that it isn’t a cached version of the API call. Thus I came up with the following…
ruleset a60x507 {
meta {
name "ran-recently-example"
description <<
ran-recently-example
>>
author "Mike Grace"
logging on
}
rule check_for_recent_run {
select when pageview ".*"
pre {
msg =<<
Because we cache our API calls for
this app for 1 minute, you are most
likely seeing cached data from the
last time you ran this.
>>;
}
if ent:recentlyRan == true within 1 minutes then {
notify("Notice", msg) with sticky = true;
}
notfired {
raise explicit event fresh_run
}
}
rule fresh_run {
select when explicit fresh_run
pre {
msg =<<
Wohoo! Looks like the cached data
expired so you should be seeing
new data if there is any provided
by the API. : )
>>;
}
{
notify("Notice",msg) with sticky = true;
}
fired {
set ent:recentlyRan;
}
}
}
- 12 selects on pageview event
- 21 if the entity flag ‘recentlyRan’ is set to true within the last minute then execute the action block
- 24 postlude block raises explicit event of ‘fresh_run’ if the action block did not fire
- 30 rule selects on the ‘fresh_run’ explicit event
- 42 set the entity flag ‘recentlyRan’ to true
App run on example.com several times over a 2 minute period with bookmarklet
This pattern is just a best guess solution that would be best solved by the API providing a timestamp of when the API call was made.
Get the bookmarklet to try it out yourself!
Gratuitous day 16 Grace face
