sysr-q / chr Goto Github PK

These links should be removed from the database after being visited.
This would probably require adding an extra database field, so it will mean that there will have to be some sort of migration tool. This could just export all the urls to a file/string, modify it, then modify the database and import.

I'm not certain what in particular is the issue (the math may just be off), but it appears to be showing incorrect percentages.

On the chr.so instance for example, /2/stats displays a visitor return ratio of 175%! (With 7 unique visitors, and -3 repeat at the time of writing).

This is definitely something which needs a fix, so it will have to be looked into.

These are separated into sections, because apparently GHM doesn't allow n-layered check lists.

• Package refactor
• chru.api.web.v1? Really, past me?
• chru.web seems redundant, since, by default, this really is a web based project.
• chru.utility.funcs:real_ip can be replaced with a ProxyFix solution.

• General code cleanup
• Documentation - it's skimpy in some areas, and beefy in others. It should be given a thorough look over.
• The adding/modification of URLs is overly complicated - what can we do about this?
• The CLI based API manager can easily be moved over to use the standard library cmd.Cmd class.

• Misc things
• Move from use of mattdaemon to some higher-level server, like gunicorn, giving the host more control over how it's run.
• Use redis rather than sqlite3
• Fix strange statistics, as highlighted by #1
• Add burn-on-read links (#2). This will need some serious scheming to implement decently.
• There's no API in the current v3 refactor - I'm not sure if anyone used the old API, so I might just add a basic JSON endpoint for shortening URLs if recaptcha is disabled.