search

beancount / fava

Fava - web interface for Beancount
https://beancount.github.io/fava/
MIT License
1.97k stars 285 forks source link

Import page shows files that can't be imported #1224

Closed mattr555 closed 3 years ago

mattr555 commented 3 years ago

I'm someone who tends to hoard a lot of files in my ~/Downloads folder, which is where I download my monthly statements, and is the import directory for bean-file/fava. I'd prefer to only see files that have an associated importer on fava's import page (though I understand the value in having other files show up in the list if you're using a different directory to collect documents before they're filed).

I'm happy to contribute a PR that adds an option to enable this. Would this be a sensible default to enable, or should we show all files by default? Don't want to break anyone's workflow. Thanks!

edit: as an alternative, we could explore adding a toggle to do this filtering on the frontend. The only quibble I would have with that is that we get a performance win by not sending down files that the user doesn't care about.

yagebu commented 3 years ago

This is not really something that I'd want to add an option for but doing this filtering in the frontend sounds like a worthwhile solution. Maybe we could display the list in two parts to allow those that don't have an importer to be more easily ignored and just toggled in the frontend.

The only quibble I would have with that is that we get a performance win by not sending down files that the user doesn't care about.

The size of the data loaded on that page is really quite negligible.

mattr555 commented 3 years ago

Maybe we could display the list in two parts to allow those that don't have an importer to be more easily ignored and just toggled in the frontend.

I like this idea! Hope to have a PR out implementing it this weekend.

The size of the data loaded on that page is really quite negligible.

Fair enough. I was noticing loading times of a couple seconds before I purged some of my old downloads, but nothing extreme.

yagebu commented 3 years ago

I like this idea! Hope to have a PR out implementing it this weekend.

Cool :)

I was noticing loading times of a couple seconds before I purged some of my old downloads, but nothing extreme.

All files in the import directory are checked by the identify part of Beancount's ingest mechanism, depending on your importers that might load and read all files (e.g., to check for the existence of some identifier in a CSV file). That might take some time if you have a lot of files in that directory of course and can't really be avoided, transmitting the result will only be a fraction of those seconds