evol/gif

A command-line image library manager for nerds.

gif's Issues

Non-Gifs?

Started playing with this, and it's well designed ("do one thing well"), thank you!

One problem: it doesn't need to assume that all input is Gif, so why do so? This would be more useful if I could also use it for png, jpg, mp4, mov... etcetera.

This could be a lot simpler than using libmagic or similar, just try to preserve the input file extension?

Store file types in the database

It would be good to store MIME types of images inside the database. At the same time, extensions of internal files would correspond to their types, so JPEG files would end in .jpg and so on.

Current behavior causes some inconvenience in certain workflows and I want to avoid it.

Refactor the codebase to support automated tests

Currently, the development of gif very much depends on me thinking and checking all corner cases. Since I'm working on it on and off in my spare time and the feature base keeps increasing, this becomes harder and harder to do. Automated tests would solve the issue for most cases. However, the codebase isn't really testable at the moment, with globals and cross-cutting concerns.

My plan is to revisit and rework the architecture and then introduce tests. That's something I think is essential if I want gif to evolve further.

Add fuzzy tag matching

The problem I have with current prefix matching is that when you have lots of images inside gif, you don't really remember exactly how you tagged your images. Fuzzy matching would help a lot.

According to some research I've done on the matter, SQLite's full text search isn't up to the task. However, we can use a separate model just for matching tags with something like https://github.com/sajari/fuzzy.

Add "-" keyword to match previously added entry

A common workflow I use is:

gif add IMAGE_URL
gif upload ID_OF_THAT_IMAGE

I propose to make this easier by having - to be a virtual ID of the last added image. The above workflow would then turn into:

gif add IMAGE_URL
gif upload -