# embeddings-sort

This program can sort images such that ones with similar motives are close together. This is accomplished by using [AI](https://github.com/minimaxir/imgbeddings) to extract the meaning of the image, and then approximating a travelling-salesperson-tour through all of them.  
As a bonus feature, this program can also sort the images by hue, brightness or color, though the results for this could be improved by using a less generalized algorithm.

The sorting can be accessed by letting the progam print the image paths in order, or by copying/symlinking the images into a new directory.

Detailed usage:

```
embeddings-sort [OPTIONS] [IMAGES]...

Arguments:
  [IMAGES]...  

Options:
  -e, --embedder <EMBEDDER>        Characteristic to sort by [default: content] [possible values: brightness, hue, color, content]
  -s, --symlink-dir <SYMLINK_DIR>  Symlink the sorted images into this directory
  -o, --copy-dir <COPY_DIR>        Copy the sorted images into this directory. Uses COW when available
  -c, --stdout                     Write sorted paths into stdout, one per line
  -0, --stdout0                    Write sorted paths into stdout, null-separated. Overrides -c
  -h, --help                       Print help
```

## Insides
The TSP approximation is done by using Prim's algorithm and doing a DFS through the resulting MST, giving a 2-approximation, which gives ok-ish results, but could be improved by using something like Christofides algorithm and doing attempts at improving the initial approximation. This is O(n²) time, however even for 10k images this should still be much quicker than the embedding step. The embeddings are therefore cached, usually in `$HOME/.cache/embeddings-sort`.