Script to fetch receipt images

[cuducos]

This is a program to download receipts to a local directory (fix #33).

$ python src/fetch_receipts.py
    This script downloads the receipt images from the Lower House server.

    Be aware that downloading everything might use more than 1 TB of disk
    space.  Because of that you have to specify one `target` directory (where
    to save the files) and optionally you can specify with `--limit` the number
    of images to be downloaded.

    If the `target` directory exists and already has some saved receipts,
    these receipts will not be downloaded again (and they will not count when
    using `--limit` either).

    In other words, if you already have 42 receipts in your target folder,
    running the command with a limit of 8 will end up in a directory with 50
    files: the 42 you already had and 8 freshly downloaded ones.


positional arguments:
  target                Directory where images will be saved.

optional arguments:
  -h, --help            show this help message and exit
  -l LIMIT, --limit LIMIT
                        Limit the number of receipts to be saved

It will download the files to a different target than data/ because it tends to use a huge amount of disk space (maybe more than 1TB).

That said people might want to point it to a external volume (using target positional argument) and/or to limit the amount of receipts to be downloaded (using --limit optional argument).

It uses a new external library (added to conda_requirements.txt) called humanize. You can manually install it with pip install humanize if you want.

Finally I recommend everyone that has a spare TB at home to save them until we can raise money to afford a virtual drive for the project with that amount of free space.

Using it to download everything to /Volumes/Narnia/Serenata:

$ python src/fetch_receipts.py /Volumes/Narnia/Serenata

Using it to download 10,000 receipts to /Volumes/Narnia/Serenata:

$ python src/fetch_receipts.py --limit 10000 /Volumes/Narnia/Serenata

[mtrovo]

I think the operation of downloading the receipts doesn't fit on the constructor of the class, maybe it's a good idea to separate it into a run method which is going to fetch the receipts and also a print_report which will print the error messages if there's something.

[cuducos]

That sounds like a good idea, thanks!

Do you wanna step in and send a PR to this branch? These days my workload is focused on Issue #34. Otherwise I'l catch up here later.

[cuducos]

Done, @mtrovo! Way better now. Thanks for the advice.

[mtrovo]

Hey @cuducos I made this change on a local branch and added some improvements to download images in parallel, can you take a look at #50.