Configurable

[rockyroer]

These updates allow the user to configure how the program works in settings.cfg:

1. They can assign a maximum number of songs to download
2. They can assign whether to add songs in ascending or descending order
3. They can type a list of what charts to create

These updates also include some reorganization of the ChartData logic so that everything depends on just the chart_id.

[aag]

A few general notes:

1. I've set up Travis CI in the project and it runs the linters and the tests on every commit. You can see the results of the runs here on this page. The checks are currently failing because there are some code formatting issues and the tests need to be updated. You can see the output of the checks by clicking the "Details" link next to each check. It's generally good practice to run all those commands locally before opening the PR, so you can fix any problems.

2. It seems pretty limiting to hard-code the chart_info data in the script. The way it is now, someone could configure a chart that exists, but isn't in the chart_info object, and the script wouldn't work. I think it would be better to include the chart name in the config file, or change billboard.py so it provides the name.

3. It seems to me like the max_number_of_songs and playlist_ordering settings should be chart-specific and not global to all charts. I might only want the top 10 songs from the R&B chart, but all of the Hot 100, but that wouldn't be possible with the current format.

Could you make a new commit that addresses those 3 things?

[rockyroer]

A few general notes:

1. I've set up Travis CI in the project and it runs the linters and the tests on every commit. You can see the results of the runs here on this page. The checks are currently failing because there are some code formatting issues and the tests need to be updated. You can see the output of the checks by clicking the "Details" link next to each check. It's generally good practice to run all those commands locally before opening the PR, so you can fix any problems.

I don't know how to run these tests locally.

2. It seems pretty limiting to hard-code the chart_info data in the script. The way it is now, someone could configure a chart that exists, but isn't in the chart_info object, and the script wouldn't work. I think it would be better to include the chart name in the config file, or change billboard.py so it provides the name.

I also don't like the hard-coded information. I was thinking of keeping the config file simple, with just list of chart-ids, and potentially putting the hardcoded data into a separate file -- like a csv file, that we could make for all possible charts. But that sounds like more work than its worth.
Honestly, I'd say the easiest thing to do would be to create a little simpler playlist description that is just the chart-id, number of songs, date and url, and not bother with a dictionary at all.

3. It seems to me like the max_number_of_songs and playlist_ordering settings should be chart-specific and not global to all charts. I might only want the top 10 songs from the R&B chart, but all of the Hot 100, but that wouldn't be possible with the current format.

I also wondered about that. What might that look like in the config file?
'hot-100', 25, DESCENDING
'country', 10, ASCENDING
Can you have a multi-line key in a cfg file?

[aag]

I don't know how to run these tests locally.

Good point. I've just pushed an update to the README with the command you need to run.

I'd say the easiest thing to do would be to create a little simpler playlist description that is just the chart-id, number of songs, date and url, and not bother with a dictionary at all.

That would be easier, but it would make the playlist title pretty ugly. I'm working on a patch to billboard.py right now that will deliver the human-readable title for each playlist, so hopefully that will get merged and we can use that.

What might that look like in the config file?

With the current config file parser, we might have to do something like this:

chart_1_id = hot-100
chart_1_name = "Hot 100"
chart_1_max_songs = 10

chart_2_id = country-songs
chart_2_name = "Hot Country Songs"
chart_2_sort = ASCENDING

Then in the load_config() function we'd have to start from 1, count up, and try to form valid config keys until a chart_X_id is missing.

Alternatively, we could switch to the TOML library, which would allow us to do something like this:

[charts]
  [chart.1]
  id = "hot-100"
  name = "Hot 100"
  max_songs = 10

  [chart.2]
  id = "country-songs"
  name = "Hot Country Songs"
  sort = ASCENDING

[aag]

I'm working on a patch to billboard.py right now that will deliver the human-readable title for each playlist, so hopefully that will get merged and we can use that.

It took a little while, but my patch got accepted and a new version of billboard.py was released. I've updated the Pipfile in the master branch, so if you merge it into your branch and do pipenv install --dev you should get the newer version. Here's how you would do the merge:

$ git checkout master
$ git pull
$ git checkout configurable
$ git merge --no-ff master

These commands first pull down the newest version of the master branch, then bring in any new things in that branch into your feature branch.

If you have uncommitted changes in your local working copy, you'll have to run git stash before the above 4 commands, and git stash pop after them.

Once you do this, you can call .title on a ChartData object to get its human-readable title and we don't have to allow the user to configure it anymore.