Support JSON log output

[jfy133]

To carry over from: https://github.com/ewels/MultiQC/issues/1005, just wanted to have this open to make the conversation 'cleaned up' in terms of requests.

As stated on the PR, @MikkelSchubert had stated an interest in improving the output log (.settings), and also making a JSON version of the log (with better documentation).

A few requests from @apeltzer and I in how we use AR2 in nf-core/eager, what would be great would be:

1. JSON format as close to MultiQC accept JSON as possible (or easily modifiable to make it work). Documentation
2. An entry for % of input reads that have had adapters removed

• This is a useful pre-screening indicator whether your DNA fragments are short, as expected for aDNA, therefore most reads have adapters

3. An entry of % of read pairs that were collapsed

• This is anther useful pre-screening indicator, as aDNA reads are so short, you expect them to merge.

Note for 2/3 both are only meant to be rough indicators, so they don't have to be 100% exact - given some of the complexity in the way AR2 works (see MultiQC thread).

[jfy133]

Ping @apeltzer for tracking

[MikkelSchubert]

An initial stab at a JSON report is now available in the master branch (v3 alpha, see comment in #52 for more information).

Right now I've just focused producing something that is useful in its own right, so it's not yet MultiQC compatible. But generic MultiQC compatible output is probably something that could be included as well since most of the work is already done (though obviously it'd be better with proper support in MultiQC).

You can see an example of the current layout here: example.json.gz. You can download and run the examples in the github repo if you want to see output in different scenarios.

Feedback would be greatly appreciated.

[jfy133]

As on the other issue, I'll look into this early next week for you !