reset multiqc between runs

[jchorl]

Currently, multiqc leverages global state to store info. If you run it twice, you get weird results - e.g. table columns get duplicated. This moves the init logic into a re-callable function, such that multiqc can reset state between runs.

• This comment contains a description of changes (with reason)
• CHANGELOG.md has been updated

I wrote a couple quick scripts to test:

$ cat other.py 
def init():
    global foo
    foo = ["bar"]

$ cat test.py 
import other

other.init()
print(other.foo)
other.foo.append("baz")
print(other.foo)
other.init()
print(other.foo)

$ python3 test.py 
['bar']
['bar', 'baz']
['bar']

[ewels]

hah, smart! Yes this has been bugging me for a while, one of those things where I was less experienced with Python when I started MultiQC and didn't know better.

I have been vaguely planning to refactor a lot of this code to use objects instead, eg. have a multiqc.Report() object with all of this state information. However that's a bigger job and is one of many things that I've been putting in a v2.0 rewrite in my head. This could be a nice simple workaround until then.. 👍🏻

[ewels]

Not tested locally, but CI is still passing and I think if anything was wrong they would fail.

If you could please just add a small note in the changelog that'd be fab, then I'll merge 👍🏻

[jchorl]

No global state would be nice :)

Updated the changelog.

[ewels]

The modules already use classes, so it's mostly a report object, a template object(?), a config object and then a MultiQC run object I guess.. If you fancy a challenge 😉