A new organisation for pyani

[widdowquinn]

As noted by Titus, it would be good to move pyani's codebase to a standalone organisation, dissociating it from my personal set of repositories.

We could put the repo under sipbs-compbiol, but that still associates it with the university. I'm inclined to set up a completely separate organisation so that the whole codebase can be handed off, if necessary. For that we need a name that isn't taken.

The obvious choice for an organisation name would be pyani - but someone has already registered that as a username, so it can't be an organisation. We need a different name.

I'm currently thinking pyani-org or something similar, but it doesn't exactly sound inspired. Please add suggestions below.

[baileythegreen]

I actually don't think it makes sense to put pyani in the name of an organisation like this. I would use the organisation as a level of abstraction; use a name that is not tied to one tool in particular, because then you allow the possibility of housing other tools under its umbrella. Currently the only thing that might eventually also be there is pdp, I guess, but that can change.

I agree with not aligning it to the university; I will think about actual suggestions for the name.

[widdowquinn]

Thanks - I look forward to the suggestions.

I don't think I want to have an umbrella organisation with multiple tools, in this instance. I had imagined this organisation would cover only pyani, its documentation, and website (and whatever else comes up tied specifically to pyani).

The motivation is that the tool itself has become generally useful, and to have it tied to my own set of repositories is a hindrance to moving it on to the community (should that be a useful thing to do; could help avoid the bus problem to an extent, too). I don't know that the community would thank us for saddling them with a bunch of additional tools along with it ;)

[baileythegreen]

Some suggestions. I'm still thinking about this, so there may be more.

• pANIply — a play on 'panoply'¹, as in, 'pyani offers a panoply of ways to run ANI'
• orgANI — pretty self-explanatory, I think²
• mANIc — microbial ANI clique³/crowd/covey/crew/club
• mANIa — microbial ANI association

It doesn't have to have ANI in the name, I know; but moving away from that I haven't yet found anything I thought worked well.

You'll notice none of my suggestions mention the language. I don't think the name of the organisation should; someone, at some point, might decide to port it to a different language, but that would be a bit weird if the organisation name mentioned Python.

Footnotes

1. Yes, I was listening to The Decemberists when I came up with this; why do you ask? (Bonus points if you know the song in question.) ↩

2. See also variants on organism/organic/organise. ↩

3. Obviously the best choice in this list. ↩

[widdowquinn]

ATM I seeing the purpose of the organisation being to house pyani and dissociate it from one individual (i.e. me) or organisation (i.e. my employer - despite the need for the copyright statement, etc.), so I'm inclined towards names that make this intuitively clear to the user. Prosaic as it is, something like pyani-org or pyani-project appeals to me more.

I don't know that porting to a different language is that likely to happen. One might argue that, by implementing the same analyses as JSpecies, pyani is a "port" of JSpecies to Python - but I'd disagree. I think it's a reimplementation of the independently-published methods, in Python. I'd see reimplementation of some of the measures in another language (with different interface) as being much more likely than porting.

[baileythegreen]

I'm currently thinking pyani-org or something similar, but it doesn't exactly sound inspired. Please add suggestions below.

I was trying to give some suggestions that might count as 'inspired'. If you're still wanting to call it pyani-org or something similar, maybe just do it.

[baileythegreen]

I think moving pyani to a separate organisation will help with an issue I seem to be having. I can't view reports from CodeCov (not including what gets generated for a commit to a PR). I get a 403 Permission Error and this link seems to indicate it is because the 'organisation' (I guess, your account here) only has one owner (and it's not me).

[widdowquinn]

There is, at least, a workaround until we make the move - the coverage plugin should generate an equivalent report when pytest is run locally (as you probably already know).

[baileythegreen]

I don't, actually. Does this happen automatically? I was wanting to look at the coverage in individual files; I found that I can see the proportikn of each that is covered, but no more detail than that.

[widdowquinn]

Have a go with pytest --cov-report=html --cov=pyani -v tests/ (there's a wee description in https://github.com/widdowquinn/pyani/blob/master/CONTRIBUTING.md#local-testing).