fix blast requirement

[bernt-matthias]

We just added this repo to the planemo monitor galaxyproject/planemo-monitor#43 and I noticed that conda struggles to find a blast+ package

[peterjc]

This rings a bell, I think some systems ended up with blast meaning legacy BLAST (e.g. Debiain), and others BLAST+ (e.g. conda).

I don't know why the linting is complaining about no .shed.yaml file, do you fancy writing one?

[bernt-matthias]

no .shed.yaml file

True. So this thing never made it to the toolshed. The question is: should it?

[peterjc]

Good point. I never ended up using it personally (we have a local mirror of the NCBI databases on our cluster, updated monthly, and our Galaxy BLAST configuration just pointed at that), and am not in a position to even try it right now (we don't have a Galaxy server currently).

[peterjc]

Belatedly should we merge this now?

[bernt-matthias]

Belatedly should we merge this now?

A few small changes would be needed (which I could do), but the biggest problem I have with the data manager is that it writes to blastdb in any case, but it should write to blastdb_p sometimes, isn't it? How abou the other datatables?

Should we add a select?

[peterjc]

I didn't notice that oversight.

[As an aside, I'm planning to ask the IUC to adopt this repository or at least the main BLAST wrappers later this year if I'm still not using Galaxy in my day job.]