Types from header

[PeterNSteinmetz]

Factors out parsing of information from header of file. Add determination of recording type from header information.

Changes indexes to indices in argument names again -- will need determination of acceptability of this spelling change for files being worked on only. See https://grammarist.com/usage/indexes-indices/ for discussion of usage, particularly in technical contexts.

This factorization is in preparation for larger changes to accommodate early file styles with different header structures. Provided here as a smaller chunk in order to avoid very much larger changes which are harder to check and understand.

[pep8speaks]

Hello @PeterNSteinmetz! Thanks for updating this PR. We checked the lines you've touched for PEP 8 issues, and found:

There are currently no PEP 8 issues detected in this Pull Request. Cheers! 🍻

Comment last updated at 2020-11-06 21:42:25 UTC

[apdavison]

Many thanks for this. A quick comment on the "indices/indexes" question. Having grown up in the UK, I prefer "indices" in general. However, we decided to use "indexes" in Neo because it is less confusing for non-native English speakers and more familiar to North Americans. It's fine to use "indices" for internal variables, but any arguments that are exposed to Neo users as part of the API should be "indexes".

[PeterNSteinmetz]

A quick comment on the "indices/indexes" question. Having grown up in the UK, I prefer "indices" in general. However, we decided to use "indexes" in Neo because it is less confusing for non-native English speakers and more familiar to North Americans. It's fine to use "indices" for internal variables, but any arguments that are exposed to Neo users as part of the API should be "indexes".

So how about the case of the argument to _get_analogsignal_chunk ? That is nominally a private method but I imagine a lot of people use that. Should those also be 'indexes' ?

[PeterNSteinmetz]

circleci is complaining that the test is failing because trying to download the file https://web.gin.g-node.org/NeuralEnsemble/ephy_testing_data/raw/master/neuralynx/Cheetah_v4.0.2/original_data/CSC14_trunc.Ncs is returning a 404 error.

Yet that file is available and downloadable in a browser or if I pull it using git from the gin-g repository.

What is going wrong there?

[JuliaSprenger]

Hi @PeterNSteinmetz, I restarted the tests and the files could now be successfully downloaded from gin. It seems this was a temporary issue with gin. The tests are succeeding now.

[JuliaSprenger]

A quick comment on the "indices/indexes" question. Having grown up in the UK, I prefer "indices" in general. However, we decided to use "indexes" in Neo because it is less confusing for non-native English speakers and more familiar to North Americans. It's fine to use "indices" for internal variables, but any arguments that are exposed to Neo users as part of the API should be "indexes".

So how about the case of the argument to _get_analogsignal_chunk ? That is nominally a private method but I imagine a lot of people use that. Should those also be 'indexes' ?

Yes, I think we should stay consistently with channel_indexes for now and discuss a potential name change in a separate issue. Could you please also use indexes for _get_analogsignal_chunk?

[PeterNSteinmetz]

I will merge my NlxFixes branch into this PR, the TypesFromHeader branch, thereby fixing the channel_indexes question, once PR #851 is merged onto main. I want to leave NlxFixes open for now until that happens.

[JuliaSprenger]

#851 is now merged in master and here. Are you planning to address the todos introduced in this PR also here or is this PR ready for review / merge from your side?

[PeterNSteinmetz]

That is it for this branch. Please merge into master.

[PeterNSteinmetz]

This is ready for merging.

[JuliaSprenger]

There are a few naming issues I found, but otherwise the PR looks ready to merge to me. Thanks!

[PeterNSteinmetz]

I think tests are still failing on hdf5io .

[PeterNSteinmetz]

OK, just merged NeuralEnsemble:master back into my TypeFromHeader branch. Let's see how this goes.