#26 Remove line numbers to fix Cheetah_v5.6.3/plain_data

This pull requests modifies the two files in ephy_testing_data/neuralynx/Cheetah_v5.6.3/plain_data which have line numbers on each line, namely CSC1.txt and CSC2.txt. It addresses Issue #22.

Removing these should allow the same code in test_neuralynxio to test reading data from ncs files since the other versions don't have these line numbers.

This PR only contains the changes to the hashes in the softlinks to the annexed data. I have synced my annex on the removeLineNumbers branch in my repository so hopefully the annexed files can be retrieved from there, as suggested by @Achilleas Koutsou.

This trick actually didn't work. I will add the files directly.

I think the annexed files have to be added directly somehow. The PR only pulls the softlinks with hash references to the new annexed files.

Whatever gets the corrected ones in the annex!

I think you didn't push the content of the files to your fork of the repository. At least I can't retrieve the content of the file from there. If you are using the gin client, can you try gin upload --to <your_fork> <filename>?

I don't have the gin client installed, but will that push them to the annex?

I have synced the annex with 'git annex sync' and it acts like that branch in the repository is synced with my local. Achilleas acted like it would be necessary for you to pull them out of the annex on my branch in my repository somehow.

OTOH, I don't really understand how that would work. When I clone my repository right now, I get the new hash when I switch to that branch, but the underlying file is missing from my annex when I try a git annex get command.

Another alternative would be for you to simply do this separate and strip the line numbers off those two files in the annex and recommit them. I assume you have write privileges on NeuralEnsemble/ephy_testing_data ?

The line numbers are easily stripped at the command line with a cut command.

I just tried the gin command line client but am running into some issues with it. Here is a transcript of some commands I have tried and the output:

gin upload --to removeLineNumbers CSC1.txt :: Adding file changes Nothing to do :: Recording changes No changes recorded :: Uploading "removeLineNumbers" unknown remote name 'removeLineNumbers': skipping [error] 1 operation failed (base) Crol:ephy_testing_data peter$ gin repos [error] user '' does not exist

I have asked Achilleas to see this thread as well.

--to is used to specify the remote you want to push to. If you only have your fork set up as remote (ie you did git clone git@gin.g-node.org:/PeterNSteinmetz/ephy_testing_data.git or gin get PeterNSteinmetz/ephy_testing_data) then you don't need to specify --to.

If you manually configured multiple remotes (e.g. git@gin.g-node.org/NeuralEnsemble/ephy_testing_data and https://gin.g-node.org/PeterNSteinmetz/ephy_testing_data then you should push to your fork.

Hmm. Interesting. I think I needed to add some servers to gin, which I did resulting in the following config.yml:

servers: neuralensemble:

web:
  protocol: https
  host: gin.g-node.org
  port: 443
git:
  user: git
  host: gin.g-node.org
  port: 22
  hostkey: gin.g-node.org,141.84.41.219 ecdsa-sha2-nistp256 AAAAE2VjZHNhLXNoYTItbmlzdHAyNTYAAAAIbmlzdHAyNTYAAABBBE5IBgKP3nUryEFaACwY4N3jlqDx8Qw1xAxU2Xpt5V0p9RNefNnedVmnIBV6lA3n+9kT1OSbyqA/+SgsQ57nHo0=

I then did gin uploads which seem to have worked, claiming to go to the 'origin' remote (that is a remote I have configured for the git client and corresponds to my gin.g-node.org/PeterNSteinmetz/ephy_testing_data repository. Below is the log of those commands:

gin upload CSC1.txt :: Adding file changes Nothing to do :: Recording changes No changes recorded :: Uploading Uploading (to: origin) "(unknown)" OK
(base) Crol:plain_data peter$ gin upload CSC2.txt :: Adding file changes Nothing to do :: Recording changes No changes recorded :: Uploading Uploading (to: origin) "(unknown)" OK

So did that work now to put the annex files into my repository on gin.g-node.org ? It did report transferring approximately the correct amount of data.

Yes, the content is now there. At least in your fork. I will have a look and see if I can also move the content to NeuralEnsemble.

Cool. I think the other PRs should merge without any use of the annex as I just used small truncated test files for those.

Moving just the content seems more complicated than it should be. For now I added your files directly using the gin-cli.