L3_TPC:B type label should be replaced with L3_TPC:C

[crisely09]

Some time ago they decided that this was an incorrect label for these MTypes, so everything that uses L3_TPC:B as MType should be changed to L3_TPC:C

Related to this Ticket.

[mgeplf]

@lecriste Isn't this something that can be controlled by changing the configuration input files? If we're not making that configuration accessible, that's something we have to change.

[lecriste]

@mgeplf, which configuration files? I can see both L3_TPC:B and L3_TPC:C present in several files in this repo.

[mgeplf]

Which step is creating the the L3_TPC densities? I'd suspect it's the excitatory-split stage, so whatever is being passed in to --cortex-all-to-exc-mtypes.

[lecriste]

It's this file. What about the other occurrences of L3_TPC:B in the repo?

[mgeplf]

I think they are there as example configs, and as part of testing. Thus, I think they can lag what the 'state of the art' configs are.

[dkeller9]

All L3_TPC:B should be changed to L3_TPC:C

[dkeller9]

I changed the csv file in the "Csv update" pull request.

[mgeplf]

We will discuss this in the atlas meeting.

[dkeller9]

@mgeplf i am getting a lint error in the "csv update" PR. Do you know why this might be occurring? https://github.com/BlueBrain/atlas-densities/actions/runs/5486872473/jobs/9997539596

[mgeplf]

I will look into it.

[mgeplf]

This is an issue with upstream dependencies, which will likely be solved shortly: https://github.com/pallets/click/issues/2558

However, this should stop us from testing the changes, to make sure the correct outputs are created. Can @dkeller9 and @lecriste verify this? In addition, I don't think it's practical that we are making configuration changes in a code repository that requires a release to get into production. I think that configuration should be tracked, versioned and validated separately.

[dkeller9]

We should just verify that the [L3_TPC:C] nrrd after the change is the same as [L3_TPC:B] before the change. @lecriste can you perform this check?

[lecriste]

@dkeller9, check performed. Out of curiosity, in which way such a label change could affect the file content?

[dkeller9]

It should not affect the file content, but it is good to verify that this is the case. If the label change had thrown a hidden error and the file was not created or nans, that would have been a problem.

[lecriste]

The new METype density is in Nexus prod: https://bbp.epfl.ch/nexus/v1/resources/bbp/atlas/_/https:%2F%2Fbbp.epfl.ch%2Fdata%2Fbbp%2Fatlas%2Fbb1364df-9b78-4dbb-bacf-ed6927f747e3

and in the CellComposition.

[crisely09]

Nice, thanks @lecriste