Nof models mismatch between modeldb API and senselab

[alexsavulescu]

With ModelDB API we get:

• http://modeldb.science/api/v1/models?modeling_application=NEURON -> 680 models <- these are run in the ModelDB CI

Senselab gives us

• https://senselab.med.yale.edu/ModelDB/ModelList?id=1882 -> 738 models

Missing accession numbers:

249705,257027,258738,264539,266498,266732,266775,266782,266799,266802,266806,266811,266814,266818,266823,266842,266848,266850,266851,266864,266871,266872,266881,266900,266901,266905,266925,266984,266989,267009,267013,267018,267050,267056,267063,267066,267067,267103,267106,267140,267142,267146,267176,267184,267189,267201,267259,267293,267298,267307,267311,267384,267391,267499,267508,267509,267511,267512

cc : @ramcdougal @nrnhines

LE 2023: 6 more models 185864 267306 267587 267594 267595 83523

[alexsavulescu]

Side note: vast majority of missing models start with 26.. (maybe this is helpful)

[alexsavulescu]

Only 3/58 models are available on GH: 249705, 266775 and 266782

[ramcdougal]

Not a bug per se... it's just that the data literally hasn't been transferred over.

There's a slight gotcha in that one of the models that hasn't been transferred unzips to be quite large and we have very little memory on modeldb.science. This causes a problem for indexing the file with the code as written.

[alexsavulescu]

it's just that the data literally hasn't been transferred over.

Could we still create repositories on GitHub for these models and upload there?

[ramcdougal]

I've added it to my near-term TODO list.

[alexsavulescu]

All missing models have been published on GH. For 267184 because files were large, I had to use git LFS. One caveat is that lfs requires an extra installation.

[ramcdougal]

ModelDB.science was synced from ModelDB.yale.edu around the end of 2022, so all of those are there now.

[olupton]

I think this can be closed now that we are consistently fetching from ModelDB.science.