[BBPBGLIB-1116] changes nakpump and ProbGABAAB_EMS solving Nan for ion specific model

[sfarina05]

Based on the Nan Issue reported here: https://bbpteam.epfl.ch/project/issues/browse/BBPBGLIB-1116, I have changed two mod file for metabolism based on the suggestions of Michel Hines and Pramod:
-nakpump.mod has now the voltage dependency as the original model of Chapman 1983
-ProbGABAAB_EMS.mod has reversal potential defined by chloride and Potassium
These two changes seem to remove the Nan and nakpump voltage dependent help in keeping ATP level in physiological range.

[WeinaJi]

I've launched the integration test with this branch https://bbpgitlab.epfl.ch/hpc/sim/blueconfigs/-/pipelines/218971

[WeinaJi]

I've launched the integration test with this branch https://bbpgitlab.epfl.ch/hpc/sim/blueconfigs/-/pipelines/218971

This change doesn't work with the multiscale run test in the neurodamus pipeline, see failed job
https://bbpgitlab.epfl.ch/hpc/sim/blueconfigs/-/jobs/1331767

[VERB]  -> Loading metypes from: /gpfs/bbp.cscs.ch/project/proj12/jenkins/subcellular/blueconfigs_test/simulation-sonataconf/base_circuit/emodels_atp_scan/emodels_atp_1p4
[VERB]  -> Loading 'asc' morphologies from: /gpfs/bbp.cscs.ch/project/proj12/jenkins/subcellular/blueconfigs_test/simulation-sonataconf/base_circuit/morphologies/fixed_ais_L23PC_20201210/ascii
[INFO]  > Instantiating cells... (1 in Rank 0)
4 special: syntax error
4  near line 1
4  {cNAC_L23BTC[0].axon[0] totalpump_nakpump(0.100000) = (0.0 * 10.1763 + 1.0)*0.000000000000400}
                                                     ^
        4 cNAC_L23BTC[0].execute("cNAC_L23BT...")
      4 cNAC_L23BTC[0].distribute_distance(SectionList[2], "totalpump_...", "(0.0 * %g ...")
    4 cNAC_L23BTC[0].biophys()
  4 cNAC_L23BTC[0].init(933, "/gpfs/bbp....", "sm080930a1...", 0)

@sfarina05 , do we need to update the above emodl path in our test? Does it mean that the new mod files won't work with old circuit?

[WeinaJi]

I've launched the integration test with this branch https://bbpgitlab.epfl.ch/hpc/sim/blueconfigs/-/pipelines/218971

This change doesn't work with the multiscale run test in the neurodamus pipeline, see failed job https://bbpgitlab.epfl.ch/hpc/sim/blueconfigs/-/jobs/1331767

[VERB]  -> Loading metypes from: /gpfs/bbp.cscs.ch/project/proj12/jenkins/subcellular/blueconfigs_test/simulation-sonataconf/base_circuit/emodels_atp_scan/emodels_atp_1p4
[VERB]  -> Loading 'asc' morphologies from: /gpfs/bbp.cscs.ch/project/proj12/jenkins/subcellular/blueconfigs_test/simulation-sonataconf/base_circuit/morphologies/fixed_ais_L23PC_20201210/ascii
[INFO]  > Instantiating cells... (1 in Rank 0)
4 special: syntax error
4  near line 1
4  {cNAC_L23BTC[0].axon[0] totalpump_nakpump(0.100000) = (0.0 * 10.1763 + 1.0)*0.000000000000400}
                                                     ^
        4 cNAC_L23BTC[0].execute("cNAC_L23BT...")
      4 cNAC_L23BTC[0].distribute_distance(SectionList[2], "totalpump_...", "(0.0 * %g ...")
    4 cNAC_L23BTC[0].biophys()
  4 cNAC_L23BTC[0].init(933, "/gpfs/bbp....", "sm080930a1...", 0)

The ngv test in the neurodamus CI is using the old emodels in /gpfs/bbp.cscs.ch/project/proj12/jenkins/subcellular/blueconfigs_test/simulation-sonataconf/base_circuit/emodels_atp_scan/emodels_atp_1p4 where not all the hoc templates contain totalpump. One option is to update the emodel path in the circuit_config.json for this test. Maybe we should do that.

[WeinaJi]

As discussed with @sfarina05 , we should update the emodels in our neurodamus CI test https://bbpgitlab.epfl.ch/hpc/sim/blueconfigs/-/blob/main/sonataconf-quick-multiscale/circuit_config.json?ref_type=heads (the value of the key "biophysical_neuron_models_dir") with the V10 version
"/gpfs/bbp.cscs.ch/project/proj137/NGVCircuits/rat_sscx_S1HL/V10/emodels/hoc"
@atemerev Could you please create a MR in the blueconfigs repo ? Thanks!

[WeinaJi]

LGTM. Once this commit is tagged and deployed on bb5, we need to update the neurodamus CI test "sonataconf-quick-multiscale" and its reference reports.
https://bbpgitlab.epfl.ch/hpc/sim/blueconfigs/-/merge_requests/123

[pramodk]

From the MOD file side, LGTM!