bhallalab/findsim

FindSim - A Framework for Integrating Neuronal Data and Signaling Models

findsim's Issues

Featureneed to plot extra plots for doseResponse or BarChart

python3.8 findSim.py ../htpipeline/Expts/Troca-Martin2011_Fig2A.json -m ../htpipeline/Models/Level1_HT_model.json --map ../htpipeline/Maps/map_ht.json --plot aTrKB.conc

File "findSim.py", line 1522, in innerMain
print("Warning: Experiment design is '{}'. Only 'TimeSeries' supports extra plots. Skipping".format( experiment.exptType ) )
NameError: name 'experiment' is not defined

few checks needed in simWrapMoose

Need cleanup in simWrapMoose

case A: entity exist in mapping file but molecule doesn't exist in model e.g "eIF4EmRNA_bind_GA":["eIF4EmRNA_eIF4GA_clx_formation"] we get seg fault, this because we are trying to delete shell
[in buildModelLookup() molecules gets root (moose.Shell('/')) path and populates self.modelLookup with key as enitity and path as shell]
and in deleteItems() trying to delete moose.Shell('/')

ideally it should ("Alert: simWrapMoose::Object in entity list but not in model '{}'".format(entity))

case B: entity entered in the experiment sheet, which doesn't exist in the mapping file then it silently ignored with out warning the user and runs e.g "eIF4EmRNA_doesnotexist"

print( "Alert: simWrapMoose::deleteItems: Object in entity list but not in mapping file '{}'".format( entity ) )

case C: checks for object.name to root need to changed to '/'

Attaching Jain2009_Fig4F_old.json
Jain2009_Fig4F_old.txt

FindSim-Bhalla1999_fig4C with sbml model result are different

Bhalla1999_fig4C with SBML model not working, its happened from older version like V1.0 to till date LegacyVersion1.2.0. Need to check

Subsets are not recognized with findSim.py

findSim.py gives the following Error when an experiment that has subsets is being used:

File "FindSim/findSim.py", line 1126, in innerMain
sw.loadModelFile( model.fileName, model.modify, scaleParam, dumpFname, paramFname )
File "/home/bhalla/nishaav/SynPlast/HTAutSim/FindSim/simWrapHillTau.py", line 210, in loadModelFile
modifyFunc( {}, "" ) # Callback.
File "FindSim/findSim.py", line 587, in modify
sw.subsetItems( self.modelSubset )
File "/home/bhalla/nishaav/SynPlast/HTAutSim/FindSim/simWrapHillTau.py", line 143, in subsetItems
raise SimError( "SimWrapHillTau::subsetItems: Entity '{}' not found".format( i ) )
simError.SimError: "SimWrapHillTau::subsetItems: Entity 'PKC_g' not found"
Failed for expt defn---------------: findSim failed for exptDefn Expts/Schaechter1993_Fig8.json: "SimWrapHillTau::subsetItems: Entity 'PKC_g' not found"

It fails to recognize the subsets listed out in experiment Modifications section. However, if the subet is set to 'all' it is able to run the experiment without any error messages. Experiment used for testing is Schaechter1993_Fig8.json

Wiki

I've added documentation by Vinod to the wiki. https://github.com/BhallaLab/FindSim/wiki

If readout molecule is passed via --plot option which is duplicate then we get Assertion Error

E.g
python findSim.py Wei-Lun2015_Fig4C_BDNF.json -m Level1_HT_model.json --map map_ht.json --plot aTrKB.conc protein.conc

protein molecule already exist in Readout, if passed again with --plot option we get
rd.displayPlots( exptFile, model._tempModelLookup, stims, hideSubplots, expt.exptType, bigFont = bigFont )
File "findSim.py", line 442, in displayPlots
assert( numPts > 0 )
AssertionError

rebuild model without Hsolve error message

In SimWrapMoose.py
buildSolver()
The SimError (raise SimError( "Hsolve already created. Please rebuild model without HSolve.)

This error message needs more information
raise SimError( "Hsolve already created. Please rebuild model without HSolve. In rdesigneur use the 'turnOffElec = True' flag." )

moose version compatibility

When importing moose, FindSim should check if available version is compatible with it or not. Current devel branch is not compatible with current stable 3.1.4 release.

Plotting issue with TimeSeries experiments

In findSim.py, under parseAndRun function, a call to 'advanceSimulation' under simWrapMoose.py is made but it is not handling the doPlot argument. This is happening in simWrapHillTau.py and therefore TimeSeries experiments works well with HillTau models but its not happening with detailed model.

Line 754-755:
if ( qe.t > currt ):
sw.advanceSimulation( qe.t - currt, doPlot = getPlots )

Unable to replicate some of the Experiments using HillTau model

With the new simWrapHillTau.py, I am unable to replicate some of the experiments using HillTau model esp. DoseResponse and BarCharts.

One subset that has shown significant differences is the PKC subset.
Schaechter1993_Fig6.json
Schaechter1993_Fig7.json
Schaechter1993_Fig8.json
Schaechter1993_Fig1B_presenceOfCa.json
Schaechter1993_Fig1B_absenceOfCa.json
Andreozzi2003_Fig1A.json

runParallel error

I'm running FindSim on my Ubuntu18.04 machine.
When running runAllParallel.py file, error happens:

............multiprocessing.pool.RemoteTraceback:
"""
Traceback (most recent call last):
File "/usr/lib/python3.6/multiprocessing/pool.py", line 119, in worker
result = (True, func(*args, **kwds))
File "/home/houchen/HouChen/GSoC_INCF/FindSim/findSim.py", line 1518, in innerMain
model.modify( modelId, erSPlist,modelWarning )
File "/home/houchen/HouChen/GSoC_INCF/FindSim/findSim.py", line 693, in modify
obj.setField( field, value )
AttributeError: 'moose.BufPool' object has no attribute 'isBuffered'
"""

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
File "runAllParallel.py", line 124, in
main()
File "runAllParallel.py", line 107, in main
results = [ i.get() for i in ret ]
File "runAllParallel.py", line 107, in
results = [ i.get() for i in ret ]
File "/usr/lib/python3.6/multiprocessing/pool.py", line 670, in get
raise self._value
AttributeError: 'moose.BufPool' object has no attribute 'isBuffered'

Do you know the reason or how to fix this?

My python version is 3.6.7.
And I used pip install pymoose to install moose.
(P.S. when building moose from repo, ctest#21 & #26 fails, I don't know weather this has something to do with errors when running FindSim.)

FindSim should be installable via pip

This is to track the progress.

Expected command

$ pip install findsim    # install stable
$ pip install findsim --pre # install nightly.

It depends on pymoose and scipy.

• PR #19
• When accepted, add this to deploy repo to automatically deploy wheel every week/day.

seg fault: dictionary stores moose Id instead of path

simWrapMoose -> buildModelLookup()
The dictionary modelLookup, the value stored is Id, instead path should be stored.
We get seg fault when we try to access the dictionary when the value objId is deleted.

Replace `imp` module with `importlib`

Module imp which has been deperecated since python3.4 is still being used in FindSim. Its should be replaced by importlib https://docs.python.org/3.7/library/importlib.html#module-importlib .

Couple of tests are failing but without assert travis ignore them

/opt/ANYONE/PY27/bin/python findSim.py ./TestTSV/fepsp.tsv --model models/synSynth7.g
Error: findSim failed for script ./TestTSV/fepsp.tsv: "findObj: No object found named: 'elec/soma'"
[INFO] Running experiment ./TestTSV/iclamp_hh13.tsv
/opt/ANYONE/PY27/bin/python findSim.py ./TestTSV/iclamp_hh13.tsv --model models/synSynth7.g
Error: findSim failed for script ./TestTSV/iclamp_hh13.tsv: "findObj: No object found named: 'elec/soma'"

FindSim prints the error but does not raise exception or return with non-zero code, therefore travis does not mark it as failure. Ideally, in case of error, findsim should return non-zero status not exception.