RAVEN not using correct extension for output file in clarg for GenericCode

[Paul Philip Sprague]

Hello,

 

I’m using RAVEN with OpenMC: an open-source Monte Carlo neutron transport code with a Python API.

 

I’ve written a python script (In this email I’ve renamed it OpenMC-script.py) that takes an input file with the “-i” argument (extension .i), builds the OpenMC input files and runs OpenMC, and then generates an output file with the name specified by the “-o” argument (extension .csv).

I’ve run into an issue where RAVEN is not using the specified file extension for my generic code output file clarg.

 

From this sample of my RAVEN input file with the offending argument highlighted in orange (I’ve only included the relevant nodes to how RAVEN is executing my generic code):

 

                ...

 

                <Files>

           <Input name="commonInput" type="">input_file.i</Input>

     </Files>

               

<Models>

           <Code name="OpenMC_Model" subType="GenericCode">

                <executable>/path/to/OpenMC-script.py</executable>

                <clargs arg="'mpirun -n 1 /path/to/OpenMC-conda-env/python -m mpi4py " type="prepend"/>

                <clargs arg=" -i " extension=".i" type="input"/>

                <clargs arg=" -o " extension=".csv" type="output"/>

                <clargs arg=" --run --mpi 24 " type="postpend"/>

           </Code>

     </Models>

 

     <Steps>

           <MultiRun name="sample" verbosity="all">

                <Input    class="Files"         type="">commonInput</Input>

                <Model    class="Models"        type="Code">OpenMC_Model</Model>

                <Sampler   class="Samplers"      type="MonteCarlo">monteCarlo</Sampler>

                <Output    class="DataObjects"   type="PointSet">samples</Output>

                <Output    class="DataObjects"   type="HistorySet">histories</Output>

           </MultiRun>

     ...

     </Steps>

 

     ...

 

Note that I have replaced “/path/to” with the correct path, and I am calling an installation of python from another environment due to dependency incompatibility between RAVEN and the OpenMC Python API

 

Expected shell execution command:

 

                $ mpirun -n 1 /path/to/OpenMC-conda-env/python -m mpi4py /path/to/OpenMC-script.py -i input_file.i -o out~input_file.csv --run --mpi 24

 

Actual shell execution command:

     $ mpirun -n 1 /path/to/OpenMC-conda-env/python -m mpi4py /path/to/OpenMC-script.py -i input_file.i -o out~input_file --run --mpi 24

 

See attached RAVEN input file for full input.

 

Because of this, I am getting an error message:

     Traceback (most recent call last):

           File "/home/sprapp/.conda/envs/raven_openmc/lib/python3.10/threading.py", line 1016, in _bootstrap_inner

                self.run()

           File "/home/sprapp/.conda/envs/raven_openmc/lib/python3.10/threading.py", line 953, in run

                self._target(*self._args, **self._kwargs)

           File "/home/sprapp/projects/raven/ravenframework/Runners/SharedMemoryRunner.py", line 115, in <lambda>

                self.thread = InterruptibleThread(target = lambda q, *arg : q.append(self.functionToRun(*arg)),

           File "/home/sprapp/projects/raven/ravenframework/Models/Code.py", line 671, in evaluateSample

                csvData = csvLoader.loadCsvFile(outFile.getAbsFile(), nullOK=False, utility=loadUtility)

           File "/home/sprapp/projects/raven/ravenframework/CsvLoader.py", line 53, in loadCsvFile

                return self._loadCsvPandas(myFile, nullOK=nullOK)

           File "/home/sprapp/projects/raven/ravenframework/CsvLoader.py", line 70, in _loadCsvPandas

                df = pd.read_csv(myFile)

           File "/home/sprapp/.conda/envs/raven_openmc/lib/python3.10/site-packages/pandas/io/parsers/readers.py", line 1026, in read_csv

                return _read(filepath_or_buffer, kwds)

           File "/home/sprapp/.conda/envs/raven_openmc/lib/python3.10/site-packages/pandas/io/parsers/readers.py", line 620, in _read

                parser = TextFileReader(filepath_or_buffer, **kwds)

           File "/home/sprapp/.conda/envs/raven_openmc/lib/python3.10/site-packages/pandas/io/parsers/readers.py", line 1620, in __init__

                self._engine = self._make_engine(f, self.engine)

           File "/home/sprapp/.conda/envs/raven_openmc/lib/python3.10/site-packages/pandas/io/parsers/readers.py", line 1880, in _make_engine

                self.handles = get_handle(

           File "/home/sprapp/.conda/envs/raven_openmc/lib/python3.10/site-packages/pandas/io/common.py", line 873, in get_handle

                handle = open(

FileNotFoundError: [Errno 2] No such file or directory: '/path/to/wd/sample/10/out~common_input.csv'

 

I believe this is a RAVEN error, not an OpenMC error, as it is coming from the python installation used by RAVEN, not the one called in the Model.

 

Any suggestions? I hope this is an easy fix but could not find anything in the user guide or manual.

 

Thanks!

 

Paul Sprague (he/him/his)

Reactor Engineering Intern   |   MFC Reactor Engineering

paul.s...@inl.gov   |   208-526-0025   |   541-200-4096

Idaho National Laboratory   |   PO Box 1625, MS 3605   |   Idaho Falls, ID   |   83415

_______________________________