This repository hosts a collection of MPI Data Science Modules

Collection currently includes

1. MPIcollections
2. MPInumpy
3. MPIpandas
4. MPIscipy

Software Requirements

Currently only Python3 fully supported

For usage

• mpi4py (https://mpi4py.readthedocs.io/en/stable/)
  • Requires working MPI implementation
  • version 3.0.0
• NumPy (https://numpy.org/)
  • version 1.10+ for full MPInumpy usage
• nltk (https://www.nltk.org/)
  • version 3.3+
• pandas (https://pandas.pydata.org/)
  • version 0.16+
• petsc4py (https://pypi.org/project/petsc4py/)
  • Requires working MPI implementation
  • version 3.12+

For tests

• coverage (https://coverage.readthedocs.io/en/coverage-5.1/)
  • version 4.5+
• nose (https://nose.readthedocs.io/en/latest/)
  • version 1.3+
• paver (https://github.com/paver/paver)
  • version 1.3+

Running tests

Locally

Results saved in mpids_local_unit.out

/bin/bash mpids_unittest.sh

On a cluster

Results saved in mpids_crill_unit.out

Note: Script configured for UH PSTL Crill Cluster

sbatch mpids_unittest_batch.sh

Cleaning up

paver clean