This repository is the working directory for the Garnet-Forest bundle of python scripts for analyzing diverse forms of 'omic' data in a network context.
This pull request adds an end-to-end integration test for Forest using the pytest package. It runs Forest on example data and compares the new output files to a saved set of gold standard output files. Instructions for running the test have been added to a new readme in the tests subdirectory.
In order to have the Forest outputs be fully deterministic, forest.py now sorts the input before sending it to msgsteiner and sorts the output before writing Cytoscape-formatted files. We have confirmed that msgsteiner's output does depend on the order of the nodes and edges in the input file, and that order varies on different hardware.
These changes close issue #8 and address the Forest-only case in issue #1. Given this model for integration testing, it would be straightforward to add new tests for Garnet-only and Garnet+Forest. This would require a fast Garnet test case.
This pull request adds an end-to-end integration test for Forest using the pytest package. It runs Forest on example data and compares the new output files to a saved set of gold standard output files. Instructions for running the test have been added to a new readme in the tests subdirectory.
In order to have the Forest outputs be fully deterministic,
forest.pynow sorts the input before sending it to msgsteiner and sorts the output before writing Cytoscape-formatted files. We have confirmed that msgsteiner's output does depend on the order of the nodes and edges in the input file, and that order varies on different hardware.These changes close issue #8 and address the Forest-only case in issue #1. Given this model for integration testing, it would be straightforward to add new tests for Garnet-only and Garnet+Forest. This would require a fast Garnet test case.