Impossible to use training sets kept in other location than INSTALLATION_DIR

[pdec]

I noticed that the below check forces users to keep their retrained reference datasets in INSTALLATION_DIR/PhiSpyModules/data directory

PhiSpy/PhiSpyModules/classification.py

Lines 34 to 36 in 540fbf2

	if not pkg_resources.resource_exists('PhiSpyModules', trainingFile):
	sys.stderr.write("FATAL: Can not find data file {}\n".format(trainingFile))
	sys.exit(-1)

The same will happen with kmers file within makeTest.py and makeTrain.py

PhiSpy/PhiSpyModules/makeTest.py

Lines 16 to 20 in 540fbf2

	kmers_file = 'data/phage_kmers_' + self._kmers_type + '_wohost.txt'
	if not pkg_resources.resource_exists:
	sys.exit("ERROR: Kmers file {} not found".format(kmers_file))
	for line in pkg_resources.resource_stream('PhiSpyModules', kmers_file):

Hence, while retraining one needs to point the PhiSpy's installation dir (e.g. /usr/local/lib/python3.6/site-packages/PhiSpy...) and can't keep separate trainSets dirs.

[linsalrob]

I agree with this issue, but will not refactor the code for the current release. We will work on this for the next release.