Improve reading of CSV file to be independent of the current working directory

[jcfr]

The reading logic should be improved so that successful loading of a csv file does not depend on the current working directory.

Instead the function pcaExplorer.readCSVFile( should be updated to resolve relative path based on the location of the csv file.

Once this is implement, use of the following context manager should not be needed:

ShapeVariationAnalyzer/ShapeVariationAnalyzer/ShapeVariationAnalyzer.py

Lines 2564 to 2566 in 19c8aa2

	with slicer.util.chdir(input_dir):
	logic.pca_exploration.loadCSVFile(filepath_in)
	logic.pca_exploration.process()