gistfile1.txt

#!/usr/bin/env python

import sys
import numpy as np
from sklearn.decomposition import PCA


if __name__=="__main__":
	nComponents = 742
	infile = sys.argv[1]
	values = np.array([ [float.fromhex(item) for item in line.split()] for line in open(infile, 'r') if not line[0]=='#'])
	pca = PCA(n_components=nComponents, svd_solver='full')
	print values.shape
	
	reduced = pca.fit_transform(values)
	print reduced.shape
	out = open(sys.argv[2], 'w')
	for row in reduced:
		for i in range(nComponents):
			out.write("%d"%row[i])
			if i==nComponents -1:
				out.write("\n")
			else:
				out.write("\t")
	out.close()
	
	id = np.identity(nComponents)
	zero = pca.inverse_transform(np.zeros(nComponents))
	eigens = pca.inverse_transform(id)
	
	print eigens.shape
	for i in range(nComponents):
		eigenOut = open("v%03d-%s"%(i,sys.argv[3]), 'w')
		vector = eigens[i];
		for j in range(32):
			for k in range(32):
				eigenOut.write("%s\t"%(vector[j*32 + k]-zero[j*32+k]))
			eigenOut.write("\n")
		eigenOut.close()