input_file = "test-full.txt"
output_file = "test-expanded-all-neighb.txt"
# Generate graph from edges file
G = ex.read_graph(c.output_dir + "edges-directed-6.txt")
# Get feature space and target words
features_index = u.read_features_file(c.output_dir + "feat-space.txt")
target_words = u.get_target_words()
# Get sentences/vectors of data to expand
sentences, data = ex.get_expansion_data(input_file, features_index)
# Get matrix of weight vectors
print "generating weight matrix..."
W, b_arr = u.get_weight_matrix(target_words)
print "expanding feature vectors..."
i = 0
for vect in data:
#if i == 5:
# break
# Store array of current feats
feats = [f.split(":")[0] for f in sentences[i][1:]]
# Get prediction values for each feature using weight matrix and bias terms
values = ex.predict_feats(W, b_arr, vect)
# Part 1:
input_file = "test-full-new.txt"
output_file = "test-local-path-1.txt"
# Generate graph from edges file
G = ex.read_graph(c.output_dir + "edges-directed-6.txt")
# Get feature space and target words
features_index = u.read_features_file(c.output_dir + "feat-space.txt")
target_words = u.get_target_words()
# Get sentences/vectors of data to expand
sentences, data = ex.get_expansion_data(input_file, features_index)
# Get matrix of weight vectors
print "generating weight matrix..."
W, b_arr = u.get_weight_matrix(target_words)
# Iterate over instances and expand them
print "expanding feature vectors..."
i = 0
for vect in data:
#if i == 3:
# break
# Store array of current feats
feats = [f.split(":")[0] for f in sentences[i][1:]]
# Get prediction values for each feature using weight matrix and bias terms
values = ex.predict_feats(W, b_arr, vect)
# create dict to hold expanded features
if __name__ == "__main__":
# Get feature vectors
print "generating feature vectors..."
features_index = u.read_features_file("output-700/feat-space.txt")
feature_size = len(features_index)
target_words = u.get_target_words()
# Now let's do SCL
print "getting targets..."
targets = u.get_target_words()
# Load target word weight vectors as rows into matrix W
print "creating W..."
W, _ = u.get_weight_matrix(targets)
# Perform SVD on W
print "performing SVD(W)..."
WT = W.T
# SVD returns U, S, V.T
U, S, VT = numpy.linalg.svd(WT, full_matrices=False)
#print "W"
#print W.shape
#print "WT"
#print WT.shape
#print ""
#print U
#print "U"