def combine_freq_patterns(support, normalize, trait="extra"):
if normalize:
output_fp = os.path.join("result", "feature", "all_freq_pat_support%d_norm.csv" % support)
input_dir = os.path.join("result", "feature", "freq_pat", "normalized", "support%d" % support)
else:
output_fp = os.path.join("result", "feature", "all_freq_pat_support%d.csv" % support)
input_dir = os.path.join("result", "feature", "freq_pat", "support%d" % support)
# input_dir = r'result\feature\freq_pat_select\support%s\%s' % (support, trait)
labels = ["uid"]
id_y = get_trait_scores()
ids = sorted(id_y.keys())
all_features = {id: [] for id in ids}
for filename in os.listdir(input_dir):
if not ".csv" in filename:
continue
feature = filename[:-4]
# print feature
labels.append(feature)
fp = os.path.join(input_dir, filename)
id_feature = read_feature(fp)
for id in all_features:
all_features[id].extend(id_feature[id])
# print id_all_features
fw = open(output_fp, "a")
labels.append(trait)
fw.write(",".join(labels) + "\n")
trait_idx = TRAITS.index(trait)
for id in ids:
line = [id]
line.extend(all_features[id])
line.append(id_y[id][trait_idx])
fw.write(",".join(line) + "\n")
fw.close()
def combine_heuristic_fp(fnames, trait="extra"):
output_fp = os.path.join("result", "feature", "combined_all_%s.csv" % trait)
input_dir = os.path.join("result", "feature")
labels = ["uid"]
id_y = get_trait_scores(trait)
ids = sorted(id_y.keys())
all_features = {id: [] for id in ids}
for fname in fnames:
fr = open(os.path.join(input_dir, fname), "rU")
labels.extend(fr.readline().split(",")[1:-1])
id_feature = {}
for line in fr.readlines():
items = line.split(",")
id = items[0]
features = items[1:-1]
all_features[id].extend(features)
labels.append(trait)
fw = open(output_fp, "a")
fw.write(",".join(labels) + "\n")
for id in ids:
line = [id]
line.extend(all_features[id])
line.append(id_y[id])
fw.write(",".join(line) + "\n")
fw.close()
def combine_heuristic_features(feature_names, trait="extra"):
output_fp = os.path.join("result", "feature", "all_heuristic_features_%s.csv" % trait)
input_dir = os.path.join("result", "feature")
labels = ["uid"]
for feature_name in feature_names:
if "-" in feature_name:
labels.extend(feature_name.split("-"))
else:
labels.append(feature_name)
id_y = get_trait_scores(trait)
ids = sorted(id_y.keys())
all_features = {id: [] for id in ids}
for feature_name in feature_names:
fp = os.path.join(input_dir, "%s.csv" % feature_name)
id_feature = read_feature(fp)
for id in all_features:
if id in id_feature:
all_features[id].extend(id_feature[id])
fw = open(output_fp, "a")
labels.append(trait)
fw.write(",".join(labels) + "\n")
for id in ids:
line = [id]
line.extend(all_features[id])
line.append(id_y[id])
fw.write(",".join(line) + "\n")
fw.close()