def compute_vector(track, db_conn):
vector = [0] * get_dimensions(db_conn)
for tag, count in track['tags']:
curr, gw = get_vector(db_conn, tag)
if curr is None:
continue
weight = gw * log1p(float(count)) / log(2)
vector = [(weight * x + y) for x, y in zip(curr, vector)]
norm = sqrt(sum([x * x for x in vector]))
if norm > 0:
return tuple([x / norm for x in vector])
else:
return tuple(vector)
def compute_vector(track, db_conn):
vector = [0] * get_dimensions(db_conn)
for tag, count in track['tags']:
curr, gw = get_vector(db_conn, tag)
if curr is None:
continue
weight = gw * log1p(float(count)) / log(2)
vector = [(weight*x + y) for x, y in zip(curr, vector)]
norm = sqrt(sum([x*x for x in vector]))
if norm > 0:
return tuple([x / norm for x in vector])
else:
return tuple(vector)
def tag_features(tag):
return get_vector(conn, tag, normalize=False)
#!/usr/bin/env python
"""Read tag feature vectors from the database."""
import argparse
import math
import sqlite3
import struct
import sys
from libunison.utils import DB_PATH, get_vector, print_vector
def _parse_args():
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument('tag')
parser.add_argument('--db', default=DB_PATH)
return parser.parse_args()
if __name__ == '__main__':
args = _parse_args()
# Convert tag name to UTF-8 and discard case.
tag = unicode(args.tag, encoding='utf-8').lower()
conn = sqlite3.connect(args.db)
vector, weight = get_vector(conn, tag)
if vector is None:
print "Tag not found."
sys.exit(0)
print_vector(vector, weight)
