def similarity_sub_coef(a, b, a_coef, b_coef):
    sim_jaccard = sm.jaccard_similarity(a, b)
    sim_euclidean = sm.dis_to_sim(sm.euclidean_distance(a_coef, b_coef))
    if(sim_jaccard < 0.5):
	return sim_jaccard
    else:
        return (sm.jaccard_similarity(a, b) + sm.dis_to_sim(sm.euclidean_distance(a_coef, b_coef)))/2
def distance_metrix():
	scores = []
	for i in range(34):
	    score = []
	    for j in range(34):
	    	dist1 = sm.bhatta_distance(id_features[i][1:],id_features[j][1:],12)
	    	dist2 = sm.euclidean_distance(id_features_2[i][1:],id_features_2[j][1:])
	    	score.append((dist1+dist2)/2)
	    scores.append(score)
	return scores
def similarity_sub(a, b):
    return sm.dis_to_sim(sm.euclidean_distance(a, b))
def similarity_sub_coef(a, b, a_coef, b_coef):
#return (sm.jaccard_similarity(a, b))
    return (sm.jaccard_similarity(a, b) + sm.dis_to_sim(sm.euclidean_distance(a_coef, b_coef)))/2