def user_post_similar(users):
post_ids = []
for post in posts_by_users(users):
post_ids.append(post.id)
idf.append(post.id, post.content.encode('U8'))
length = len(post_ids)
similar = np.zeros(length**2, np.float)
similar = similar.reshape(length, length)
cache = {}
for numi, i in enumerate(post_ids):
for numj, j in enumerate(post_ids):
key = '%s-%s'%(i, j)
if key not in cache and '%s-%s'%(j, i) not in cache:
cache[key] = idf.similar(i, j)
else:
key = '%s-%s'%(j, i)
similar[numi][numj] = cache[key]
return post_ids, similar
def user_comment_smilar(users):
authors = []
for user in users:
author = get_author_id_by_user_id(user)
if author:
authors.append(author)
comments = []
post_ids, similar = user_post_similar(users)
for comment in comment_by_users(users):
comments.append([comment.id, comment.content.encode('U8')])
length = len(comments)
comment_rela = np.zeros(length*len(post_ids), np.float)
comment_rela = comment_rela.reshape(length, len(post_ids))
for num, (i, j) in enumerate(comments):
idf.append(i, j)
for x, p in enumerate(post_ids):
comment_rela[num][x] = idf.similar(i, p)
return post_ids, similar, [i[0] for i in comments], comment_rela
