def get_reddit_posts():
try:
all_message_replies = []
for subreddit in COMMENT_SUBREDDIT_LIST:
message_replies = get_message_replies(subreddit =subreddit, max_replies= 500, submission_count = 300, min_reply_score = 20)
all_message_replies += message_replies
raw_data = list([mr.get_raw_data() for mr in all_message_replies])
write_data_to_cache(raw_data, "raw_data_cache.p")
except Exception:
log.exception("Could not save posts.")
def get_reddit_posts():
try:
all_message_replies = []
for subreddit in COMMENT_SUBREDDIT_LIST:
message_replies = get_message_replies(subreddit=subreddit,
max_replies=500,
submission_count=300,
min_reply_score=20)
all_message_replies += message_replies
raw_data = list([mr.get_raw_data() for mr in all_message_replies])
write_data_to_cache(raw_data, "raw_data_cache.p")
except Exception:
log.exception("Could not save posts.")
return max_ind
def find_input_match(input_text, reply_matrix, match_mat, vectorizer,
reply_text):
match_mat_rows = find_input_match_vec(input_text, match_mat, vectorizer)
max_cos = find_cosine_match(match_mat, match_mat_rows, reply_matrix)
return max_cos
def create_reply_matrix(reply_text, vectorizer):
return vectorizer.transform(reply_text)
if __name__ == '__main__':
message_replies = get_message_replies()
table_data = list(
chain.from_iterable([mr.get_table_data() for mr in message_replies]))
pd_frame = pd.DataFrame(
np.array(table_data),
columns=["message", "message_score", "reply", "reply_score"])
pd_frame['reply_score'] = pd_frame['reply_score'].map(lambda x: int(x))
pd_frame['message_score'] = pd_frame['message_score'].map(lambda x: int(x))
match_mat, vectorizer = create_match_matrix(table_data, pd_frame)
reply_matrix = create_reply_matrix(pd_frame['reply'], vectorizer)
input_set = list(set(pd_frame['message']))
replies = []
match_sum += np.sum(small_mat[z,match_mat_cols])
if match_sum>= max_val:
max_val = match_sum
max_ind = i
return max_ind
def find_input_match(input_text, reply_matrix, match_mat, vectorizer, reply_text):
match_mat_rows = find_input_match_vec(input_text, match_mat, vectorizer)
max_cos = find_cosine_match(match_mat, match_mat_rows, reply_matrix)
return max_cos
def create_reply_matrix(reply_text, vectorizer):
return vectorizer.transform(reply_text)
if __name__ == '__main__':
message_replies = get_message_replies()
table_data = list(chain.from_iterable([mr.get_table_data() for mr in message_replies]))
pd_frame = pd.DataFrame(np.array(table_data),columns=["message", "message_score", "reply", "reply_score"])
pd_frame['reply_score'] = pd_frame['reply_score'].map(lambda x : int(x))
pd_frame['message_score'] = pd_frame['message_score'].map(lambda x : int(x))
match_mat, vectorizer = create_match_matrix(table_data, pd_frame)
reply_matrix = create_reply_matrix(pd_frame['reply'], vectorizer)
input_set = list(set(pd_frame['message']))
replies = []
reply_accurate = []
up_to = len(input_set)
for z in xrange(0,up_to):
