def correct_positive():
"""correct number of positive tweets."""
classify = uva.check50.py.run("tweet.py").module.classify
import helpers
dates, tweets = helpers.read_tweets("trump.txt")
positives = helpers.read_words("positive_words.txt")
negatives = helpers.read_words("negative_words.txt")
with uva.check50.py.capture_stdout() as stdout:
classify(tweets, positives, negatives)
out = stdout.getvalue()
check_classify(out, "positive", 538)
return out
def n_days():
"""correct number of bad days."""
bad_days = uva.check50.py.run("tweet.py").module.bad_days
import helpers
dates, tweets = helpers.read_tweets("trump.txt")
positives = helpers.read_words("positive_words.txt")
negatives = helpers.read_words("negative_words.txt")
with uva.check50.py.capture_stdout() as stdout:
bad_days(dates, tweets, positives, negatives)
out = stdout.getvalue()
matches = re.findall("[\d]+[^\n^\d]*[\d]+[^\n^\d]*[\d]+", out)
if not matches or len(matches) != 31:
n = len(matches) if matches else 0
raise check50.Failure(f"expected 31 bad days, but found {n} dates!")
return out
def best_words():
"""correct top five positive words."""
positive_word = uva.check50.py.run("tweet.py").module.positive_word
import helpers
dates, tweets = helpers.read_tweets("trump.txt")
positives = helpers.read_words("positive_words.txt")
with uva.check50.py.capture_stdout() as stdout:
positive_word(tweets, positives)
out = stdout.getvalue()
top5 = ["great", "trump", "thank", "good", "honor"]
for word in top5:
if word not in out:
raise check50.Mismatch(f"{word}", out)
for word in set(positives) - set(top5) - {"positive", "top"}:
if word and word in out:
raise check50.Failure(f"Did not expect {word} in top 5!")
return out
# detemine sentiment for a single day
def determineDaySentiment(acc, sentiment):
return [
acc[0], acc[1] +
(1 if sentiment[1] > 0 else 0 if sentiment[1] == 0 else -1)
]
# detemine sentiments for all days
daySentiments = [
reduce(determineDaySentiment, list(day[1]), [day[0], 0])
for day in groupedSentiments
]
# print bad days
print("Trump's bad days:")
[print(" ", day[0]) for day in daySentiments if day[1] < 0]
if __name__ == "__main__":
# get the dates and tweets from tweet_filename
dates, tweets = helpers.read_tweets("trump.txt")
# get the lists of negative and positive words
positives = helpers.read_words("positive_words.txt")
negatives = helpers.read_words("negative_words.txt")
classify(tweets, positives, negatives)
positive_word(tweets, positives)
bad_days(dates, tweets, positives, negatives)
import collections
from scipy.sparse import csr_matrix
import numpy as np
import time
import helpers
import config
#np.random.seed(42)
# Read tags
tags, tag2idx, tag_count = helpers.read_tags()
# Read words
words, word2idx, word_count = helpers.read_words()
# Clusters
K = tag_count
# Initialize cluster centers
mu = np.random.rand(K, word_count)
# Get chunks
chunk_reader = helpers.ChunkReader(
post_filename=config.paths.TRAIN_DATA_IDX,
chunk_size=config.data.CHUNK_SIZE) # TODO: Change
chunks = [chunk for chunk in chunk_reader]
#with open(config.paths.TRAIN_DATA_IDX, 'r') as f:
for iteration in range(0, config.algorithm.MAX_ITER):
start = time.time()