def dump_csv(self, filename, columns=DEFAULT_CSV_COLUMNS):
"""
Dumps the matching tweets to a CSV file specified by `filename`.
The default columns are ['id_str', 'user.screen_name', 'timestamp', 'text'].
Columns are specified by their path in the tweet dictionary, so that
'user.screen_name' will grab tweet['user']['screen_name']
If `filename` ends with `.gz`, it will be a gzipped file. Else it will be a plaintext (utf8) file.
Gzip typically achieves 3x-5x compression on this type of data, depending on the columns chosen and the structure of the data.
Example:
########
collection.since(one_hour_ago).dump_csv('my_tweets.csv', columns=['timestamp', 'text'])
"""
if filename.endswith('.gz'):
outfile = gzip.open(filename, 'w')
else:
outfile = open(filename, 'w')
try:
writer = UnicodeWriter(outfile)
writer.writerow(columns)
for tweet in self:
writer.writerow(self._make_row(tweet, columns))
finally:
outfile.close()
def it_to_csv(it_of_its, outfile, mode="w"):
"""
Export an iterable of iterables to a CSV. Row x Column.
(eg: [["id", "score"], [1, 0.9], [2, 0.73], ...] will give
id,score
1,0.9
2,0.73
Can set 'mode' for appending ("a"). Default: write ("w")
"""
with open(outfile, mode) as handle:
writer = UnicodeWriter(handle)
for row in it_of_its:
writer.writerow(list(row))
def it_to_csv(it_of_its, outfile, mode="w"):
"""
Export an iterable of iterables to a CSV. Row x Column.
(eg: [["id", "score"], [1, 0.9], [2, 0.73], ...] will give
id,score
1,0.9
2,0.73
Can set 'mode' for appending ("a"). Default: write ("w")
"""
with open(outfile, mode) as handle:
writer = UnicodeWriter(handle)
for row in it_of_its:
writer.writerow(list(row))
def dump_csv(self, filename, columns=DEFAULT_CSV_COLUMNS):
"""
Dumps the matching tweets to a CSV file specified by `filename`.
The default columns are ['id_str', 'user.screen_name', 'timestamp', 'text'].
Columns are specified by their path in the tweet dictionary, so that
'user.screen_name' will grab tweet['user']['screen_name']
If `filename` ends with `.gz`, it will be a gzipped file. Else it will be a plaintext (utf8) file.
Gzip typically achieves 3x-5x compression on this type of data, depending on the columns chosen and the structure of the data.
Example:
########
collection.since(one_hour_ago).dump_csv('my_tweets.csv', columns=['timestamp', 'text'])
"""
if filename.endswith('.gz'):
outfile = gzip.open(filename, 'w')
else:
outfile = open(filename, 'w')
try:
writer = UnicodeWriter(outfile)
writer.writerow(columns)
for tweet in self:
writer.writerow(self._make_row(tweet, columns))
finally:
outfile.close()
text = line[TEXT_INDEX]
clean = clean_text(text)
tweet_texts.append(clean)
tweet_labels.append(label)
sys.stdout.write('+')
sys.stdout.write('\n')
dictionary = defaultdict(lambda: 0)
for text in tweet_texts:
for word in text.split():
dictionary[word] += 1
sys.stdout.write('.')
with open(CLEAN_DATA_FILENAME, 'wb') as outfile:
writer = UnicodeWriter(outfile)
writer.writerow(['label', 'clean_text'])
for label, text in zip(tweet_labels, tweet_texts):
writer.writerow([label, text])
sys.stdout.write('-')
MIN_D = 20
MAX_D = .5 * len(dictionary)
with open(DICTIONARY_COUNT_FILENAME, 'wb') as outfile:
writer = UnicodeWriter(outfile)
writer.writerow(['word', 'count'])
for word in sorted(dictionary.keys()):
if dictionary[word] < MIN_D or dictionary[word] > MAX_D:
continue
writer.writerow([word, dictionary[word]])
text = line[TEXT_INDEX]
clean = clean_text(text)
tweet_texts.append(clean)
tweet_labels.append(label)
sys.stdout.write('+')
sys.stdout.write('\n')
dictionary = defaultdict(lambda: 0)
for text in tweet_texts:
for word in text.split():
dictionary[word] += 1
sys.stdout.write('.')
with open(CLEAN_DATA_FILENAME, 'wb') as outfile:
writer = UnicodeWriter(outfile)
writer.writerow(['label', 'clean_text'])
for label,text in zip(tweet_labels, tweet_texts):
writer.writerow([label, text])
sys.stdout.write('-')
MIN_D = 20
MAX_D = .5*len(dictionary)
with open(DICTIONARY_COUNT_FILENAME, 'wb') as outfile:
writer = UnicodeWriter(outfile)
writer.writerow(['word', 'count'])
for word in sorted(dictionary.keys()):
if dictionary[word] < MIN_D or dictionary[word] > MAX_D:
continue
writer.writerow([word, dictionary[word]])
