def train(self, **kwargs):
self.job.training(self.tagset, streammanager(self.interface,self.training_subscription), **kwargs)
def classify(self, iterator=None, sc=None):
if iterator is None:
iterator = streammanager(self.interface,self.input_subscription)
# classified will be able to be used to push output to output_subscription
classified = self.job.run_classifier(iterator, stop_condition = sc)
self.output(classified)
def train(self, **kwargs):
self.job.training(
self.tagset,
streammanager(self.interface, self.training_subscription),
**kwargs)
def classify(self, iterator=None, sc=None):
if iterator is None:
iterator = streammanager(self.interface, self.input_subscription)
# classified will be able to be used to push output to output_subscription
classified = self.job.run_classifier(iterator, stop_condition=sc)
self.output(classified)
#accept file arg
try:
filetype = sys.argv[1]
filename = sys.argv[2]
except:
print("Script run with invalid arguments")
print("Usage: python3 tsnetest.py file_type data_file.tsv")
print("Example Usage: python3 tsnetest.py neel NEEL2016-training.tsv")
sys.exit(2)
text = []
urls = []
#read through message objects, saving text in text array
for message in streammanager(filetype, filename):
text.append(message.text)
urls.append(message.url)
#fit and transform the text into vector form using TF-IDF
vectors = TfidfVectorizer().fit_transform(text)
print(repr(vectors))
#reduce dimensionality to 50 before running tsne
X_reduced = TruncatedSVD(n_components=50, random_state=0).fit_transform(vectors)
#run tsne, convert to two dimensions
X_embedded = mytsne.TSNE(n_components=2, perplexity=40, verbose=2, urls=urls, text=text).fit_transform(X_reduced)
trust = mytsne.trustworthiness(vectors, X_embedded)
print("Trustworthiness: {}".format(trust))
#accept file arg
try:
filetype = sys.argv[1]
filename = sys.argv[2]
except:
print("Script run with invalid arguments")
print("Usage: python3 tsnetest.py file_type data_file.tsv")
print("Example Usage: python3 tsnetest.py neel NEEL2016-training.tsv")
sys.exit(2)
text = []
urls = []
#read through message objects, saving text in text array
for message in streammanager(filetype, filename):
text.append(message.text)
urls.append(message.url)
#fit and transform the text into vector form using TF-IDF
vectors = TfidfVectorizer().fit_transform(text)
print(repr(vectors))
#reduce dimensionality to 50 before running tsne
X_reduced = TruncatedSVD(n_components=50,
random_state=0).fit_transform(vectors)
#run tsne, convert to two dimensions
X_embedded = mytsne.TSNE(n_components=2,
perplexity=40,
verbose=2,
