def execute_waiting():
"""
Looks for jobs not currently being executed in the pipeline and processes one if the pipeline is
currently empty.
"""
from providentia import app
from providentia.classifier import sentiment
# only run analysis if classifier is ready
if sentiment.classify("test") is None:
logging.debug('Classifier not ready, going back to sleep.')
return
# look to process a job
with app.app_context():
if tbl_benchmark.is_job_being_processed():
logging.debug('A job is being processed, going back to sleep')
else:
logging.debug('No job being processed, looking for jobs to execute')
unstarted_jobs = tbl_benchmark.get_unstarted_jobs()
if unstarted_jobs is None:
logging.debug('No jobs available, going back to sleep.')
else:
logging.debug('Found unstarted jobs!')
random.shuffle(unstarted_jobs)
start_job(unstarted_jobs.pop())
def add_review(self, text, stars):
self.review_count += 1
self.stars += stars
if sentiment.classify(text) == "pos":
self.positive_count += 1
else:
self.negative_count += 1
def __init__(self, business_id, text, stars):
self.business_id = business_id
self.total_stars = stars
self.positive_count = 0
self.negative_count = 0
if sentiment.classify(text) == "pos":
self.positive_count = 1
else:
self.negative_count = 1
def add_review(self, text, cool, funny, useful):
self.review_count += 1
# Count number of words in the review without punctuation
tokenizer = RegexpTokenizer(r'\w+')
tokens = tokenizer.tokenize(text)
self.length += len(tokens)
self.cool += cool
self.funny += funny
self.useful += useful
if sentiment.classify(text) == "pos":
self.positive_count += 1
else:
self.negative_count += 1
def classify():
"""Classify the given text"""
data = request.get_json()
text = data['text']
result = {}
from providentia.classifier import sentiment
if sentiment.get_classifier() is None:
return Response({"message": "Classifier not ready yet!"}, status=503)
else:
result['result'] = sentiment.classify(text)
return Response(json.dumps(result),
status=200,
mimetype='application/json')
def add_sentiment(self, text):
if sentiment.classify(text) == "pos":
self.positive_count += 1
else:
self.negative_count += 1