from app.config import config
from tweepy.error import TweepError
from mongoengine.errors import NotUniqueError, OperationError
from pymongo.errors import DuplicateKeyError
import random
# Logging
from app.logger import logger
logger = logger(__name__)
# Load the classifier and markov.
# Loaded here so we can keep it in memory.
# accessible via app.brain.CLS or app.brain.MKV
CLS = Classifier()
MKV = Markov(ramble=config().ramble,
ngram_size=config().ngram_size,
spasm=config().spasm)
def ponder():
"""
Fetch tweets from the Muses
and memorize them;
i.e. train classifier or Markov on them.
"""
logger.info('Pondering new twitter data...')
# Each of these are just a list
# of tweets as strings.
def setUp(self):
self.clf = Classifier(filepath=test_filepath)
def setUp(self):
self.clf = Classifier(filepath=test_filepath)
class ClassifierTest(unittest.TestCase):
def setUp(self):
self.clf = Classifier(filepath=test_filepath)
def tearDown(self):
# Clean up test pickle.
try:
os.remove(test_filepath)
except FileNotFoundError:
pass
def test_train(self):
self.clf.train(docs, labels, save=False)
self.assertEqual(list(self.clf.clf.class_count_), [3.0, 3.0])
def test_online_train(self):
self.clf.train(docs, labels, save=False)
self.clf.train(new_docs, [1, 1, 1], save=False)
self.assertEqual(list(self.clf.clf.class_count_), [3.0, 6.0])
def test_classify(self):
self.clf.train(docs, labels, save=False)
probs = self.clf.classify(["foo dog bar"])[0]
self.assertEqual(1, round(probs[1]))
def test_save_and_load(self):
self.clf.train(docs, labels, save=True)
clf = self.clf.load()
self.assertEqual(list(clf.class_count_), [3.0, 3.0])
class ClassifierTest(unittest.TestCase):
def setUp(self):
self.clf = Classifier(filepath=test_filepath)
def tearDown(self):
# Clean up test pickle.
try:
os.remove(test_filepath)
except FileNotFoundError:
pass
def test_train(self):
self.clf.train(docs, labels, save=False)
self.assertEqual(list(self.clf.clf.class_count_), [3.0, 3.0])
def test_online_train(self):
self.clf.train(docs, labels, save=False)
self.clf.train(new_docs, [1, 1, 1], save=False)
self.assertEqual(list(self.clf.clf.class_count_), [3.0, 6.0])
def test_classify(self):
self.clf.train(docs, labels, save=False)
probs = self.clf.classify(['foo dog bar'])[0]
self.assertEqual(1, round(probs[1]))
def test_save_and_load(self):
self.clf.train(docs, labels, save=True)
clf = self.clf.load()
self.assertEqual(list(clf.class_count_), [3.0, 3.0])