def test_comments(self):
""" Tests the insertion and deletion of comments """
bat = queries.find_media_by_creator('Dario')
queries.insert_comment(bat[0].media_id, 11, 7, 0.9, 0.4, -0.5, 0.6, 0.3)
comments = queries.find_comments_for_media(bat[0].media_id)
# can we find the comments?
self.assertEqual(comments[0].item_id, bat[0].media_id, \
'Retrieved incorrect id')
self.assertEqual(comments[0].relevancy, 7, \
msg='Retrieved incorrect relevancy')
self.assertAlmostEqual(float(comments[0].pleasantness), 0.9, \
msg='Retrieved incorrect pleasantness')
self.assertAlmostEqual(float(comments[0].attention), 0.4, \
msg='Retrieved incorrect attention')
self.assertAlmostEqual(float(comments[0].sensitivity), -0.5, \
msg='Retrieved incorrect sensitivity')
self.assertAlmostEqual(float(comments[0].aptitude), 0.6, \
msg='Retrieved incorrect aptitude')
self.assertAlmostEqual(float(comments[0].polarity), 0.3, \
msg='Retrieved incorrect polarity')
# is the number of comments incremented
bat = queries.find_media_by_creator('Dario')
self.assertEqual(bat[0].number_of_comments, 1,
'Retrieved incorrect number of comments')
def test_clean(self):
""" Tests cleaning the database of comments and emotions """
bat = queries.find_media_by_asin('0440419395')
queries.insert_comment(bat.media_id, 11, 7, 0.9, 0.4, -0.5, 0.6, 0.3)
queries.clean_media(bat.media_id)
comments = queries.find_comments_for_media(bat.media_id)
emotions = queries.find_emotions_for_media(bat.media_id)
self.assertEqual(len(comments), 0)
self.assertEqual(len(emotions), 0)
def calculateVectorsForAllComments(dictFromJSON, g):
compound_emotion_dict = collections.defaultdict(int)
sentic_emotion_dict = collections.defaultdict(int)
processed_comments = list()
overall_rating = 0.0
# the product model from the DB
product = queries.find_media_by_asin(dictFromJSON["asin"])
tokenized_docs = buildListOfTokenizedDocuments(dictFromJSON)
for comment in dictFromJSON["comments"]:
vectorized_comment = calculateVector(tokenizeDocument(comment["text"]), tokenized_docs)
vectorized_desc = calculateVector(tokenizeDocument(dictFromJSON["description"]), tokenized_docs)
comment["vector_space"] = vectorized_comment
relevancy = getCosine(vectorized_comment, vectorized_desc)
if relevancy < 0.15:
continue
comment["relevancy"] = relevancy
# add emotional score
try:
comment_emotion = emotions(comment["text"], g)
except ConceptError:
print("Not enough concepts to do anything useful - skipping this product")
continue
comment["emotion_vector"] = comment_emotion.emotion_vector
compound_emotions = comment_emotion.get_compound_emotion()
sentic_values = comment_emotion.get_all_sentic_values()
sentic_values = [value.name for value in sentic_values if value is not None]
compound_emotions_list = []
for compound_emotion, strength in compound_emotions:
compound_emotions_list.append(
{"compound_emotion": compound_emotion.name, "strength": strength.name}
)
comment["compound_emotions"] = compound_emotions_list
comment["sentic_emotions"] = sentic_values
# add all compound_emotions to the default dictFromJSON
for compound in comment["compound_emotions"]:
compound_emotion_dict[compound["compound_emotion"]] += 1
for sentic in comment["sentic_emotions"]:
sentic_emotion_dict[sentic] += 1
overall_rating += float(comment["rating"])
# add the comment to the database
# insert_comment(item_id, relevancy, pleasantness, attention, sensitivity, aptitude, polarity, date)
queries.insert_comment(product.media_id,
comment["unixtime"],
comment["relevancy"],
comment["emotion_vector"]["pleasantness"],
comment["emotion_vector"]["attention"],
comment["emotion_vector"]["sensitivity"],
comment["emotion_vector"]["aptitude"],
comment["emotion_vector"]["polarity"])
comment["text"] = html.unescape(comment["text"])
processed_comments.append(comment)
popular_compound_emotions = []
if len(compound_emotion_dict) == 0:
raise NoEmotionsFoundError("No compound emotions found")
for i in range(0, 3):
popular_emotion = max(compound_emotion_dict, key=compound_emotion_dict.get)
popular_compound_emotions.append(popular_emotion)
compound_emotion_dict.pop(popular_emotion)
popular_sentic_emotions = []
if len(sentic_emotion_dict) == 0:
raise NoEmotionsFoundError("No sentic emotions found")
for i in range(0, 3):
popular_sentic = max(sentic_emotion_dict, key=sentic_emotion_dict.get)
popular_sentic_emotions.append(popular_sentic)
sentic_emotion_dict.pop(popular_sentic)
# add emotion to the database
queries.insert_media_emotion(product.media_id, popular_sentic)
dictFromJSON["popular_compound_emotions"] = popular_compound_emotions
dictFromJSON["popular_sentic_emotions"] = popular_sentic_emotions
if len(processed_comments) > 0:
rating = overall_rating / len(processed_comments)
rating = float("{0:.2f}".format(rating))
dictFromJSON["overall_rating"] = rating
dictFromJSON["comments"] = sort_list_of_dicts(processed_comments)
# get rid of escaped html characters in description
dictFromJSON["description"] = html.unescape(dictFromJSON["description"])
return dictFromJSON