def test_cosine_similarity(self):
# Assert document cosine similarity.
v1 = self.model.similarity(self.model[0], self.model[1])
v2 = self.model.similarity(self.model[0], self.model[2])
v3 = self.model.similarity(self.model[0], vector.Document("cats cats"))
self.assertAlmostEqual(v1, 0.20, places=2)
self.assertAlmostEqual(v2, 0.00, places=2)
self.assertAlmostEqual(v3, 0.45, places=2)
# Assert that Model.similarity() is aware of LSA reduction.
self.model.reduce(2)
v1 = self.model.similarity(self.model[0], self.model[1])
v2 = self.model.similarity(self.model[0], self.model[2])
self.assertAlmostEqual(v1, 1.00, places=2)
self.assertAlmostEqual(v2, 0.00, places=2)
self.model.lsa = None
print("pattern.vector.Model.similarity()")
def test_information_gain(self):
# Assert information gain weights.
# Example from http://www.comp.lancs.ac.uk/~kc/Lecturing/csc355/DecisionTrees_given.pdf
v = vector.Corpus([
vector.Document({"wind": 1}, type=False),
vector.Document({"wind": 0}, type=True),
vector.Document({"wind": 0}, type=True),
vector.Document({"wind": 0}, type=True),
vector.Document({"wind": 1}, type=True),
vector.Document({"wind": 1}, type=False),
vector.Document({"wind": 1}, type=False)
])
self.assertAlmostEqual(v.information_gain("wind"), 0.52, places=2)
print "patten.vector.Corpus.information_gain()"
def get_all_artists_all_words_to_file():
f = open(common_words_file, "w")
docs = ""
for dir in os.listdir(basedir):
print "artist" + dir
if dir != '.git':
# push the name of the artist onto the exclude words list
name1 = re.split('-', dir)
for n in name1:
rap_exclude_words.append(n)
docs += get_artist_docs(dir)
corpus = vec.Document(docs, exclude=rap_exclude_words, stop_words=False)
for ln in corpus.keywords(top=20):
f.write("%0.08f\t%s" % ln)
print "%0.08f\t%s" % ln
f.close()
def test_cosine_similarity(self):
# Assert document cosine similarity.
v1 = self.model.similarity(self.model[0], self.model[1])
v2 = self.model.similarity(self.model[0], self.model[2])
v3 = self.model.similarity(self.model[0], vector.Document("cats cats"))
self.assertAlmostEqual(v1, 0.20, places=2)
self.assertAlmostEqual(v2, 0.00, places=2)
self.assertAlmostEqual(v3, 0.45, places=2)
# Assert that Model.similarity() is aware of LSA reduction.
try:
import numpy
self.model.reduce(2)
v1 = self.model.similarity(self.model[0], self.model[1])
v2 = self.model.similarity(self.model[0], self.model[2])
self.assertAlmostEqual(v1, 1.00, places=2)
self.assertAlmostEqual(v2, 0.00, places=2)
self.model.lsa = None
except ImportError, e:
pass
def test_document_vector(self):
# Assert Vector properties.
# Test copy.
v = vector.Document("the cat sat on the mat").vector
v = v.copy()
# Test properties.
self.assertTrue(isinstance(v, dict))
self.assertTrue(isinstance(v, vector.Vector))
self.assertTrue(isinstance(v.id, int))
self.assertEqual(sorted(v.features), ["cat", "mat", "sat"])
self.assertEqual(v.weight, vector.TF)
self.assertAlmostEqual(v.norm, 0.58, places=2)
self.assertAlmostEqual(v["cat"], 0.33, places=2)
self.assertAlmostEqual(v["sat"], 0.33, places=2)
self.assertAlmostEqual(v["mat"], 0.33, places=2)
# Test copy + update.
v = v({"cat": 1, "sat": 1, "mat": 1})
self.assertEqual(sorted(v.features), ["cat", "mat", "sat"])
self.assertAlmostEqual(v["cat"], 1.00, places=2)
self.assertAlmostEqual(v["sat"], 1.00, places=2)
self.assertAlmostEqual(v["mat"], 1.00, places=2)
print "pattern.vector.Document.vector"
def test_classifier_vector(self):
# Assert Classifier._vector() (translates input from train() and classify() to a Vector).
v = vector.Classifier()._vector
self.assertEqual(("cat", {
"cat": 0.5,
"purs": 0.5
}), v(vector.Document("the cat purs", type="cat")))
self.assertEqual(("cat", {
"cat": 0.5,
"purs": 0.5
}), v({
"cat": 0.5,
"purs": 0.5
}, type="cat"))
self.assertEqual(("cat", {
"cat": 0.5,
"purs": 0.5
}), v(["cat", "purs"], type="cat"))
self.assertEqual(("cat", {
"cat": 0.5,
"purs": 0.5
}), v("cat purs", type="cat"))
print "pattern.Classifier._vector()"
def test_information_gain(self):
# Assert information gain weights.
# Example from
# http://www.comp.lancs.ac.uk/~kc/Lecturing/csc355/DecisionTrees_given.pdf
m = vector.Model([
vector.Document({"wind": 1}, type=False),
vector.Document({"wind": 0}, type=True),
vector.Document({"wind": 0}, type=True),
vector.Document({"wind": 0}, type=True),
vector.Document({"wind": 1}, type=True),
vector.Document({"wind": 1}, type=False),
vector.Document({"wind": 1}, type=False)], weight=None
)
self.assertAlmostEqual(m.information_gain("wind"), 0.52, places=2)
# Example from http://rutcor.rutgers.edu/~amai/aimath02/PAPERS/14.pdf
m = vector.Model([
vector.Document({"3": 1}, type=True),
vector.Document({"3": 5}, type=True),
vector.Document({"3": 1}, type=False),
vector.Document({"3": 7}, type=True),
vector.Document({"3": 2}, type=False),
vector.Document({"3": 2}, type=True),
vector.Document({"3": 6}, type=False),
vector.Document({"3": 4}, type=True),
vector.Document({"3": 0}, type=False),
vector.Document({"3": 9}, type=True)], weight=None
)
self.assertAlmostEqual(m.ig("3"), 0.571, places=3)
self.assertAlmostEqual(m.gr("3"), 0.195, places=3)
print("patten.vector.Model.information_gain()")
print("patten.vector.Model.gain_ratio()")
def lsa_apply(df):
print("Building model")
m = pv.Model([pv.Document(a) for a in df['abstract']], weight=pv.TFIDF)
print("Returning reduction")
return m.reduce(2)
def get_lsa(texts):
docs = [pv.Document(a) for a in texts]
model = pv.Model(docs, weight=pv.TFIDF)
lsa = model.reduce(2)
return lsa
def lsa_apply(df):
m = pv.Model([pv.Document(a) for a in df['abstract']], weight=pv.TFIDF)
return m.reduce(2)
def create_models(group):
docs = [pv.Document(item, threshold=1) for item in group]
return pv.Model(docs, weight=pv.TFIDF)
def get():
res = es.search(index="posts", body={
"query": {
"match_all": {}
},
'size': 10000
})
# sid = res['_scroll_id']
# scroll_size = len(res['hits']['hits'])
things = {}
boat_docs = {}
c = collections.Counter()
d = collections.Counter()
for item in res['hits']['hits']:
raw_doc = item['_source']['source']
original_boat_name = item['_source']['boat']
# for boat in boats:
# if raw_doc.find(boat) > -1:
t = TextBlob(raw_doc)
d = vector.Document(raw_doc, threshold=1, stopwords=False)
things.setdefault(original_boat_name, collections.Counter())
things[original_boat_name].update([i[1] for i in d.keywords(top=10)])
# boat_docs.setdefault(original_boat_name, collections.Counter())
# for i in t.ngrams(3):
# boat_docs[original_boat_name].update([' '.join(i)])
# if t.sentiment.polarity > 0:
# print(t.sentiment)
# print(raw_doc)
# doc = nlp(raw_doc)
# for chunk in doc.noun_chunks:
# things.setdefault(original_boat_name, collections.Counter())
# things[original_boat_name].update(chunk)
# for sent in doc.sents:
# print(TextBlob(str(sent)).sentiment.polarity, sent)
# print(raw_doc)
# print('-'*40)
# print(raw_doc)
final = {}
filterwords = ['boat', 'boats', 'sail', 'sailing', 'template']
for boat_name, ist in things.items():
print(boat_name)
final.setdefault(boat_name, [])
commons = ist.most_common()
for i in commons:
if i[0] not in filterwords and len(i[0]) > 1 and boat_name.lower().find(i[0]) == -1 and i[1] > 1 and not parse_int(i[0]):
final[boat_name].append({
'word': i[0],
'count': i[1],
})
for name, common in final.items():
post_data = {
'body': {
'keywords': common,
'boat': name
},
"index": "boats",
'doc_type': 'boat'
}
res = es.index(**post_data)
