def nn(model,
text,
vectors,
query,
loaded_custom_model,
k=5): #$ Added custom model parameter
"""
Return the nearest neighbour sentences to query
text: list of sentences
vectors: the corresponding representations for text
query: a string to search
"""
if loaded_custom_model: #$
qf = penseur_utils.encode(model, [query], verbose=False) #$
else: #$
qf = encode(model, [query], verbose=False)
qf /= norm(qf)
scores = numpy.dot(qf, vectors.T).flatten()
sorted_args = numpy.argsort(scores)[::-1]
sentences = [text[a] for a in sorted_args[:k]]
# print 'QUERY: ' + query #$
# print 'NEAREST: ' #$
sorted_sentences = [] #$
# for i, s in enumerate(sentences): #$
# print s, sorted_args[i] #$
for i in xrange(len(sentences)): #$
sorted_sentences.append(sentences[i]) #$
return sorted_sentences #$
def encode(self, sentences, verbose=False): self.sentences = sentences if self.loaded_custom_model: self.vectors = penseur_utils.encode(self.model, sentences, verbose) else: self.vectors = skipthoughts.encode(self.model, sentences, verbose) return self.vectors
def vector(model, text, vectors, query, loaded_custom_model): #$
if loaded_custom_model: #$
qf = penseur_utils.encode(model, [query], verbose=False) #$
else: #$
qf = encode(model, [query], verbose=False) #$
return qf / norm(qf) #$
def encode_single_sentence(self, text, verbose = False): if self.loaded_custom_model: vec = penseur_utils.encode(self.model, [text], verbose) else: vec = skipthoughts.encode(self.model, [text], verbose) return vec
def encode(self, sentences):
self.sentences = sentences
if self.loaded_custom_model:
self.vectors = penseur_utils.encode(self.model, sentences)
else:
self.vectors = skipthoughts.encode(self.model, sentences)