Example of Doc2Vec and Google Cloud Natural Language API

gistfile1.py

from gensim.models.doc2vec import Doc2Vec
from gensim.models.doc2vec import TaggedDocument

from google.cloud import language

def wakachigaki(text):
    client = language.Client()
    document = client.document_from_text(text)
    annotations = document.annotate_text()
    words = []
    for token in annotations.tokens:
        words.append(token.text_content)
    return words

training_docs = [
    TaggedDocument(words=wakachigaki(u"This is a pen."), tags=['d1']),
    TaggedDocument(words=wakachigaki(u"This is an apple."), tags=['d2']),
    TaggedDocument(words=wakachigaki(u"This is a pineapple."), tags=['d3']),
    TaggedDocument(words=wakachigaki(u"That is a pen."), tags=['d4']),
]

model = Doc2Vec(documents=training_docs, min_count=1, dm=0)

print(model.docvecs['d1'])

print(model.docvecs.most_similar('d1'))
print(model.docvecs.most_similar('d2'))
print(model.docvecs.most_similar('d3'))
print(model.docvecs.most_similar('d4'))