def run_model()

in Medical_Text_Analysis_Resources/docker_containers/model_container/model_scripts/hosted_model.py [0:0]

• 14 lines of code
• 2 McCabe index (conditional complexity)

    def run_model(self,input_sentence,the_paragraph):
        #get the embedder
        word_embedding_model = models.Transformer('emilyalsentzer/Bio_ClinicalBERT')
        # Apply mean pooling to get one fixed sized sentence vector
        pooling_model = models.Pooling(word_embedding_model.get_word_embedding_dimension(),
                                   pooling_mode_mean_tokens=True,
                                   pooling_mode_cls_token=False,
                                   pooling_mode_max_tokens=False)
        embedder = SentenceTransformer(modules=[word_embedding_model, pooling_model])
        # create a corpus of every individual sentence within the paragraph
        corpus=self.break_up_by_sentence(the_paragraph)
        corpus_embeddings = embedder.encode(corpus)
        # define the input sentence as the query
        queries = [input_sentence]
        query_embedding = embedder.encode(queries)
        # calculate the distance between the query embedding and the corpus embeddings
        distances = scipy.spatial.distance.cdist(query_embedding, corpus_embeddings, "cosine")[0]
        to_return=self.get_best_n_sentences(corpus,distances)
        return(to_return)