pipeline.py

# from scipy.special import softmax
import tensorflow as tf

class PreTrainedPipeline():
    def __init__(self, path):
        # define the best model TODO
        sequence_input = tf.keras.Input(shape=(300), name='input')
        x = tf.keras.layers.Dense(2048, activation="LeakyReLU")(sequence_input)
        x = tf.keras.layers.Dense(1024, activation="LeakyReLU")(x)
        x = tf.keras.layers.Dense(512, activation="LeakyReLU")(x)
        x = tf.keras.layers.Dense(128, activation="LeakyReLU")(x)
        x = tf.keras.layers.Dense(512, activation="LeakyReLU")(x)
        x = tf.keras.layers.Dense(1024, activation="LeakyReLU")(x)
        x = tf.keras.layers.Dense(2048, activation="LeakyReLU")(x)
        outputs = tf.keras.layers.Dense(300, activation="tanh")(x)

        model = tf.keras.Model(sequence_input, outputs)

        model.compile(optimizer="Adamax", loss="cosine_similarity")

        # model.load_weights("path to model file") TODO

        self.model = model

    def __call__(self, inputs: str):
        return [
            [ # Sample output, call the model here TODO
                {'label': 'POSITIVE', 'score': 0.05},
                {'label': 'NEGATIVE', 'score': 0.03},
                {'label': 'معنی', 'score': 0.92},
                {'label': f'{inputs}', 'score': 0},
            ]
        ]
 
    # def RevDict(sent,flag,model): 
    #     """ 
    #     This function recieves a sentence from the user, and turns back top_10 (for flag=0) or top_100 (for flag=1) predictions. 
    #     the input sentence will be normalized, and stop words will be removed 
    #     """ 
        
    #     normalizer = Normalizer() 
    #     X_Normalized = normalizer.normalize(sent) 
    #     X_Tokens = word_tokenize(X_Normalized) 
    #     stopwords = [normalizer.normalize(x.strip()) for x in codecs.open(r"stopwords.txt",'r','utf-8').readlines()] 
    #     X_Tokens = [t for t in X_Tokens if t not in stopwords] 
    #     preprocessed = [' '.join(X_Tokens)][0] 
    #     sent_ids = sent2id([preprocessed]) 
    #     output=np.array((model.predict(sent_ids.reshape((1,20))).tolist()[0])) 
    #     distances=distance.cdist(output.reshape((1,300)), comparison_matrix, "cosine")[0] 
    #     min_index_100 = distances.argsort()[:100] 
    #     min_index_10 = distances.argsort()[:10] 
    
    #     temp=[] 
    #     if flag == 0: 
    #         for i in range(10): 
    #             temp.append(id2h[str(min_index_10[i])]) 
    #     elif flag == 1: 
    #         for i in range(100): 
    #             temp.append(id2h[str(min_index_100[i])]) 
        
    #     for i in range(len(temp)): 
    #         print(temp[i])

    # def sent2id(sents): 
    #     sents_id=np.zeros((len(sents),20)) 
    #     for j in tqdm(range(len(sents))): 
    #         for i,word in enumerate(sents[j].split()): 
    #             try: 
    #                 sents_id[j,i] = t2id[word] 
    #             except: 
    #                 sents_id[j,i] = t2id['UNK'] 
    #             if i==19: 
    #                 break 
    #     return sents_id