sentimentorr.py

import pandas as pd
from transformers import AutoTokenizer, AutoModelForSequenceClassification
from torch.nn.functional import softmax as softmax 
import numpy as np
import torch


  
tokenizer = AutoTokenizer.from_pretrained("joeddav/distilbert-base-uncased-go-emotions-student")

model = AutoModelForSequenceClassification.from_pretrained("joeddav/distilbert-base-uncased-go-emotions-student")

labels = ["admiration","amusement","anger","annoyance","approval","caring","confusion","curiosity","desire","disappointment", "disapproval", "disgust", "embarrassment", "excitement", "fear", "gratitude", "grief", "joy", "love", "nervousness", "optimism", "pride", "realization", "relief", "remorse", "sadness", "surprise", "neutral"
]

labels7larg=['admiration',
 'amusement',
 'anger',
 'annoyance',
 'approval',
 'caring',
 'confusion',
 'curiosity',
 'desire',
 'disappointment',
 'disapproval',
 'disgust',
 'embarrassment',
 'excitement',
 'fear',
 'gratitude',
 'grief',
 'joy',
 'love',
 'nervousness',
 'optimism',
 'pride',
 'realization',
 'relief',
 'remorse',
 'sadness',
 'surprise',
 'neutral',
 'larg']
 


def sentimentor(mmm):
    
    dfda = pd.Series(mmm)

    vecs = np.array([model(**tokenizer(txt, return_tensors="pt", padding=True))[0][0].detach().numpy() for txt in dfda])


    
    ds=pd.DataFrame(columns=labels7larg)
    for iii in range(len(vecs)):
        softt=softmax(torch.from_numpy(vecs[iii]), dim=0)
        kki=pd.DataFrame(softt, index=labels)
        ji=kki.nlargest(1, 0)
        dv=pd.DataFrame(columns=labels)

        dv.loc[iii,:]=softt

        dv.loc[iii,'larg']=ji.index[0]

        ds=pd.concat([ds, dv])

    if len(ds)==len(dfda):
        concatt = pd.concat([dfda,ds], axis=1)
        xsd=concatt['larg']
        return xsd[0]
    else:
        print('eshteba kardi')