# Importando as dependencias:

import gradio as gr
from transformers import pipeline


# Importando o modelo:

from transformers import AutoProcessor, AutoModelForSeq2SeqLM

processor = AutoProcessor.from_pretrained("Salesforce/blip-image-captioning-base")
model = AutoModelForSeq2SeqLM.from_pretrained("Salesforce/blip-image-captioning-base")

# Carregar o modelo pré-treinado
distilled_student_sentiment_classifier = pipeline(
    model="lxyuan/distilbert-base-multilingual-cased-sentiments-student", 
    return_all_scores=True
)

# Definir a função para a interface do Gradio
def analyze_sentiment(text):
    result = distilled_student_sentiment_classifier(text)
    # Formatar o resultado para exibir de forma amigável
    formatted_result = {item['label']: item['score'] for item in result[0]}
    return formatted_result

# Criar a interface do Gradio
iface = gr.Interface(
    fn=analyze_sentiment, 
    inputs="text", 
    outputs="json",
    title="Análise de Sentimento",
    description="Digite um texto para analisar seu sentimento."
)

# Lançar a interface do Gradio
iface.launch()