Update app.py

app.py

@@ -1,12 +1,22 @@
 import streamlit as st
-from transformers import pipeline
+from transformers import pipeline, RobertaTokenizerFast, TFRobertaForSequenceClassification, AutoTokenizer, AutoModelForSequenceClassification
 
 # Sentiment Analysis Pipeline
-pipe = pipeline('sentiment-analysis')
+sentiment_pipe = pipeline('sentiment-analysis')
 
 # Toxicity Classifier
-model_path = "citizenlab/distilbert-base-multilingual-cased-toxicity"
-toxicity_classifier = pipeline("text-classification", model=model_path, tokenizer=model_path)
+model_path_toxic = "citizenlab/distilbert-base-multilingual-cased-toxicity"
+toxicity_classifier = pipeline("text-classification", model=model_path_toxic, tokenizer=model_path_toxic)
+
+# Emotion Analysis
+tokenizer_emotion = RobertaTokenizerFast.from_pretrained("arpanghoshal/EmoRoBERTa")
+model_emotion = TFRobertaForSequenceClassification.from_pretrained("arpanghoshal/EmoRoBERTa")
+emotion = pipeline('sentiment-analysis', model=model_emotion, tokenizer=tokenizer_emotion)
+
+# User Needs Analysis
+tokenizer_needs = AutoTokenizer.from_pretrained("thusken/nb-bert-base-user-needs")
+model_needs = AutoModelForSequenceClassification.from_pretrained("thusken/nb-bert-base-user-needs")
+user_needs = pipeline('text-classification', model=model_needs, tokenizer=tokenizer_needs)
 
 st.title("Plataforma de Diálogos Participativos")
 
@@ -14,31 +24,53 @@ st.title("Plataforma de Diálogos Participativos")
 text = st.text_area("Añade el texto a evaluar")
 
 # Create columns for buttons
-col1, col2 = st.columns(2)
+col1, col2, col3, col4 = st.columns(4)
 
-# Place each button in a separate column to make them appear on the same row
+# Place each button in a separate column
 run_sentiment_analysis = col1.button("Evaluar Sentimiento")
 run_toxicity_analysis = col2.button("Evaluar Toxicidad")
+run_emotion_analysis = col3.button("Evaluar Emoción")
+run_user_needs_analysis = col4.button("Evaluar Necesidades del Usuario")
 
 # Container for output
 output_container = st.container()
 
-# Check if the sentiment analysis button has been pressed and if there's text in the text area
+# Sentiment analysis
 if run_sentiment_analysis and text:
     with output_container:
-        sentiment_output = pipe(text)
-        st.write("Resultado del análisis de sentimiento:")
-        st.json(sentiment_output)
+        sentiment_output = sentiment_pipe(text)
+        label = sentiment_output[0]['label']
+        score = round(sentiment_output[0]['score'] * 100, 2)
+        st.markdown(f"**Resultado del análisis de sentimiento:**\n\n- **Etiqueta:** {label}\n- **Confianza:** {score}%")
 elif run_sentiment_analysis and not text:
     st.warning("Por favor, añade un texto para evaluar el sentimiento.")
 
-# Check if the toxicity analysis button has been pressed and if there's text in the text area
+# Toxicity analysis
 if run_toxicity_analysis and text:
     with output_container:
         toxicity_output = toxicity_classifier(text)
-        st.write("Resultado del análisis de toxicidad:")
-        st.json(toxicity_output)
+        label = toxicity_output[0]['label']
+        score = round(toxicity_output[0]['score'] * 100, 2)
+        st.markdown(f"**Resultado del análisis de toxicidad:**\n\n- **Etiqueta:** {label}\n- **Confianza:** {score}%")
 elif run_toxicity_analysis and not text:
     st.warning("Por favor, añade un texto para evaluar la toxicidad.")
 
+# Emotion analysis
+if run_emotion_analysis and text:
+    with output_container:
+        emotion_output = emotion(text)
+        label = emotion_output[0]['label']
+        score = round(emotion_output[0]['score'] * 100, 2)
+        st.markdown(f"**Resultado del análisis de emoción:**\n\n- **Etiqueta:** {label}\n- **Confianza:** {score}%")
+elif run_emotion_analysis and not text:
+    st.warning("Por favor, añade un texto para evaluar la emoción.")
 
+# User needs analysis
+if run_user_needs_analysis and text:
+    with output_container:
+        needs_output = user_needs(text)
+        label = needs_output[0]['label']
+        score = round(needs_output[0]['score'] * 100, 2)
+        st.markdown(f"**Resultado del análisis de necesidades del usuario:**\n\n- **Etiqueta:** {label}\n- **Confianza:** {score}%")
+elif run_user_needs_analysis and not text:
+    st.warning("Por favor, añade un texto para evaluar las necesidades del usuario.")