import streamlit as st
from transformers import pipeline, AutoModelForCausalLM, AutoTokenizer

# Load chatbot model
chatbot_model = "microsoft/DialoGPT-medium"
tokenizer = AutoTokenizer.from_pretrained(chatbot_model)
model = AutoModelForCausalLM.from_pretrained(chatbot_model)

# Load emotion detection model
emotion_pipeline = pipeline("text-classification", model="j-hartmann/emotion-english-distilroberta-base")

st.title("🧠 Mental Health Chatbot")

# Chat history
if "chat_history" not in st.session_state:
    st.session_state.chat_history = []

# User Input
user_input = st.text_input("You:", key="user_input")

if st.button("Send"):
    if user_input:
        # Generate chatbot response
        input_ids = tokenizer.encode(user_input + tokenizer.eos_token, return_tensors="pt")
        output = model.generate(input_ids, max_length=200, pad_token_id=tokenizer.eos_token_id)
        response = tokenizer.decode(output[:, input_ids.shape[-1]:][0], skip_special_tokens=True)

        # Detect emotion
        emotion_result = emotion_pipeline(user_input)
        emotion = emotion_result[0]["label"]

        # Store chat history
        st.session_state.chat_history.append(("You", user_input))
        st.session_state.chat_history.append(("Bot", response))

        # Display chat
        for sender, msg in st.session_state.chat_history:
            st.write(f"**{sender}:** {msg}")

        # Display emotion
        st.write(f"🧠 **Emotion Detected:** {emotion}")



# import streamlit as st
# from transformers import pipeline, AutoTokenizer

# # ✅ Load Emotion Recognition Model
# emotion_pipeline = pipeline("text-classification", model="ahmettasdemir/distilbert-base-uncased-finetuned-emotion")

# # ✅ Load Stress Detection Model
# stress_pipeline = pipeline("text-classification", model="j-hartmann/emotion-english-distilroberta-base")

# # ✅ Load Mental Disorder Detection Model
# mental_bert_pipeline = pipeline("text-classification", model="nlpconnect/vit-gpt2-image-captioning")

# # ✅ Load PHQ-9 Depression Severity Classifier
# phq9_pipeline = pipeline("text-classification", model="PHQ-9 Depression Classifier")

# # ✅ Load Chatbot Model (DeepSeek)
# deepseek_model = "deepseek-ai/deepseek-llm-7b"
# deepseek_tokenizer = AutoTokenizer.from_pretrained(deepseek_model)
# deepseek_pipeline = pipeline("text-generation", model=deepseek_model, tokenizer=deepseek_tokenizer)

# # 🏥 Streamlit UI
# st.title("🧠 Mental Health Assistant Bot")

# user_input = st.text_input("How are you feeling today?", "")

# if st.button("Submit"):
#     if user_input:
#         # ✅ Emotion Analysis
#         emotion_result = emotion_pipeline(user_input)[0]
#         st.write(f"**Emotion Detected:** {emotion_result['label']} ({emotion_result['score']:.2f})")

#         # ✅ Stress Level Analysis
#         stress_result = stress_pipeline(user_input)[0]
#         st.write(f"**Stress Level:** {stress_result['label']} ({stress_result['score']:.2f})")

#         # ✅ Mental Health Condition Detection
#         mental_health_result = mental_bert_pipeline(user_input)[0]
#         st.write(f"**Possible Mental Health Condition:** {mental_health_result['label']} ({mental_health_result['score']:.2f})")

#         # ✅ AI Chatbot Response
#         deepseek_response = deepseek_pipeline(user_input, max_length=100, do_sample=True)[0]['generated_text']
#         st.write(f"🤖 **Chatbot:** {deepseek_response}")