import streamlit as st
import numpy as np
import pickle
from langchain.schema import HumanMessage, SystemMessage, AIMessage
from langchain_groq import ChatGroq
from dotenv import load_dotenv
import os

# Set page configuration
st.set_page_config(
    page_title="Agricultural AI Assistant and Crop Recommendation",
    layout="wide"
)

# Load environment variables
load_dotenv()
os.environ['GROQ_API_KEY'] = os.getenv("GROQ_API_KEY")
groq_api_key = os.getenv("GROQ_API_KEY")
chat = ChatGroq(groq_api_key=groq_api_key, model_name="llama-3.3-70b-versatile")

# Load the model and scaler
model = pickle.load(open('model.pkl', 'rb'))
ms = pickle.load(open('minmaxscaler.pkl', 'rb'))

# Custom CSS for styling
st.markdown("""
<style>
    body {
        background: #BCBBB8;
    }
    .title {
        text-align: center;
        color: mediumseagreen;
    }
    .warning {
        color: red;
        font-weight: bold;
        text-align: center;
    }
    .container {
        background: #edf2f7;
        font-weight: bold;
        padding: 20px;
        border-radius: 15px;
        margin-top: 20px;
    }
    .stButton>button {
        background-color: #007bff;
        color: white;
        font-size: 16px;
        font-weight: bold;
        border: none;
        border-radius: 5px;
        padding: 10px 20px;
    }
    .stTextInput>div>input {
        border-radius: 5px;
        border: 1px solid #007bff;
        padding: 10px;
    }
</style>
""", unsafe_allow_html=True)

# Initialize session state for chatbot messages
if 'flow_messages' not in st.session_state:
    st.session_state['flow_messages'] = [
        SystemMessage(content="You are a highly intelligent and friendly agricultural assistant. Provide accurate and relevant answers about crops, farming, and agricultural practices.")
    ]

# Define the chatbot response function
def get_response(question):
    st.session_state['flow_messages'].append(HumanMessage(content=question))
    answer = chat(st.session_state['flow_messages'])
    st.session_state['flow_messages'].append(AIMessage(content=answer.content))
    return answer.content

# App features
st.markdown('<h1 class="title">Agricultural AI Assistant 🌾</h1>', unsafe_allow_html=True)
st.sidebar.header("Features")
features = st.sidebar.radio("Choose a feature:", ("Crop Recommendation", "Conversational Q&A"))

if features == "Crop Recommendation":
    st.write("""
    ### Provide the necessary agricultural parameters:
    """)

    # Input fields for the parameters
    N = st.number_input('Nitrogen', min_value=0, max_value=150, step=1)
    P = st.number_input('Phosphorus', min_value=0, max_value=100, step=1)
    K = st.number_input('Potassium', min_value=0, max_value=100, step=1)
    temp = st.number_input('Temperature (°C)', min_value=-10.0, max_value=60.0, step=0.1)
    humidity = st.number_input('Humidity (%)', min_value=0.0, max_value=100.0, step=0.1)
    ph = st.number_input('pH', min_value=0.0, max_value=14.0, step=0.1)
    rainfall = st.number_input('Rainfall (mm)', min_value=0.0, max_value=1000.0, step=1.0)

    # Button to trigger prediction
    if st.button('Get Recommendation'):
        # Feature list and transformation
        feature_list = [N, P, K, temp, humidity, ph, rainfall]
        single_pred = np.array(feature_list).reshape(1, -1)

        # Apply scaling
        scaled_features = ms.transform(single_pred)

        # Make prediction
        prediction = model.predict(scaled_features)

        # Dictionary to map predictions to crop names
        crop_dict = {
            1: "Rice", 2: "Maize", 3: "Jute", 4: "Cotton", 5: "Coconut", 6: "Papaya", 7: "Orange",
            8: "Apple", 9: "Muskmelon", 10: "Watermelon", 11: "Grapes", 12: "Mango", 13: "Banana",
            14: "Pomegranate", 15: "Lentil", 16: "Blackgram", 17: "Mungbean", 18: "Mothbeans",
            19: "Pigeonpeas", 20: "Kidneybeans", 21: "Chickpea", 22: "Coffee"
        }

        # Display the result
        if prediction[0] in crop_dict:
            crop = crop_dict[prediction[0]]
            result = f"**{crop}** is the best crop to be cultivated with the provided data."
            st.success(result)
        else:
            result = "Sorry, we could not determine the best crop to be cultivated with the provided data."
            st.error(result)

elif features == "Conversational Q&A":
    st.write("""
    ### Ask any question about crops, farming, and agriculture:
    """)
    user_input = st.text_input("Your Question:")
    if st.button("Ask Question"):
        if user_input.strip():
            response = get_response(user_input)
            st.subheader("The Response is:")
            st.write(response)
        else:
            st.warning("Please enter a question!")