import streamlit as st
import importlib
import os
import json
from datetime import datetime
from dotenv import load_dotenv
from openai import OpenAI

# Load environment variables from .env file
load_dotenv()

# Set page title and layout
st.set_page_config(page_title="Data Science Tutor", layout="wide")

# Hide Streamlit's default page navigation menu
st.markdown("""
<style>
/* Hide the default "Pages" menu in the top-left sidebar */
[data-testid="stSidebarNav"] {
    display: none;
}
</style>
""", unsafe_allow_html=True)

# Sidebar with image above the CRISP-DM Steps
st.sidebar.image(
    "data2.jpeg",  # Replace with your file path or URL
    use_container_width=True
)

# Sidebar navigation
st.sidebar.title("CRISP-DM Steps")
sections = {
    "Main Page": None,
    "1. Business Understanding": "1_Business_understanding",
    "2. Data understanding": "2_Data_understanding",
    "3. Data Preparation": "3_Data_preparation",
    "4. Feature Engineering": "4_Feature_engineering",
    "5. Modeling": "5_Modeling",
    "6. Evaluation": "6_Evaluation",
    "7. Deployment & Testing": "7_Deployment",
    "8. ML, Deep Learning & Transformers": "8_Models"
}

# By default, make the first item (Main Page) selected.
selected_section = st.sidebar.radio("Select a topic:", list(sections.keys()), index=0)

# If the user selects “Main Page,” just show your introduction content.
if sections[selected_section] is None:
    st.title("🚀 Welcome to the Data Science Tutor!")
    st.markdown(
        """
        <div style="color: #2FA4E7; margin-top: 1rem;">
            <h2>About This App</h2>
            <p>
                This application is designed to guide you through the CRISP-DM process
                for data science projects. Each section in the sidebar highlights a
                different step in the process, providing structured lessons, best
                practices, and hands-on examples.
            </p>
            <h3>App Sections</h3>
            <ul>
                <li><strong>1. Business Understanding</strong> – Clarify project objectives, requirements, and success criteria.</li>
                <li><strong>2. Data Understanding</strong> – Explore data sources, structures, and initial insights.</li>
                <li><strong>3. Data Preparation</strong> – Clean, integrate, and transform the data for modeling.</li>
                <li><strong>4. Feature Engineering</strong> – Engineer and select relevant features for better models.</li>
                <li><strong>5. Modeling</strong> – Develop, train, and tune predictive models.</li>
                <li><strong>6. Evaluation</strong> – Assess performance metrics and refine models.</li>
                <li><strong>7. Deployment & Testing</strong> – Deploy models into production environments and validate.</li>
                <li><strong>8. ML, Deep Learning & Transformers</strong> – Delve deeper into advanced methods and architectures.</li>
            </ul>
        </div>
        """,
        unsafe_allow_html=True
    )
else:
    # Otherwise, load the selected module from the pages folder
    module_name = f"pages.{sections[selected_section]}"
    module = importlib.import_module(module_name)
    module.run()


# OpenAI API Section
api_key = os.getenv("OPENAI_API_KEY")
if not api_key:
    st.sidebar.title("Ask AI")
    api_key = st.sidebar.text_input("Enter your OpenAI API Key", type="password")

client = OpenAI()

# Create side menus for toggles from the pages/ list
st.sidebar.title("Focus Areas")
focus_areas = [
   "Compare Models",
    "Delivery Speed vs Costs",
    "Ethical Considerations in AI",
    "Interpretability & Explainability",
    "Business Impact & ROI",
    "Scalability & Performance",
    "Data Privacy & Security"
]
selected_focus_areas = [area for area in focus_areas if st.sidebar.checkbox(area)]

# Main chat section
st.title("Data Science Tutor Chat")
st.image("https://miro.medium.com/v2/resize:fit:100/format:webp/1*NfE0G4nEj4xX7Z_8dSx83g.png")

# Initialize conversation in the session state
if "messages" not in st.session_state:
    st.session_state["messages"] = [
        {"role": "assistant", "content": "How can I assist you with Data Science today?"}
    ]

# Initialize context prompt added state
if "context_prompt_added" not in st.session_state:
    st.session_state["context_prompt_added"] = False

st.write("---")
st.subheader("Chat")

for msg in st.session_state["messages"]:
    st.chat_message(msg["role"]).write(msg["content"])

if prompt := st.chat_input("Enter your question here:"):
    # Add context to the messages if toggles are selected
    focus_context = ""
    if selected_focus_areas:
        focus_context = f"Focus on {', '.join(selected_focus_areas)} in your response."

    # Add prompting instructions
    prompting_instructions = """
    You are a Data Science tutor. Your responses should be focused on Data Science related subjects. 
    If a question is not related to Data Science, respond with: "Sorry, I can only assist with Data Science related questions."
    Ensure that your responses are not harmful, denigratory, or discriminatory for security and privacy reasons.
    """
    # Add context based on the selected section
    section_context = f"The user is currently viewing the {selected_section} section. "

    # If the context prompt hasn't been added yet, build & inject it once;
    # otherwise, just add the user's raw question.
    if not st.session_state["context_prompt_added"]:
        st.session_state["messages"].append({"role": "user", "content": f"{section_context}{prompt}\n{focus_context}\n{prompting_instructions}"})
        st.session_state["context_prompt_added"] = True
    else:
        st.session_state["messages"].append({"role": "user", "content": f"{section_context}{prompt}"})

    # Display the latest user message in the chat
    st.chat_message("user").write(st.session_state["messages"][-1]["content"])

    # Now call GPT-4 with the entire conversation
    completion = client.chat.completions.create(
            model="gpt-4",
            messages=st.session_state["messages"]
    )
    response_text = completion.choices[0].message.content.strip()

    st.session_state["messages"].append({"role": "assistant", "content": response_text})
    st.chat_message("assistant").write(response_text)

    # Log the conversation
    log_entry = {
        "timestamp": datetime.now().isoformat(),
        "user_query": prompt,
        "assistant_response": response_text,
        "focus_areas": selected_focus_areas,
        "selected_section": selected_section
    }
    log_file_path = os.path.join("logs", "conversation_logs.json")
    os.makedirs(os.path.dirname(log_file_path), exist_ok=True)
    if os.path.exists(log_file_path):
        with open(log_file_path, "r") as log_file:
            logs = json.load(log_file)
    else:
        logs = []
    logs.append(log_entry)
    with open(log_file_path, "w") as log_file:
        json.dump(logs, log_file, indent=4)