import gradio as gr
import openai
import base64
import os
import sqlite3
from datetime import datetime
from dotenv import load_dotenv

load_dotenv()

client = openai.OpenAI(api_key=os.getenv("OPENAI_API_KEY"))

def init_db():
    conn = sqlite3.connect("complaints.db")
    c = conn.cursor()
    c.execute("""
        CREATE TABLE IF NOT EXISTS complaints (
            id INTEGER PRIMARY KEY AUTOINCREMENT,
            complaint_text TEXT,
            image_path TEXT,
            ai_response TEXT,
            created_at TEXT,
            status TEXT DEFAULT 'resolved'
        )
    """)
    conn.commit()
    conn.close()

def save_complaint(complaint_text, ai_response):
    conn = sqlite3.connect("complaints.db")
    c = conn.cursor()
    c.execute("""
        INSERT INTO complaints 
        (complaint_text, image_path, ai_response, created_at, status)
        VALUES (?, ?, ?, ?, ?)
    """, (complaint_text, "uploaded", ai_response, 
          str(datetime.now()), "resolved"))
    conn.commit()
    conn.close()

def encode_image(image_path):
    with open(image_path, "rb") as f:
        return base64.b64encode(f.read()).decode("utf-8")

def analyze_complaint(image, complaint_text):
    if image is None:
        return "Please upload a product image."
    if not complaint_text:
        return "Please describe your complaint."
    try:
        base64_image = encode_image(image)
        response = client.chat.completions.create(
            model="gpt-4o",
            messages=[
                {
                    "role": "system",
                    "content": """You are a helpful customer support agent.
                    Analyze the product image and complaint together.
                    Provide:
                    1. Acknowledgment of the issue
                    2. What you can see in the image
                    3. Recommended solution
                    4. Next steps for the customer"""
                },
                {
                    "role": "user",
                    "content": [
                        {
                            "type": "text",
                            "text": f"Customer complaint: {complaint_text}"
                        },
                        {
                            "type": "image_url",
                            "image_url": {
                                "url": f"data:image/jpeg;base64,{base64_image}"
                            }
                        }
                    ]
                }
            ],
            max_tokens=500
        )
        ai_response = response.choices[0].message.content
        save_complaint(complaint_text, ai_response)
        return f"🤖 AI Response\n\n{ai_response}"
    except Exception as e:
        return f"Error: {str(e)}"

def get_history():
    try:
        conn = sqlite3.connect("complaints.db")
        c = conn.cursor()
        c.execute("""
            SELECT id, complaint_text, ai_response, 
                   created_at, status 
            FROM complaints 
            ORDER BY id DESC LIMIT 5
        """)
        rows = c.fetchall()
        conn.close()
        if not rows:
            return "No complaints yet."
        history = f"Total recent complaints: {len(rows)}\n\n"
        for row in rows:
            history += f"ID: {row[0]}\n"
            history += f"Complaint: {row[1]}\n"
            history += f"Status: {row[4]}\n"
            history += f"Time: {row[3]}\n"
            history += "-" * 40 + "\n"
        return history
    except Exception as e:
        return f"Error: {str(e)}"

init_db()

with gr.Blocks(
    title="AI Customer Support Agent",
    theme=gr.themes.Soft()
) as demo:

    gr.Markdown("""
    # 🤖 AI Customer Support Agent
    ### Upload a product image and describe your complaint
    *Powered by GPT-4V — Multimodal AI*
    """)

    with gr.Row():
        with gr.Column():
            image_input = gr.Image(
                type="filepath",
                label="📸 Upload Product Image"
            )
            complaint_input = gr.Textbox(
                label="📝 Describe Your Complaint",
                placeholder="Example: I received a damaged product.",
                lines=4
            )
            submit_btn = gr.Button(
                "🚀 Analyze Complaint",
                variant="primary"
            )
        with gr.Column():
            response_output = gr.Textbox(
                label="🤖 AI Response",
                lines=12,
                interactive=False
            )

    gr.Markdown("---")

    with gr.Row():
        history_btn = gr.Button("📋 View Complaint History")
        history_output = gr.Textbox(
            label="Complaint History",
            lines=8,
            interactive=False
        )

    submit_btn.click(
        fn=analyze_complaint,
        inputs=[image_input, complaint_input],
        outputs=response_output
    )

    history_btn.click(
        fn=get_history,
        inputs=[],
        outputs=history_output
    )

    gr.Markdown("""
    ### How it works:
    1. Upload a photo of your damaged/wrong product
    2. Describe your complaint in text
    3. AI analyzes BOTH image and text together
    4. Get an instant resolution response
    """)

if __name__ == "__main__":
    demo.launch()