app.py

import gradio as gr
import pandas as pd
import numpy as np
import plotly.express as px
import plotly.graph_objects as go
from plotly.subplots import make_subplots
import io
import json
import warnings
import google.generativeai as genai
import os
from contextlib import redirect_stdout

# --- Configuration ---
warnings.filterwarnings('ignore')
CSS = """
/* --- Phoenix UI Custom CSS --- */
/* Stat Card Styling */
.stat-card {
    border-radius: 12px !important;
    padding: 20px !important;
    background: #f7fafc; /* light gray background */
    border: 1px solid #e2e8f0;
    box-shadow: 0 4px 6px -1px rgba(0, 0, 0, 0.1), 0 2px 4px -1px rgba(0, 0, 0, 0.06);
    text-align: center;
}
.stat-card-title { font-size: 16px; font-weight: 500; color: #4a5568; margin-bottom: 8px; }
.stat-card-value { font-size: 32px; font-weight: 700; color: #2d3748; }

/* General Layout & Feel */
.gradio-container { font-family: 'Inter', sans-serif; }
.gr-button { box-shadow: 0 1px 2px 0 rgba(0,0,0,0.05); }

/* Sidebar Styling */
.sidebar {
    background-color: #f9fafb;
    padding: 15px;
    border-right: 1px solid #e5e7eb;
    min-height: 100vh;
}
.sidebar .gr-button {
    width: 100%;
    text-align: left !important;
    background: none !important;
    border: none !important;
    box-shadow: none !important;
    color: #374151 !important;
    font-size: 16px !important;
    padding: 12px 10px !important;
    margin-bottom: 8px !important;
    border-radius: 8px !important;
}
.sidebar .gr-button:hover { background-color: #e5e7eb !important; }
.sidebar .gr-button.selected { background-color: #d1d5db !important; font-weight: 600 !important; }

/* AI Co-pilot Styling */
.code-block { border: 1px solid #e5e7eb; border-radius: 8px; }
.explanation-block { background-color: #f0f9ff; border-left: 4px solid #3b82f6; padding: 12px; }
"""

# --- Helper Functions ---
def safe_exec(code_string: str, local_vars: dict):
    """Safely execute a string of Python code and capture its output."""
    output_buffer = io.StringIO()
    try:
        with redirect_stdout(output_buffer):
            exec(code_string, globals(), local_vars)
        
        stdout = output_buffer.getvalue()
        fig = local_vars.get('fig')
        result_df = local_vars.get('result_df')
        return stdout, fig, result_df, None
    except Exception as e:
        return None, None, None, f"Execution Error: {str(e)}"

# --- Core Data Processing & State Management ---
def load_and_process_file(file_obj, state_dict):
    """Loads a CSV, processes it, and updates the entire UI state."""
    if file_obj is None:
        return state_dict, "Please upload a file.", *[gr.update(visible=False)] * 4
    try:
        df = pd.read_csv(file_obj.name, low_memory=False)
        for col in df.select_dtypes(include=['object']).columns:
            try:
                df[col] = pd.to_datetime(df[col], errors='raise')
            except (ValueError, TypeError):
                continue
        
        metadata = extract_dataset_metadata(df)
        state_dict = {
            'df': df,
            'metadata': metadata,
            'filename': os.path.basename(file_obj.name),
            'dashboard_plots': []
        }
        
        status_msg = f"✅ **{state_dict['filename']}** loaded successfully."
        
        # Update UI elements with new data context
        cockpit_update = gr.update(visible=True)
        deep_dive_update = gr.update(visible=False)
        copilot_update = gr.update(visible=False)
        welcome_update = gr.update(visible=False)
        
        # Stat cards
        rows, cols = metadata['shape']
        quality = metadata['data_quality']
        
        return (state_dict, status_msg, welcome_update, cockpit_update, deep_dive_update, copilot_update,
                gr.update(value=f"{rows:,}"), gr.update(value=cols), gr.update(value=f"{quality}%"), 
                gr.update(value=f"{len(metadata['datetime_cols'])}"),
                gr.update(choices=metadata['columns']), gr.update(choices=metadata['columns']), gr.update(choices=metadata['columns']))
    except Exception as e:
        return state_dict, f"❌ **Error:** {e}", *[gr.update()] * 11

def extract_dataset_metadata(df: pd.DataFrame):
    rows, cols = df.shape
    numeric_cols = df.select_dtypes(include=np.number).columns.tolist()
    categorical_cols = df.select_dtypes(include=['object', 'category']).columns.tolist()
    datetime_cols = df.select_dtypes(include=['datetime64', 'datetime64[ns]']).columns.tolist()
    data_quality = round((df.notna().sum().sum() / (rows * cols)) * 100, 1) if rows * cols > 0 else 0
    return {
        'shape': (rows, cols), 'columns': df.columns.tolist(),
        'numeric_cols': numeric_cols, 'categorical_cols': categorical_cols,
        'datetime_cols': datetime_cols, 'dtypes': df.dtypes.to_string(),
        'head': df.head().to_string(), 'data_quality': data_quality
    }

# --- Page Navigation ---
def switch_page(page_name):
    """Controls visibility of main content pages."""
    if page_name == "cockpit":
        return gr.update(visible=True), gr.update(visible=False), gr.update(visible=False)
    elif page_name == "deep_dive":
        return gr.update(visible=False), gr.update(visible=True), gr.update(visible=False)
    elif page_name == "co-pilot":
        return gr.update(visible=False), gr.update(visible=False), gr.update(visible=True)
    return gr.update(visible=True), gr.update(visible=False), gr.update(visible=False)

# --- Page 1: Data Cockpit ---
def get_ai_suggestions(state_dict, api_key):
    """Generates proactive analytical suggestions from the AI."""
    if not api_key: return "Enter your Gemini API key to get suggestions.", gr.update(visible=False)
    if not state_dict: return "Upload data first.", gr.update(visible=False)
    
    metadata = state_dict['metadata']
    prompt = f"""
    Based on the following dataset metadata, generate 3 to 5 specific, actionable, and interesting analytical questions a user might want to ask. Frame them as questions.
    - **Columns:** {', '.join(metadata['columns'])}
    - **Numeric:** {', '.join(metadata['numeric_cols'])}
    - **Categorical:** {', '.join(metadata['categorical_cols'])}
    - **Datetime:** {', '.join(metadata['datetime_cols'])}
    
    Return ONLY a JSON list of strings. Example: ["What is the trend of sales over time?", "Which category has the highest average price?"]
    """
    try:
        genai.configure(api_key=api_key)
        model = genai.GenerativeModel('gemini-1.5-flash')
        response = model.generate_content(prompt)
        suggestions = json.loads(response.text)
        
        # Create a button for each suggestion
        buttons = [gr.Button(s, variant="secondary", visible=True) for s in suggestions]
        # Pad with hidden buttons to always have 5 outputs
        buttons += [gr.Button(visible=False)] * (5 - len(buttons))

        return gr.update(visible=False), *buttons
    
    except Exception as e:
        return f"Could not generate suggestions: {e}", *[gr.update(visible=False)]*5

def handle_suggestion_click(question_text):
    """When a suggestion button is clicked, switch to the co-pilot page and populate the input."""
    return (
        gr.update(visible=False), # Hide cockpit
        gr.update(visible=False), # Hide deep dive
        gr.update(visible=True),  # Show co-pilot
        question_text # Populate the chat input
    )

# --- Page 2: Deep Dive Dashboard ---
def add_plot_to_dashboard(state_dict, x_col, y_col, plot_type):
    """Generates a plot and adds it to the state-managed dashboard."""
    if not x_col: return state_dict, gr.update()
    
    df = state_dict['df']
    title = f"{plot_type.capitalize()}: {y_col} by {x_col}" if y_col else f"Distribution of {x_col}"
    fig = None
    
    try:
        if plot_type == 'histogram': fig = px.histogram(df, x=x_col, title=title)
        elif plot_type == 'box': fig = px.box(df, x=x_col, y=y_col, title=title)
        elif plot_type == 'scatter': fig = px.scatter(df, x=x_col, y=y_col, title=title, trendline="ols")
        elif plot_type == 'bar': 
            counts = df[x_col].value_counts().nlargest(20)
            fig = px.bar(counts, x=counts.index, y=counts.values, title=f"Top 20 Categories for {x_col}")
            fig.update_xaxes(title=x_col)
        
        if fig:
            state_dict['dashboard_plots'].append(fig)
        
        # Rebuild the accordion with all plots
        accordion_children = [gr.Plot(fig, visible=True) for fig in state_dict['dashboard_plots']]
        return state_dict, gr.Accordion(label="Your Dashboard Plots", children=accordion_children, open=True)
    except Exception as e:
        gr.Warning(f"Plotting Error: {e}")
        return state_dict, gr.update()

def clear_dashboard(state_dict):
    """Clears all plots from the dashboard."""
    state_dict['dashboard_plots'] = []
    return state_dict, gr.Accordion(label="Your Dashboard Plots", children=[])

# --- Page 3: AI Co-pilot ---
def respond_to_chat(user_message, history, state_dict, api_key):
    """Handles the advanced chat interaction with the AI Co-pilot."""
    if not api_key:
        history.append((user_message, "I need a Gemini API key to function. Please provide it in the sidebar."))
        return history, *[gr.update(visible=False)] * 4
    if not state_dict:
        history.append((user_message, "Please upload a dataset first."))
        return history, *[gr.update(visible=False)] * 4

    history.append((user_message, None))

    metadata = state_dict['metadata']
    prompt = f"""
    You are 'Phoenix Co-pilot', an expert AI data analyst. Your goal is to help a user analyze a pandas DataFrame named `df`.
    
    **Instructions:**
    1.  Carefully understand the user's question.
    2.  Formulate a plan (thought process).
    3.  Write Python code to execute that plan.
    4.  The code can use pandas (pd), numpy (np), and plotly.express (px).
    5.  **For plots, assign the figure to a variable `fig` (e.g., `fig = px.histogram(...)`).**
    6.  **For table-like results, assign the final DataFrame to a variable `result_df` (e.g., `result_df = df.describe()`).**
    7.  Do not modify the original `df`. Use `df.copy()` if needed.
    8.  Provide a brief, user-friendly explanation of the result.
    9.  Respond **ONLY** with a single, raw JSON object with keys: "thought", "code", "explanation".

    **DataFrame Metadata:**
    - Columns and dtypes: {metadata['dtypes']}
    - First 5 rows: {metadata['head']}

    **User Question:** "{user_message}"

    **Your JSON Response:**
    """
    
    try:
        genai.configure(api_key=api_key)
        model = genai.GenerativeModel('gemini-1.5-flash')
        response = model.generate_content(prompt)
        
        response_text = response.text.strip().replace("```json", "").replace("```", "")
        response_json = json.loads(response_text)
        
        thought = response_json.get("thought", "Thinking...")
        code_to_run = response_json.get("code", "")
        explanation = response_json.get("explanation", "Here is the result.")
        
        stdout, fig_result, df_result, error = safe_exec(code_to_run, {'df': state_dict['df'], 'px': px, 'pd': pd, 'np': np})

        bot_message = f"🤔 **Thought:** *{thought}*"
        history[-1] = (user_message, bot_message)
        
        # Prepare outputs, making them visible only if they contain content
        output_updates = [gr.update(visible=False, value=None)] * 4 # [explanation, code, plot, table]
        
        if explanation: output_updates[0] = gr.update(visible=True, value=f"**Phoenix Co-pilot:** {explanation}")
        if code_to_run: output_updates[1] = gr.update(visible=True, value=code_to_run)
        if fig_result: output_updates[2] = gr.update(visible=True, value=fig_result)
        if df_result is not None: output_updates[3] = gr.update(visible=True, value=df_result)
        if stdout:
            # Append stdout to explanation if it exists
            new_explanation = (output_updates[0]['value'] if output_updates[0]['visible'] else "") + f"\n\n**Console Output:**\n```\n{stdout}\n```"
            output_updates[0] = gr.update(visible=True, value=new_explanation)
        if error:
            error_explanation = f"**Phoenix Co-pilot:** I encountered an error. Here's the details:\n\n`{error}`"
            output_updates[0] = gr.update(visible=True, value=error_explanation)

        return history, *output_updates

    except Exception as e:
        error_msg = f"A critical error occurred: {e}. The AI may have returned an invalid response. Please try rephrasing your question."
        history[-1] = (user_message, error_msg)
        return history, *[gr.update(visible=False)] * 4

# --- Gradio UI Definition ---
def create_gradio_interface():
    with gr.Blocks(theme=gr.themes.Monochrome(primary_hue="indigo", secondary_hue="blue"), css=CSS, title="Phoenix AI Data Explorer") as demo:
        global_state = gr.State({})

        with gr.Row():
            # --- Sidebar ---
            with gr.Column(scale=1, elem_classes="sidebar"):
                gr.Markdown("# 🚀 Phoenix UI")
                gr.Markdown("AI Data Explorer")
                
                # Navigation buttons
                cockpit_btn = gr.Button("📊 Data Cockpit", elem_classes="selected")
                deep_dive_btn = gr.Button("🔍 Deep Dive Builder")
                copilot_btn = gr.Button("🤖 AI Co-pilot")
                
                gr.Markdown("---")
                file_input = gr.File(label="📁 Upload New CSV", file_types=[".csv"])
                status_output = gr.Markdown("Status: Awaiting data...")
                gr.Markdown("---")
                api_key_input = gr.Textbox(label="🔑 Gemini API Key", type="password", placeholder="Enter key here...")
                suggestion_btn = gr.Button("Get Smart Suggestions", variant="secondary")

            # --- Main Content Area ---
            with gr.Column(scale=4):
                
                # Welcome Page (Visible initially)
                with gr.Column(visible=True) as welcome_page:
                    gr.Markdown("# Welcome to the AI Data Explorer (Phoenix UI)", elem_id="welcome-header")
                    gr.Markdown("Please **upload a CSV file** and **enter your Gemini API key** in the sidebar to begin.")
                    # CORRECTED: Uses a local file, 'workflow.png', which must be in the same directory.
                    gr.Image(value="workflow.png", label="Workflow", show_label=False, show_download_button=False, container=False)

                # Page 1: Data Cockpit (Hidden initially)
                with gr.Column(visible=False) as cockpit_page:
                    gr.Markdown("## 📊 Data Cockpit")
                    with gr.Row():
                        with gr.Column(elem_classes="stat-card"):
                            gr.Markdown("<div class='stat-card-title'>Rows</div>", elem_classes="stat-card-content")
                            rows_stat = gr.Textbox("0", show_label=False, elem_classes="stat-card-value")
                        with gr.Column(elem_classes="stat-card"):
                            gr.Markdown("<div class='stat-card-title'>Columns</div>", elem_classes="stat-card-content")
                            cols_stat = gr.Textbox("0", show_label=False, elem_classes="stat-card-value")
                        with gr.Column(elem_classes="stat-card"):
                            gr.Markdown("<div class='stat-card-title'>Data Quality</div>", elem_classes="stat-card-content")
                            quality_stat = gr.Textbox("0%", show_label=False, elem_classes="stat-card-value")
                        with gr.Column(elem_classes="stat-card"):
                            gr.Markdown("<div class='stat-card-title'>Date/Time Cols</div>", elem_classes="stat-card-content")
                            time_cols_stat = gr.Textbox("0", show_label=False, elem_classes="stat-card-value")

                    suggestion_status = gr.Markdown(visible=True)
                    with gr.Accordion(label="✨ AI Smart Suggestions", open=True):
                         suggestion_buttons = [gr.Button(visible=False) for _ in range(5)]

                # Page 2: Deep Dive Dashboard Builder (Hidden initially)
                with gr.Column(visible=False) as deep_dive_page:
                    gr.Markdown("## 🔍 Deep Dive Dashboard Builder")
                    gr.Markdown("Create a custom dashboard by adding multiple plots to investigate your data.")
                    with gr.Row():
                        plot_type_dd = gr.Dropdown(['histogram', 'bar', 'scatter', 'box'], label="Plot Type", value='histogram')
                        x_col_dd = gr.Dropdown([], label="X-Axis / Column")
                        y_col_dd = gr.Dropdown([], label="Y-Axis (for Scatter/Box)")
                    with gr.Row():
                        add_plot_btn = gr.Button("Add to Dashboard", variant="primary")
                        clear_plots_btn = gr.Button("Clear Dashboard")
                    dashboard_accordion = gr.Accordion(label="Your Dashboard Plots", open=True)

                # Page 3: AI Co-pilot (Hidden initially)
                with gr.Column(visible=False) as copilot_page:
                    gr.Markdown("## 🤖 AI Co-pilot")
                    gr.Markdown("Ask complex questions in natural language. The Co-pilot will write and execute code to find the answer.")
                    chatbot = gr.Chatbot(height=400, label="Conversation with Co-pilot", show_copy_button=True)
                    
                    with gr.Accordion("Co-pilot's Response Details", open=True):
                        copilot_explanation = gr.Markdown(visible=False, elem_classes="explanation-block")
                        copilot_code = gr.Code(language="python", visible=False, label="Executed Python Code", elem_classes="code-block")
                        copilot_plot = gr.Plot(visible=False, label="Generated Visualization")
                        copilot_table = gr.Dataframe(visible=False, label="Generated Table", wrap=True)

                    with gr.Row():
                        chat_input = gr.Textbox(label="Your Question", placeholder="e.g., 'What is the correlation between age and salary?'", scale=4)
                        chat_submit_btn = gr.Button("Submit", variant="primary")

        # --- Event Handlers ---
        pages = [cockpit_page, deep_dive_page, copilot_page]
        nav_buttons = [cockpit_btn, deep_dive_btn, copilot_btn]
        
        for i, btn in enumerate(nav_buttons):
            btn.click(
                lambda i=i: (gr.update(visible=i==0), gr.update(visible=i==1), gr.update(visible=i==2)),
                outputs=pages
            ).then(
                lambda i=i: [gr.update(elem_classes="selected" if j==i else "") for j in range(len(nav_buttons))],
                outputs=nav_buttons
            )

        file_input.upload(
            fn=load_and_process_file,
            inputs=[file_input, global_state],
            outputs=[global_state, status_output, welcome_page, cockpit_page, deep_dive_page, copilot_page,
                     rows_stat, cols_stat, quality_stat, time_cols_stat,
                     x_col_dd, y_col_dd, plot_type_dd]
        ).then(
            lambda: (gr.update(visible=True), gr.update(visible=False), gr.update(visible=False)), outputs=pages
        ).then(
            lambda: (gr.update(elem_classes="selected"), gr.update(elem_classes=""), gr.update(elem_classes="")), outputs=nav_buttons
        )
        
        suggestion_btn.click(
            get_ai_suggestions,
            [global_state, api_key_input],
            [suggestion_status, *suggestion_buttons]
        )
        
        for btn in suggestion_buttons:
            btn.click(
                handle_suggestion_click,
                inputs=[btn],
                outputs=[cockpit_page, deep_dive_page, copilot_page, chat_input]
            ).then(
                lambda: (gr.update(elem_classes=""), gr.update(elem_classes=""), gr.update(elem_classes="selected")),
                outputs=nav_buttons
            )

        add_plot_btn.click(add_plot_to_dashboard, [global_state, x_col_dd, y_col_dd, plot_type_dd], [global_state, dashboard_accordion])
        clear_plots_btn.click(clear_dashboard, [global_state], [global_state, dashboard_accordion])

        chat_submit_btn.click(
            respond_to_chat, 
            [chat_input, chatbot, global_state, api_key_input], 
            [chatbot, copilot_explanation, copilot_code, copilot_plot, copilot_table]
        ).then(lambda: "", outputs=[chat_input])
        chat_input.submit(
            respond_to_chat, 
            [chat_input, chatbot, global_state, api_key_input], 
            [chatbot, copilot_explanation, copilot_code, copilot_plot, copilot_table]
        ).then(lambda: "", outputs=[chat_input])

    return demo

if __name__ == "__main__":
    app = create_gradio_interface()
    app.launch(debug=True)