app.py

import os
import gradio as gr
from sqlalchemy import text, inspect
from smolagents import tool, CodeAgent, HfApiModel
import pandas as pd
import tempfile
from database import engine, initialize_database

# Ensure the database initializes (won't crash if empty)
initialize_database()

# SQL Execution Tool (FIXED - Defined BEFORE Use)
@tool
def sql_engine(query: str) -> str:
    """
    Executes an SQL SELECT query and returns the results.

    Args:
        query (str): The SQL query string to execute. Only SELECT queries are allowed.

    Returns:
        str: A formatted string containing the query results, or an error message if the query fails.
    """
    try:
        with engine.connect() as con:
            rows = con.execute(text(query)).fetchall()
        if not rows:
            return "No results found."
        return "\n".join([", ".join(map(str, row)) for row in rows])
    except Exception as e:
        return f"Error: {str(e)}"

# Initialize CodeAgent for SQL query generation (Moved Below `sql_engine`)
agent = CodeAgent(
    tools=[sql_engine],
    model=HfApiModel(model_id="Qwen/Qwen2.5-Coder-32B-Instruct"),
)

# Function to execute an uploaded SQL script
def execute_sql_script(file_path):
    """
    Executes an uploaded SQL file to initialize the database.

    Args:
        file_path (str): Path to the SQL file.

    Returns:
        str: Success message or error description.
    """
    try:
        with engine.connect() as con:
            with open(file_path, "r") as f:
                sql_script = f.read()
                con.execute(text(sql_script))
        return "SQL file executed successfully. Database updated."
    except Exception as e:
        return f"Error: {str(e)}"

# Function to get table names dynamically
def get_table_names():
    """
    Returns a list of tables available in the database.

    Returns:
        list: List of table names.
    """
    inspector = inspect(engine)
    return inspector.get_table_names()

# Function to get table schema dynamically
def get_table_schema(table_name):
    """
    Returns a list of column names for a given table.

    Args:
        table_name (str): Name of the table.

    Returns:
        list: List of column names.
    """
    inspector = inspect(engine)
    columns = inspector.get_columns(table_name)
    return [col["name"] for col in columns]

# Function to fetch data dynamically from any table
def get_table_data(table_name):
    """
    Retrieves all rows from a specified table as a Pandas DataFrame.

    Args:
        table_name (str): Name of the table.

    Returns:
        pd.DataFrame: Table data or an error message.
    """
    try:
        with engine.connect() as con:
            result = con.execute(text(f"SELECT * FROM {table_name}"))
            rows = result.fetchall()

        columns = get_table_schema(table_name)

        if not rows:
            return pd.DataFrame(columns=columns)

        return pd.DataFrame(rows, columns=columns)
    except Exception as e:
        return pd.DataFrame({"Error": [str(e)]})

# Function to handle SQL file uploads and execute them
def handle_file_upload(file):
    """
    Handles SQL file upload, executes SQL, and updates database schema.

    Args:
        file (File): Uploaded SQL file.

    Returns:
        tuple: Execution result message and updated table data.
    """
    temp_file_path = tempfile.mkstemp(suffix=".sql")[1]
    with open(temp_file_path, "wb") as temp_file:
        temp_file.write(file.read())
    
    result = execute_sql_script(temp_file_path)
    tables = get_table_names()

    if tables:
        table_data = {table: get_table_data(table) for table in tables}
    else:
        table_data = {"Error": ["No tables found after execution. Ensure your SQL file creates tables."]}

    return result, table_data

# Function to handle natural language to SQL conversion
def query_sql(user_query: str) -> str:
    """
    Converts a user's natural language query into an SQL query.

    Args:
        user_query (str): The question asked by the user.

    Returns:
        str: The results of the executed SQL query.
    """
    tables = get_table_names()
    if not tables:
        return "Error: No tables found. Please upload an SQL file first."

    schema_info = "Available tables and columns:\n"
    
    for table in tables:
        columns = get_table_schema(table)
        schema_info += f"Table '{table}' has columns: {', '.join(columns)}.\n"

    schema_info += "Generate a valid SQL SELECT query using ONLY these column names. DO NOT return anything other than the SQL query itself."

    generated_sql = agent.run(f"{schema_info} Convert this request into SQL: {user_query}")

    if not isinstance(generated_sql, str) or not generated_sql.strip().lower().startswith(("select", "show", "pragma")):
        return "Error: Only SELECT queries are allowed."

    return sql_engine(generated_sql)

# Function to handle query input
def handle_query(user_input: str) -> str:
    """
    Handles user input and returns the SQL query result.

    Args:
        user_input (str): User's natural language query.

    Returns:
        str: The query result or error message.
    """
    return query_sql(user_input)

# Gradio UI
with gr.Blocks() as demo:
    gr.Markdown("## SQL Query Interface")

    with gr.Row():
        user_input = gr.Textbox(label="Ask a question about the data")
        query_output = gr.Textbox(label="Result")

    user_input.change(fn=handle_query, inputs=user_input, outputs=query_output)

    gr.Markdown("## Upload SQL File to Execute")
    file_upload = gr.File(label="Upload SQL File")
    upload_output = gr.Textbox(label="Execution Result")

    # Dynamic table display
    table_output = gr.Dataframe(label="Database Tables (Dynamic)")

    file_upload.change(fn=handle_file_upload, inputs=file_upload, outputs=[upload_output, table_output])

if __name__ == "__main__":
    demo.launch(server_name="0.0.0.0", server_port=7860, share=True)