notebooks/Text-2-SQL-Mistral-Nemo-12B-SageMaker.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "acc61fcb-e80c-435d-991c-582f7d3179fc",
   "metadata": {},
   "source": [
    "# Text-to-SQL Query Generator Using Mistral Nemo 12B and Amazon SageMaker\n",
    "\n",
    "## Introduction\n",
    "This notebook demonstrates how to build a tool that converts plain English questions into SQL queries. Follow the steps below to see how we use the Mistral Nemo 12B language model, Amazon SageMaker, and other additional AWS services to help with the text-to-SQL conversion.  \n",
    "\n",
    "\n",
    "## Dataset\n",
    "The notebook uses an [Online Retail Transaction Dataset](https://archive.ics.uci.edu/dataset/352/online+retail) from a UK-based online retailer. The dataset includes:\n",
    "\n",
    "- Over 500,000 transactions from 2010-2011\n",
    "- Customer purchase records with product details, quantities and prices\n",
    "- International transactions across multiple countries\n",
    "- Detailed invoice and stock information\n",
    "\n",
    "The dataset is stored in Amazon S3 and made queryable through Amazon Athena's serverless SQL engine for efficient data analysis.\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "928bbebd-f56d-4aa0-bca2-819460263ddf",
   "metadata": {},
   "source": [
    "## Architecture\n"
   ]
  },
  {
   "attachments": {
    "25296473-fbe6-486e-b7eb-df7fa6877236.png": {
     "image/png": "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"
    }
   },
   "cell_type": "markdown",
   "id": "babbae18-0fc0-44bc-8765-d99943527358",
   "metadata": {},
   "source": [
    "![image.png](attachment:25296473-fbe6-486e-b7eb-df7fa6877236.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f543bd50-ccfe-4e1a-aa36-332c5b964188",
   "metadata": {},
   "source": [
    "## AWS Services Used  \n",
    "\n",
    "Here's a list of the main AWS services we'll be using:\n",
    "\n",
    "\n",
    "**[Amazon S3 (Simple Storage Service)](https://aws.amazon.com/s3/)**\n",
    "- Provides highly durable, secure object storage with 99.999999999% durability \n",
    "- Cost-effective with multiple storage tiers\n",
    "- Storage location for Athena query results\n",
    "\n",
    "**[AWS Glue](https://aws.amazon.com/glue/)**\n",
    "- Serverless ETL service for data preparation\n",
    "- Converts raw data into optimized Parquet format\n",
    "- Manages data catalog and schema definitions\n",
    "- Enables seamless data discovery and querying\n",
    "\n",
    "**[Amazon Athena](https://aws.amazon.com/athena/)**\n",
    "- Serverless interactive query service\n",
    "- Analyzes data directly in S3 using standard SQL\n",
    "- Pay-per-query pricing model\n",
    "- Integrates with AWS Glue catalog\n",
    "\n",
    "**[Amazon SageMaker](https://aws.amazon.com/sagemaker/)**\n",
    "- Fully managed ML service for model deployment\n",
    "- Hosts and serves the Mistral Nemo 12B language model\n",
    "- Provides auto-scaling and high availability\n",
    "- Handles the real-time natural language processing requests  \n",
    "\n",
    "## Prerequisites\n",
    "\n",
    "1. **AWS Account Setup**\n",
    "   - Active AWS account with appropriate IAM permissions\n",
    "   - Basic familiarity with AWS console navigation\n",
    "\n",
    "2. **Data Preparation**\n",
    "   - [The dataset](https://archive.ics.uci.edu/dataset/352/online+retail) converted to Apache Parquet format in Amazon S3\n",
    "   - Follow the [AWS Glue ETL guide](https://docs.aws.amazon.com/prescriptive-guidance/latest/patterns/three-aws-glue-etl-job-types-for-converting-data-to-apache-parquet.html) for step-by-step data transformation instructions\n",
    "   - Ensure proper S3 bucket permissions are configured\n",
    "\n",
    "3. **AWS Glue Configuration** \n",
    "   - AWS Glue Data Catalog properly configured and populated\n",
    "   - Athena integration enabled following the [official documentation](https://docs.aws.amazon.com/athena/latest/ug/data-sources-glue.html)\n",
    "   - Verify table metadata and sample queries work as expected\n",
    "\n",
    "4. **Model Deployment**\n",
    "   - Mistral Nemo 12B model successfully deployed on Amazon SageMaker\n",
    "   - Follow this [deployment notebook](https://github.com/aws-samples/mistral-on-aws/blob/main/notebooks/Mistral-Nemo-Instruct-2407_deploy_sample.ipynb) for detailed setup instructions\n",
    "   - Confirm model endpoint is active and responding to inference requests\n",
    "   - Monitor endpoint metrics and costs through CloudWatch\n",
    "\n",
    "Note: Please ensure all prerequisites are completed and tested before proceeding with the main implementation.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c1ab2e35-952c-4663-9dc1-373697a1e823",
   "metadata": {},
   "source": [
    "## Setup Dependencies  \n",
    "`boto3` is the official AWS SDK (Software Development Kit) for Python that provides programmatic access to AWS services and resources. It enables you to:\n",
    "- Interact with AWS services like S3, EC2, SageMaker, Lambda, DynamoDB and more\n",
    "- Create, configure and manage AWS resources \n",
    "- Build AWS-powered applications and automation workflows\n",
    "- Handle authentication, request retries, and error handling\n",
    "\n",
    "\n",
    " The `json` library is a built-in Python module used for encoding and decoding JSON (JavaScript Object Notation) data. It is essential for:\n",
    "- Parsing JSON responses returned by AWS services\n",
    "- Converting Python dictionaries/objects to JSON strings for API requests \n",
    "- Serializing and deserializing data between Python and JSON formats\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "a67af4e7-8db5-4d48-8af1-712e8cb42e82",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import required libraries\n",
    "import boto3\n",
    "import json"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dd2a2ddf-a594-48ed-9105-ad0d7736d2ea",
   "metadata": {},
   "source": [
    "# AWS clients\n",
    "AWS clients are used to interact with different AWS services:\n",
    "- Athena client: Used to query data stored in S3 using SQL\n",
    "- Glue client: Used for ETL (Extract, Transform, Load) jobs and data catalog management  \n",
    "- SageMaker Runtime client: Used to invoke deployed machine learning model endpoints"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "c8afa789-c46b-4cce-a54e-e3c8c41bcbab",
   "metadata": {},
   "outputs": [],
   "source": [
    "athena_client = boto3.client('athena')     # For running SQL queries\n",
    "glue_client = boto3.client('glue')         # For managing data catalogs\n",
    "smr = boto3.client('sagemaker-runtime')    # For running Mistral Nemo 12B on SageMaker"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "36ce764b",
   "metadata": {},
   "source": [
    "# Helper functions for AWS Glue and Athena database operations\n",
    "\n",
    "`get_table_schema`\n",
    "- Takes a database name as input\n",
    "- Uses AWS Glue Data Catalog API to retrieve metadata about all tables in the database\n",
    "- Returns a dictionary containing table names and their column definitions\n",
    "\n",
    "`format_schema_for_prompt`\n",
    "- Takes the schema dictionary from get_table_schema\n",
    "- Formats it into a human-readable string with table names and column details\n",
    "- Used to prepare schema information for prompts and documentation\n",
    "\n",
    "`execute_athena_query`\n",
    "- Takes a SQL query string as input\n",
    "- Executes the query using Amazon Athena service\n",
    "- Stores query results in a specified S3 bucket location\n",
    "- Returns query execution details and results location\n",
    "\n",
    "IMPORTANT: Replace {INSERT-YOUR-S3-BUCKET} with your own S3 bucket name.  \n",
    "The bucket will be used to store Athena query results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "1046ad36-589f-4b28-b432-8c76b672a831",
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_table_schema(database_name):\n",
    "    tables = glue_client.get_tables(DatabaseName=database_name)['TableList']\n",
    "    schema = {}\n",
    "    for table in tables:\n",
    "        table_name = table['Name']\n",
    "        columns = table['StorageDescriptor']['Columns']\n",
    "        schema[table_name] = [f\"{col['Name']} ({col['Type']})\" for col in columns]\n",
    "    return schema\n",
    "\n",
    "def format_schema_for_prompt(schema):\n",
    "    formatted = \"Database Schema:\\n\"\n",
    "    for table, columns in schema.items():\n",
    "        formatted += f\"Table: {table}\\n\"\n",
    "        formatted += \"Columns:\\n\"\n",
    "        for column in columns:\n",
    "            formatted += f\"- {column}\\n\"\n",
    "        formatted += \"\\n\"\n",
    "    return formatted\n",
    "\n",
    "def execute_athena_query(query, database):\n",
    "    response = athena_client.start_query_execution(\n",
    "        QueryString=query,\n",
    "        QueryExecutionContext={\n",
    "            'Database': database\n",
    "        },\n",
    "        ResultConfiguration={\n",
    "            'OutputLocation': 's3://{INSERT-YOUR-S3-BUCKET}/'  # Replace with your S3 bucket\n",
    "        }\n",
    "    )\n",
    "    return response['QueryExecutionId']"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5a846267-390d-4764-b6a3-f7fabeef45ec",
   "metadata": {},
   "source": [
    "# Configuration Variables for Athena Database and Endpoint Connections  \n",
    "\n",
    "`database_name` is the name of the Athena database to query  \n",
    "\n",
    "`endpoint_name` is the name of the SageMaker endpoint for model inference\n",
    "\n",
    "These variables should be updated with your specific configuration values before running the code.\n",
    "\n",
    "IMPORTANT: \n",
    "1. Replace {INSERT-ATHENA-DATABASE-NAME} with your own Athena Database Name.\n",
    "2. Replace {INSERT-YOUR-SAGEMAKER-ENDPOINT-NAME} with your own SageMaker endpoint name.  \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "19fa1b2a-6aca-48a2-b801-ebaff5c4fa81",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Set up variables\n",
    "database_name = \"{INSERT-ATHENA-DATABASE-NAME}\"  # Replace with your Athena database name\n",
    "endpoint_name = \"{INSERT-YOUR-SAGEMAKER-ENDPOINT-NAME}\"  # Replace with your SageMaker endpoint name"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "452c77da",
   "metadata": {},
   "source": [
    "# Get the schema and prepare the prompt\n",
    "This section prepares the prompt for generating SQL queries\n",
    "1. Retrieves the database schema using get_table_schema()\n",
    "2. Formats the schema into a prompt-friendly format\n",
    "3. Takes a user query as input\n",
    "4. Constructs the final prompt that will be sent to the model\n",
    "\n",
    "IMPORTANT: \n",
    "Replace {INSERT-YOUR-QUESTION} with your own question."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "54fcb581-de2a-46da-a215-7a73e9c68f1c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Generated SQL:\n",
      "SELECT description, SUM(quantity) as total_quantity\n",
      "FROM etl_aws_text_to_sql_iad\n",
      "GROUP BY description\n",
      "ORDER BY total_quantity DESC\n",
      "LIMIT 1;\n"
     ]
    }
   ],
   "source": [
    "# Get the schema and prepare the prompt\n",
    "schema = get_table_schema(database_name)\n",
    "schema_prompt = format_schema_for_prompt(schema)\n",
    "\n",
    "user_query = \"{INSERT-YOUR-QUESTION}\" # Enter Your question\n",
    "prompt = f\"\"\"Given the following Athena database schema:\n",
    "\n",
    "{schema_prompt}\n",
    "\n",
    "Generate a SQL query to answer the following question: {user_query}\n",
    "\n",
    "The SQL query should be valid for Athena and use only the tables and columns provided in the schema.\n",
    "Return only the SQL query, without any explanations or additional text.\"\"\"\n",
    "\n",
    "# Invoke SageMaker endpoint\n",
    "body = {\n",
    "    \"inputs\": f\"<s> [INST]{prompt}[/INST]\",\n",
    "    \"parameters\": {\n",
    "        \"max_new_tokens\": 3000,\n",
    "        \"return_full_text\": False,\n",
    "        \"temperature\": 0.1,\n",
    "        \"top_p\": 0.9,\n",
    "        \"top_k\": 50,\n",
    "    }\n",
    "}\n",
    "\n",
    "resp = smr.invoke_endpoint(\n",
    "    EndpointName=endpoint_name,\n",
    "    ContentType='application/json',\n",
    "    Body=json.dumps(body)\n",
    ")\n",
    "\n",
    "# Process the response and generate SQL\n",
    "response_body = json.loads(resp['Body'].read().decode())\n",
    "generated_sql = response_body[0]['generated_text'].strip()\n",
    "\n",
    "print(\"Generated SQL:\")\n",
    "print(generated_sql)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "51723634",
   "metadata": {},
   "source": [
    "# Execute the query against the specified Athena database\n",
    "This function submits a SQL query to Athena and returns:\n",
    "\n",
    "- query_execution_id: A unique identifier for tracking the query execution\n",
    "- If the query fails to execute, an Exception will be raised with error details  \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "9167ceda-5eac-4ec8-91e5-2eca5356f0ba",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Query execution started. Execution ID: 7180bca8-7ddf-4787-8f7d-3ec5fbe62364\n"
     ]
    }
   ],
   "source": [
    "try:\n",
    "    query_execution_id = execute_athena_query(generated_sql, database_name)\n",
    "    print(f\"Query execution started. Execution ID: {query_execution_id}\")\n",
    "except Exception as e:\n",
    "    print(f\"Error executing query: {str(e)}\")\n",
    "    print(\"Please review and modify the generated SQL before execution.\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cc81013a",
   "metadata": {},
   "source": [
    "# Check query execution status\n",
    "Athena queries run asynchronously, meaning they continue executing in the background.\n",
    "You can run this cell repeatedly to monitor the progress of your query.\n",
    "\n",
    "Common statuses include:\n",
    "- QUEUED: Query is waiting to begin execution\n",
    "- RUNNING: Query is actively executing\n",
    "- SUCCEEDED: Query completed successfully\n",
    "- FAILED: Query encountered an error  \n",
    "\n",
    "\n",
    "Note: This cell can be run multiple times to check the status.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "e46fb09f-c500-45c9-ba34-759b381ca8f8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Query status: SUCCEEDED\n"
     ]
    }
   ],
   "source": [
    "def get_query_execution_status(query_execution_id):\n",
    "    response = athena_client.get_query_execution(QueryExecutionId=query_execution_id)\n",
    "    return response['QueryExecution']['Status']['State']\n",
    "status = get_query_execution_status(query_execution_id)\n",
    "print(f\"Query status: {status}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4addbe21",
   "metadata": {},
   "source": [
    "# Get the query results\n",
    "Note: Only run this cell after the query has completed successfully, otherwise you may get timeout errors or incomplete results.\n",
    "\n",
    "Prerequisites:\n",
    "- An Athena query must have been executed and completed successfully\n",
    "- The query execution ID must be available"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "729ef42d-deba-43c6-886b-84468b231e7a",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'Data': [{'VarCharValue': 'description'}, {'VarCharValue': 'total_quantity'}]}\n",
      "{'Data': [{'VarCharValue': 'WORLD WAR 2 GLIDERS ASSTD DESIGNS'}, {'VarCharValue': '53847'}]}\n"
     ]
    }
   ],
   "source": [
    "def get_query_results(query_execution_id):\n",
    "    response = athena_client.get_query_results(QueryExecutionId=query_execution_id)\n",
    "    return response['ResultSet']['Rows']\n",
    "results = get_query_results(query_execution_id)\n",
    "for row in results:\n",
    "    print(row)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "afadafdd",
   "metadata": {},
   "source": [
    "# Explain the Query Results  \n",
    "\n",
    "Let's now use Mistral Nemo 12B to generate an explaination of the query results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "7942b7ca-399a-4cf7-8c3c-f4fbcac41b1c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Query Analysis and Results Explanation\n",
      "=====================================\n",
      "Original Question: What is the top selling product?\n",
      "\n",
      "Generated SQL Query:\n",
      "SELECT description, SUM(quantity) as total_quantity\n",
      "FROM etl_aws_text_to_sql_iad\n",
      "GROUP BY description\n",
      "ORDER BY total_quantity DESC\n",
      "LIMIT 1;\n",
      "\n",
      "Results Interpretation:\n",
      "\n",
      "Found Results:\n",
      "Product: WORLD WAR 2 GLIDERS ASSTD DESIGNS\n",
      "Total Quantity Sold: 53847\n",
      "\n",
      "Note: Results are based on the available data in the Athena database.\n"
     ]
    }
   ],
   "source": [
    "def explain_results(results, user_query, generated_sql):\n",
    "    print(\"Query Analysis and Results Explanation\")\n",
    "    print(\"=====================================\")\n",
    "    print(f\"Original Question: {user_query}\")\n",
    "    print(f\"\\nGenerated SQL Query:\\n{generated_sql}\")\n",
    "    print(\"\\nResults Interpretation:\")\n",
    "    \n",
    "    # Skip the header row and process results\n",
    "    if len(results) > 1:  # Check if we have results beyond the header\n",
    "        header = [field['VarCharValue'] for field in results[0]['Data']]\n",
    "        data = []\n",
    "        for row in results[1:]:  # Skip header row\n",
    "            data.append([field['VarCharValue'] for field in row['Data']])\n",
    "            \n",
    "        print(\"\\nFound Results:\")\n",
    "        for row in data:\n",
    "            # For the specific \"top selling product\" query, use this context otherwise use\n",
    "            if \"description\" in header and \"total_quantity\" in header:\n",
    "                desc_index = header.index(\"description\")\n",
    "                qty_index = header.index(\"total_quantity\")\n",
    "                print(f\"Product: {row[desc_index]}\")\n",
    "                print(f\"Total Quantity Sold: {row[qty_index]}\")\n",
    "            else:\n",
    "                # Generic result display\n",
    "                for i, value in enumerate(row):\n",
    "                    print(f\"{header[i]}: {value}\")\n",
    "    else:\n",
    "        print(\"No results found for this query.\")\n",
    "    \n",
    "    print(\"\\nNote: Results are based on the available data in the Athena database.\")\n",
    "\n",
    "# Example usage:\n",
    "explain_results(results, user_query, generated_sql)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "51f31b74-282a-4fd9-bcf6-3e98af4e0560",
   "metadata": {},
   "source": [
    "## Troubleshooting Tips\n",
    "\n",
    "Common Issues:\n",
    "1. \"Access Denied\" errors\n",
    "   - Check your AWS credentials\n",
    "   - Verify IAM permissions  \n",
    "\n",
    "     \n",
    "2. \"Endpoint not found\" errors\n",
    "   - Ensure your SageMaker endpoint is running\n",
    "   - Check the endpoint name  \n",
    "\n",
    "\n",
    "3. \"Database not found\" errors\n",
    "   - Verify your Athena database name\n",
    "   - Check if tables are properly created  \n",
    "   \n",
    "     "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5db88464",
   "metadata": {},
   "source": [
    "## License\n",
    "This project uses [Mistral Nemo 12B model](https://huggingface.co/nvidia/Mistral-NeMo-12B-Instruct) under the Apache 2.0 license.  \n",
    "The [Online Retail Dataset](https://archive.ics.uci.edu/dataset/352/online+retail) uses the Creative Commons Attribution 4.0 International (CC BY 4.0) license."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.14"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}