A natural language interface for querying datasets using LLMs
Project description
QueryBot
A powerful web application that enables users to upload various data formats, execute SQL queries, and download results through an intuitive interface.
Table of Contents
Overview
QueryBot is built with FastAPI for the backend and a responsive HTML/Bootstrap frontend. It provides a seamless experience for data analysis by allowing users to query multiple data sources through a unified SQL interface.
Architecture
QueryBot follows a modern client-server architecture with real-time data processing capabilities:
flowchart TD
%% Define subgraphs with better spacing
subgraph FE["Frontend Layer"]
direction TB
UI["Web Interface<br/>/static/index.html"]
JS["JavaScript Client<br/>/static/js/script.js"]
UI --- JS
end
subgraph BE["Application Layer"]
direction TB
API["FastAPI Server<br/>/app.py"]
DuckDB["DuckDB Engine<br/>In-Memory Processing"]
LLM["LLM Service<br/>Query Generation"]
API --- DuckDB
API --- LLM
end
subgraph DS["Data Layer"]
direction TB
Local["Local Storage<br/>CSV | Parquet | Excel"]
Remote["Cloud Storage<br/>S3 | Remote URLs"]
DB["Databases<br/>MySQL | SQLite"]
end
%% Define interactions with better labels
FE --> |"[A] File Upload Request"| BE
FE --> |"[B] Query Execution"| BE
FE --> |"[C] Chart Generation"| BE
BE --> |"Schema Inference"| DS
BE --> |"Data Retrieval"| DS
%% Return flow
DS --> |"Raw Data"| BE
BE --> |"Processed Results"| FE
%% Styling
classDef default font-family:arial,stroke-width:2px
classDef layer fill:none,stroke-width:2px
class FE,BE,DS layer
Workflow
-
Data Ingestion
- Users can upload local files or provide paths to remote data sources
- Supported formats include CSV, Parquet, Excel, SQLite, and MySQL databases
- S3 bucket integration for cloud storage access
- Smart file format detection and schema inference
-
Query Processing
- Natural language queries are converted to SQL using LLM services
- Multiple LLM models supported (GPT-4.1-nano, GPT-4.1-mini, o4-mini, Gemini 2.5 Flash)
- Customizable system prompts for specialized workflows
- Automatic handling of complex data types and date formats
-
Data Processing
- DuckDB engine for high-performance query execution
- Zero-copy data access using DuckDB's native read functions
- Direct file reading without temporary tables
- Unified approach for all file types (CSV, Parquet, Excel, SQLite, MySQL)
- Automatic column name handling and type inference
- Support for complex SQL operations including joins and aggregations
-
Visualization
- Dynamic chart generation using Chart.js
- Context-aware visualization recommendations
- Interactive data exploration capabilities
- Export functionality for further analysis
Features
-
Enterprise Data Integration:
- Native support for CSV, Parquet, SQLite, Excel, JSON, and DuckDB files
- Zero-copy data access using DuckDB's native read functions
- Direct file reading without temporary tables
- Smart file format detection and handling
- Automatic schema inference with type validation
- Multi-dataset querying with automatic join optimization
- Direct DuckDB access for improved performance
-
AI-Powered Query Processing:
- Natural language to SQL conversion using state-of-the-art LLMs
- Multi-model support with cost-optimized options:
- GPT-4.1-nano
- GPT-4.1-mini
- o4-mini
- Gemini 2.5 Flash
- Context-aware query generation with dataset understanding
- Intelligent column name handling with special character support
- Customizable system prompts for specialized workflows
- User-editable system prompts for fine-tuned control
-
Advanced Analytics Engine:
- Complex temporal analysis with multi-format date handling
- Advanced text analytics with case-insensitive pattern matching
- Window functions and complex aggregations
- Custom analytical workflows through system prompts
- Automatic query optimization and execution planning
- Enhanced date format handling
-
Interactive Data Visualization:
- Dynamic chart generation with Chart.js integration
- Context-aware visualization recommendations
- Custom visualization descriptions
- Real-time data exploration
- Export capabilities with CSV formatting
-
Enterprise-Grade Architecture:
- Query history persistence with result caching
- Dataset-specific query suggestions
- Comprehensive error handling and logging
- Environment-based configuration management
- Scalable FastAPI architecture with DuckDB backend
- Custom JSON encoder for complex data types
- Secure API key management
-
User Experience:
- Responsive design for desktop and mobile
- Dark theme support
- Recent files history
- Recent queries per file
- Toast notifications for operations
- Loading indicators
- Error handling with detailed messages
-
Development Features:
- FastAPI backend
- DuckDB for efficient data processing
- Custom JSON encoder for complex data types
- Environment variable configuration
- Extensible architecture
Getting Started
Installation
Note: Installation is optional as
uvxwill handle the necessary setup.
pip install querybot
Running the Application
- Start the application with a single command using
uv:uvx querybot - Open your web browser and navigate to http://localhost:8001 to access the QueryBot interface.
Usage
-
Upload Data: Use the interface to specify data sources by:
- Entering file paths (multiple paths can be separated by commas, without quotes)
- Uploading files directly through the browser
-
Supported Data Formats:
- CSV files (
.csv) - Parquet files (
.parquet) - SQLite databases (
.db) - Excel spreadsheets (
.xlsx) - External MySQL databases (from relational-data.org)
- CSV files (
-
Execute Queries: Write SQL queries in the editor and click "Run Query" to see results.
-
Export Results: Download query results as CSV files for further analysis or reporting.
Project Structure
/querybot
│
├── querybot # Main package directory
│ ├── app.py # FastAPI application entry point
│ ├── __init__.py # Package initialization
│ ├── static # Static assets
│ │ ├── index.html # Main frontend interface
│ │ └── js # JavaScript resources
│ │ └── script.js # Frontend functionality
├── pyproject.toml # Project metadata and dependencies
├── .gitignore # Git ignore configuration
├── uv.lock # Dependency lock file
├── README.md # Project documentation
Development
Prerequisites
- Python 3.8+
Setting Up Development Environment
-
Clone the repository:
git clone https://github.com/gramener/querybot.git cd querybot
-
Install dependencies:
pip install -e ".[dev]"
Deployment
Modify the pyproject.toml file to change the version number.
uv build
uv publish
This is deployed to pypi as s.anand@gramener.com
License
This project is licensed under the MIT License - see the LICENSE file for details.
Additional Resources
Query Result
Contribution Guidelines
We welcome contributions! Please see our CONTRIBUTING.md for more details.
FAQ
Q: What data formats are supported? A: The application supports CSV, Parquet, SQLite, Excel, and MySQL formats, with direct DuckDB integration and S3 storage support.
Q: How do I report an issue? A: Please use the Issue Tracker to report any issues you encounter.
Q: Can I customize the AI's behavior? A: Yes, you can edit the system prompt to customize how the AI interprets and processes your queries.
Q: Does QueryBot support cloud storage? A: Yes, QueryBot supports accessing files directly from S3 buckets.
Q: How does QueryBot handle different date formats? A: QueryBot includes enhanced date format handling that can automatically recognize and process various date formats in your data.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file querybot-0.3.3.tar.gz.
File metadata
- Download URL: querybot-0.3.3.tar.gz
- Upload date:
- Size: 26.6 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: uv/0.6.5
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
e5eccd7d60dc2ce05e64a95d6df167119a3e6e36d50b504f42df1c76fa35fa49
|
|
| MD5 |
bea053c973196bf2893ac392a1d0ce0d
|
|
| BLAKE2b-256 |
96675154e4ea6f4b2ac0d510271d7d10fa1a284a3fdda33462b14818be1d36cf
|
File details
Details for the file querybot-0.3.3-py3-none-any.whl.
File metadata
- Download URL: querybot-0.3.3-py3-none-any.whl
- Upload date:
- Size: 25.2 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: uv/0.6.5
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
a93a4860bdf5750d2068f676f7304cba05b220f8d1f254ff9b08fbf683e015e3
|
|
| MD5 |
7a613f8cc5c16132ce3fa47115c0a3cd
|
|
| BLAKE2b-256 |
f7b61863310b320fae63b8f4d3b2cc8536a08b5da1d251b254556d9a318c5d68
|