Prompt2query

# Prompt2Query

Prompt2Query is a desktop application built using tkinter that enables users to input natural language prompts, generate corresponding code, and execute it using the llm_pandas library. This tool is specifically designed for interacting with data using Python code, providing an intuitive interface to query datasets with natural language.

Flow chart

%%{init: {'theme': 'black', 'themeVariables': { 'fontSize': '16px'}, "securityLevel": "loose"}}%%
graph TD
    A[User Input] --> B{Mode Selection}
    
    B -->|Pandas| C[Pandas Mode]
    C --> C1[LLMHandler]
    C1 --> C2[Generate Code]
    C2 --> C3[Execute Code]
    C3 --> C4[Generate Natural Language]
    C4 --> R[Display Results]
    
    B -->|SQL| D[SQL Mode]
    D --> D1[LLMMySQLHandler]
    D1 --> D2[Generate SQL]
    D2 --> D3[Execute SQL]
    D3 --> D4[Format Results]
    D4 --> R
    
    B -->|REPandas| E[Recursive Mode]
    E --> E1[Generate Plan]
    E1 --> E2[Execute Phases]
    E2 --> E3{Phase Complete?}
    E3 -->|No| E2
    E3 -->|Yes| R
    
    B -->|Default| F[Default Mode]
    F --> F1[LLM Processing]
    F1 --> R
    
    G[File Operations] --> G1{Action Type}
    G1 -->|Load| G2[Load CSV]
    G1 -->|Save| G3[Save Results]
    G2 --> C1
    G3 --> H[File System]
    
    I[History Management] --> I1[Track Selected Items]
    I1 --> I2[Update Context]
    I2 --> C1
    I2 --> D1

Loading

Features

• Natural Language to Code: Convert user queries written in natural language to executable code using llm_pandas.
• Multiple Modes: Currently supports both Pandas and SQL modes for generating queries.
• File Attachment: Load CSV files to query data.
• Code Execution: Execute the generated code directly within the application.
• Scrollable Result Logs: View query results and logs in a clean, organized, scrollable interface.
• Interactive Interface: Buttons and dropdowns for a seamless user experience.
• Query History: Track and review previous queries and their corresponding code and results.

Installation

1. Clone the repository:

   git clone https://github.com/Abhigyan126/Prompt2query.git

2. Navigate into the project directory:

   cd prompt2query

3. Install the required dependencies:

   pip install tkinter google-generativeai

4. Run the application:

   python Begin.py

Usage

• Load a Dataset: Use the "Attachment" button to upload a CSV file.
• Enter Query: Type your query in the input box at the bottom right of the window.
• Execute: Choose between Pandas or SQL mode from the dropdown, and click the "Execute" button to generate and run the corresponding code.
• View Results: The generated code and results will appear in the respective panes on the left-hand side.
• Log: All executed queries and results are logged for easy reference.

Dependencies

• tkinter: For the graphical user interface.
• llm_pandas: To handle natural language queries and generate code.
• datetime: For timestamping query logs.

Project Structure

.
├── Icons                  # Folder containing application icons
├── Begin.py               # Main application script
├── llm_pandas             # Library to generate code
├── llm.py                 # Library to comunicate with LLM here in this cose gemini by google
└── README.md              # Project documentation