Skip to main content

A tool to help with copying and pasting code context into LLM chats

Project description

LLM Code Context

LLM Code Context is a Python-based tool designed to streamline the process of sharing code context with Large Language Models (LLMs). It allows developers to easily select, format, and copy relevant code snippets and project structure information, enhancing the quality of interactions with AI assistants in coding tasks.

This project was developed with significant input from Claude 3 Opus and Claude 3.5 Sonnet. All of the code that makes it into the repo is human curated (by me ๐Ÿ˜‡, @restlessronin).

Features

  • File Selection: Offers a command-line interface for selecting files from your project.
  • Intelligent Ignoring: Respects .gitignore rules and additional custom ignore patterns to exclude irrelevant files.
  • Customizable Output: Uses Jinja2 templates for flexible formatting of the selected code context.
  • Folder Structure Visualization: Generates a textual representation of your project's folder structure.
  • Clipboard Integration: Automatically copies the generated context to your clipboard for easy pasting.
  • Configuration Management: Supports user-specific and project-specific configurations for a tailored experience.

Installation

Using pipx (Recommended)

pipx is a tool to help you install and run end-user CLI applications written in Python.

  1. If you haven't installed pipx yet, follow the installation instructions in the pipx documentation.
  2. Once pipx is installed, you can install LLM Code Context:
    pipx install llm-code-context
    

This will install LLM Code Context in an isolated environment and make its commands available in your shell.

Usage

LLM Code Context offers various command-line tools, each designed for a specific task. All commands should be run from the root directory of your project, where the .gitignore file is located. This is crucial because:

  1. The default file selection process starts from this root directory and selects all files that are not ignored by .gitignore rules.
  2. It ensures that file selection and ignore rules are applied correctly and consistently.

Here are the main commands:

  1. Select files:

    lcc-select
    
  2. Generate context from selected files:

    lcc-genfiles
    
  3. Generate folder structure diagram:

    lcc-dirtree
    

Typical workflow:

  1. Navigate to your project's root directory in the terminal.
  2. Edit the project configuration file .llm-code-context/config.json to add any files to the "gitignores" key that should be in git but may not be useful for code context (e.g., "LICENSE" and "poetry.lock", maybe even "README.md").
  3. Run lcc-select to choose the files you want to include in your context. You can look at .llm-code-context/scratch.json to see what files are currently selected. If you prefer, you can edit the scratch file directly, before the next step.
  4. Run lcc-genfiles to process the selected files and copy the formatted context to your clipboard.
  5. Paste the context into your conversation with the LLM.

Configuration

LLM Code Context uses three configuration files:

  1. User Configuration (located in a platform-specific directory determined by platformdirs):
{
  "templates_path": "/path/to/your/templates"
}
  1. Project Configuration (.llm-code-context/config.json in your project root). Example:
{
  "template": "all-file-contents.j2",
  "gitignores": [".git", "LICENSE"]
}
  1. Scratch Configuration (.llm-code-context/scratch.json in your project root). Example:
    • Keeps track of the currently selected files.

You can edit these files manually or use the provided interfaces to update them.

Project Structure

โ””โ”€โ”€ llm-code-context.py
    โ”œโ”€โ”€ .gitignore
    โ”œโ”€โ”€ .llm-code-context
    โ”‚   โ”œโ”€โ”€ .gitignore
    โ”‚   โ””โ”€โ”€ config.json
    โ”œโ”€โ”€ LICENSE
    โ”œโ”€โ”€ MANIFEST.in
    โ”œโ”€โ”€ README.md
    โ”œโ”€โ”€ poetry.lock
    โ”œโ”€โ”€ pyproject.toml
    โ”œโ”€โ”€ src
    โ”‚   โ””โ”€โ”€ llm_code_context
    โ”‚       โ”œโ”€โ”€ __init__.py
    โ”‚       โ”œโ”€โ”€ config_manager.py
    โ”‚       โ”œโ”€โ”€ context_generator.py
    โ”‚       โ”œโ”€โ”€ file_selector.py
    โ”‚       โ”œโ”€โ”€ folder_structure_diagram.py
    โ”‚       โ”œโ”€โ”€ git_ignorer.py
    โ”‚       โ”œโ”€โ”€ initializer.py
    โ”‚       โ”œโ”€โ”€ pathspec_ignorer.py
    โ”‚       โ”œโ”€โ”€ template_processor.py
    โ”‚       โ””โ”€โ”€ templates
    โ”‚           โ””โ”€โ”€ all-file-contents.j2
    โ””โ”€โ”€ tests
        โ””โ”€โ”€ test_pathspec_ignorer.py

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

This project is licensed under the Apache License, Version 2.0. See the LICENSE file for details.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

llm_code_context-0.0.5.tar.gz (12.4 kB view details)

Uploaded Source

Built Distribution

llm_code_context-0.0.5-py3-none-any.whl (13.9 kB view details)

Uploaded Python 3

File details

Details for the file llm_code_context-0.0.5.tar.gz.

File metadata

  • Download URL: llm_code_context-0.0.5.tar.gz
  • Upload date:
  • Size: 12.4 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: poetry/1.8.2 CPython/3.12.4 Darwin/23.6.0

File hashes

Hashes for llm_code_context-0.0.5.tar.gz
Algorithm Hash digest
SHA256 7507a9514699e30478d2a3ae355c7a843d02d2ac0916427f943fb6466169f1cc
MD5 e6ad6d26c3d9e56f877fc17415dd06b2
BLAKE2b-256 3df23e1107d5c9fbc4d2c9b7bed229f4f1bc95bfd9b67fab3d862aa3f491a2b0

See more details on using hashes here.

File details

Details for the file llm_code_context-0.0.5-py3-none-any.whl.

File metadata

File hashes

Hashes for llm_code_context-0.0.5-py3-none-any.whl
Algorithm Hash digest
SHA256 aaad3ab2a22c16293e16ccb0b177d5b1de664c187e722a2170bb75eb113533eb
MD5 38645917d9226f4c5b3faa710c057f5e
BLAKE2b-256 e02823ba2d6146120eaa4ca7cc8f1e71af608e8811bbc2818a62f6dc55913ad5

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page