CheatSheet

A productivity self-hosted tool to manage and browse cheatsheet and snippets (in a Markdown format).

This project is based on Bookmarker code base.

Search

Finding a cheatsheet is very easy. The search is a multi-line search where each line is matched independently (AND relation between the lines), either in a fuzzy or a full-match fashion.

Semantic Search and Answer Generation

In addition to the classic fuzzy/keyword-based search, a semantic search is also supported. Simply describe what you are looking for in a natural language and the most relevant cheatsheet will be presented.

It also generates an answer to your search query based on the semantic search result using a large language model (LLM) that runs locally. This allows for not just finding relevant content, but also receiving a direct, AI-generated response to your query for a more comprehensive and streamlined experience.

Privacy

Snippets are stored locally in a SQLite database. This is useful when you don't want to expose sensitive data (e.g., in your workplace) with third-party tools. Note that it is recommended to periodically back up the SQLite file.

Installation

Requirements:

1. Python 3.7 or higher

Linux / macOS:

1. Create a directory named venv
2. Run:

python -m venv <route/to/venv-directory>
source <route/to/venv-directory>/bin/activate
pip install -r requirements.txt
pip install -r requirements_ai.txt

Note: If you are not interested in running an LLM locally, you can configure enable_ai to false and skip the last command.

Running the application

1. Activate the virtual environment: source venv/bin/activate
2. Run python src/main.py
3. From the browser: http://localhost:8000

Tests

This project is well tested with unittests and end-to-end tests. The code coverage is 98%.

The build.py script runs the following static code analysis and test suites:

• pycodestyle (pep8)
• pylint
• Unit tests
• End-to-end tests

To run build.py:

1. Activate the virtual environment: source venv/bin/activate
2. Install dev dependencies: pip install -r requirements_dev.txt
3. Run: PYTHONPATH=src python build.py

To check coverage:

1. PYTHONPATH=src coverage run --concurrency=multiprocessing build.py
2. coverage combine
3. coverage html
4. Open htmlcov/index.html in your browser