SemantiCLI

Semantic search in the command line.

look up strings using semantic search, see cosine similarity between strings.

Installation

From PyPI:

pip install semanticli

From source:

pip install .

Quick Usage

smnti --help
smnti "Search Me" -f text.txt
cat Alice.txt | smnti "Deck of cards"
smnti -cs 0.6 -c 200 -o 10 -s \n -n "Big Ship" -f treasure.txt

Command-Line Options

-V or --version: Print current version.

--debug: Print debug logs.

-cs or --cos-sim: Cosine similarity greater or equal threshold for printing a match (between 0-1), default is 0.8.

-v or --invert-match: Print anything below the cosine similiarity threshold.

-n or --line: Print file name, line numbers, and similarity.

-s sep_str or --separator sep_str: Separation string to divide with, default is \n\n.

-c size or --chunk size: Chunk size number, might be larger if didn't encounter the separator, default is 4000.

-o size or --chunk-overlap size: Overlap size number between chunks, default is 200.

-lm or --list-models: List installed embedding models.

-m model or --model model: Use specified model for this search only, default is Xenova/all-MiniLM-L6-v2.

-im model or --install-model model: Install specified embeddings model from hugging face.

-rm model or --remove-model model: Uninstall specified model.

note:

SemantiCLI is using ONNX runtime for model inference. it can only install models that have "model.onnx" in their repo at hugginface.

Development Setup

1. Clone the repository:

git clone https://github.com/yairzaf/semanticli.git
cd semanticli

2. Set up development environment:

make dev

3. Activate the virtual environment:

# Linux/Mac:
source venv/bin/activate

# Windows:
.\venv\Scripts\activate

Available Make Commands

• make help - Show available commands
• make venv - Create virtual environment
• make install - Install package and dependencies
• make dev - Set up development environment (venv + install)
• make clean - Remove virtual environment and build artifacts
• make test - Run tests
• make lint - Run linter checks
• make format - Format code using black
• make binary-linux - Create standalone Linux binary
• make binary-windows - Create standalone Windows binary (requires Wine on Linux/WSL)

Contributing

We welcome contributions! Please follow these steps:

1. Fork the repository
2. Create your feature branch from develop:

git checkout -b feature/amazing-feature develop

3. Set up development environment:

make dev

4. Make your changes and follow these guidelines:

   • Write meaningful commit messages following Conventional Commits
   • Add tests for new functionality
   • Update documentation as needed
   • Follow the existing code style

5. Run tests and linting:

make test
make lint

6. Push to your fork:

git push origin feature/amazing-feature

7. Create a Pull Request to the develop branch

Branch Naming Convention

• feature/* - New features
• bugfix/* - Bug fixes
• docs/* - Documentation changes
• refactor/* - Code refactoring
• release/* - Release preparation

Commit Message Format

feat: add new export command
fix: resolve argument parsing error
docs: update installation guide
test: add unit tests for export feature
refactor: restructure command handling

After Pull Request Merge

Clean up your branches:

git checkout develop
git pull origin develop
git branch -d feature/amazing-feature
git push origin --delete feature/amazing-feature

Building Binaries

Create standalone executable:

For Linux:

make binary-linux
# Binary will be in dist/semanticli

For Windows (from Linux/WSL):

# Install Wine if needed:
sudo apt-get install wine

make binary-windows
# Binary will be in dist/semanticli.exe

License

This project is licensed under the MIT License - see the LICENSE file for details.