Shortcuts

Developer guidelines

Initial developer setup

git clone https://github.com/torchsynth/torchsynth
cd torchsynth
pip3 install -e ".[dev]"

Make sure you have pre-commit hooks installed:

pre-commit install

This helps us avoid checking dirty jupyter notebook cells into the repo.

Examples

In Python or Jupyter Notebook (if you want to see pretty plots), run examples/examples.py.

Python 3.9 on OSX

Unfortunately, Python 3.9 on OSX Big Sur won’t work, because librosa repends upon numba which isn’t packaged for 3.9 yet. In which case you’ll have to create a Python 3.7 conda environment. (You might also need to downgrade LLVM to 10 or 9.):

conda install -c conda-forge ipython librosa matplotlib numpy matplotlib scipy jupytext
conda install -c anaconda ipykernel
python -m ipykernel install --user --name=envname

and change the kernel to envname.

Tests

Unit testing is performed using pytest.

To run tests, run pytest from the project root:

TORCHSYNTH_DEBUG=True pytest

To run tests with a coverage report:

pytest --cov=./torchsynth

Examples also serves as an integration test:

python examples/examples.py
Read the Docs v: stable
Versions
latest
stable
v1.0.1
v1.0.0
Downloads
pdf
html
epub
On Read the Docs
Project Home
Builds

Free document hosting provided by Read the Docs.