lasse-py is a Python library used at LASSE / UFPA for digital signal processing and quantization research, developed by the LASSE group to support simulation and analysis of modern communication systems.
lasse-py provides a modular set of tools for scalar quantization, bit-level processing, and signal modeling — making it ideal for academic research, prototyping, and 6G experimental platforms.
Some of the packages are: audio, dsp, io (input / output), statistics, stochastic (processes), transforms (DCT, Fourier, etc) and util, in which one can find related modules.
If you do not have an environment yet, it is a good idea to follow the hints at https://github.com/lasseufpa/python_template and install one.
Go to the folder in which the lasse-py Python library was installed.
Make sure your Python environment has all dependencies using conda or similar software. For instance, for conda:
conda install --file requirements.txt
Then install the lasse-py library using
python setup.py install
The lasse-py library uses isort, black, etc., and was created using the template at
https://github.com/lasseufpa/python_template.
Github was configured to execute the command below each time there is a commit:
conda run pre-commit run --all-files
and before using git commit you should run the same command locally and check if all stages indicate passed, as follows
seed isort known_third_party.............................................Passed isort....................................................................Passed black....................................................................Passed Flake8...................................................................Passed pyright..................................................................Passed
Hence, follow the instructions at https://github.com/lasseufpa/python_template to create an environment and development platform (with VS Code, etc.) and then be able to commit changes to lasse-py.
If you installed using
python setup.py install
after each code modification, you will have to reinstall using the same command to update the code.
Hence, when changing the code, it is more convenient to set the environment variable PYTHONPATH to find the code. For instance, when using Windows' cmd:
set PATH=c:\mylib\lasse-py
in case the folder lasse is located at c:\mylib\lasse-py\lasse.
If you did not follow the instructions at https://github.com/lasseufpa/python_template, before doing git commit, you need to locally check whether your code is compliant, otherwise github may block the commit.
For instance, using conda, you may install
conda install -c conda-forge isort flake8 click
and complement with pip:
pip install pre-commit black
More information at https://guicommits.com/organize-python-code-like-a-pro/
We use:
pip install pipreqs
And from the project folder:
pipreqs .
(which will parse the files and save a new requirements.txt. You may move it if the file already exists)
https://github.com/librosa/librosa - audio and music analysis
https://github.com/keunwoochoi/kapre - for processing audio in GPU when using Keras
https://docs.scipy.org/doc/scipy/reference/signal.html - signal processing
https://github.com/crflynn/stochastic - stochastic (random) processes
We need to evaluate code to record and play sound:
https://github.com/spatialaudio/python-sounddevice/ - bindings for the PortAudio library and a few convenience functions to play and record NumPy arrays containing audio signals
https://github.com/PortAudio/portaudio - audio I/O library in C that runs on Windows, Mac and Linux
We need to evaluate taking in account a good class definition for signals to be used in PHY simulations:
https://github.com/veeresht/CommPy/network/dependencies
https://github.com/rwnobrega/komm/network/dependencies
https://github.com/darcamo/pyphysim/blob/master/notebooks/TDL_and_OFDM.ipynb
https://github.com/MeowLucian/Multipath_Simulation
https://github.com/kirlf/ModulationPy
akmimo
https://github.com/aldebaro/dsp-class-ufpa/blob/main/stochastic_proc.ipynb