Johns-Frends

Andrew Tate
John Leeds
Luke Samuels
Rhys Sorenson-Graff

Contributions:

Andrew Tate: Andrew worked on the initial setup of our repository and our first functions to play audio. He also worked on the playback speed audio modification.
John Leeds: John contributed to the organization of the repository, testing, and code review.
Luke Samuels: Luke helped refactor to separate our source code and tests. Luke also worked on the reverse audio modification.
Rhys Sorenson-Graff: Rhys researched and contributed conversion of .wav files to numpy arrays. This is how we originally intended to handle audio modifications before discovering the built in audioop library.

Installation:

Create a virtual environment with python3.11 -m venv venv
Activate it with venv/Scripts/activate (Windows) or source venv/bin/activate (Linux and Mac).
Install dependencies with python -m pip install -r requirements.txt.
Note: Deactivate the virtual environment with deactivate.
Source: Python Virtual Environments: A Primer

Getting Started:

First, initialize the database by running python src/sqlite_init.py
From here, you can add whatever sounds you want with python src/cli.py add [path to sound] [name (optional)]
Once you have added sounds to the archive, you can play them with python src/cli.py play [name]
You can optionally specify audio effects to apply such as reversing the sound, changing the volume, changing the speed, or playing multiple sounds in parallel
Other commands: rename, list, remove, clean
For more information, run python src/cli.py -h or python src/cli.py [command] -h
For more details on the technical design, see design.md

Testing:

Tests are written with the built in unittest library and can be executed with python -m unittest.

Our automatic tests are located in the tests/ directory, and we have written tests for as many of our commands as possible. There are tests for adding, renaming, and removing sounds as well as tests for adding and removing tags.

Our audio functions are all tested manually.

Note that when creating new tests, the file must begin with test_ in order for them to be discovered.

Challenges:

One of the biggest challenges we faced was creating audio modifications (such as changing the speed or volume). Because the simpleaudio documentation references using numpy arrays, we thought we would need to modify the .wav files using a similar process. However, we eventually discovered the audioop library that is built into Python, which had all of the features that we wanted.

In particular, we struggled to implement modifying playback speed. When using pydub (which seems to be commonly suggested online), our audio was corrupted when we exported it and then opened it with wave. Interestingly, the .wav file was fine when using a built-in audio player. We were able to work around this by first exporting it as an mp3, reading that as a pydub AudioSegment, exporting that as a wav, and then reading that. Obviously, this is not how we would like to handle this, but we can change it later if necessary.

We also are currently deciding how to store sounds in the archive. Right now, we just are throwing all the sounds in the sounds/ directory, but this is subject to change.

ProfJanetDavis / Johns-Frends