Setting up buzzdetect for the first time

1. Get Conda

Install the package manager Conda.

2. Download buzzdetect

There are two ways to acquire buzzdetect. You can download an archive (.zip or .tar.gz) from our latest release, or you can use git to clone the current state of the repository to your machine. While we try to ensure the repository always reflects fully functioning code, some bugs may slip through the cracks. The releases will likely be a more stable experience, but use git if you want to get the latest improvements.

Release (stable)

You can download a release of buzzdetect as a .zip or .tar.gz folder through our GitHub Releases page. See the “Assets” accordion menu for each release and download the “Source code” asset in whichever format you want.

Git (development)

On Windows, you’ll need to install Git if you haven’t already.

  1. Open a terminal wherever you want to store the buzzdetect folder.

  2. Git clone the buzzdetect files from the GitHub repo.

    • Command: git clone "https://github.com/OSU-Bee-Lab/buzzdetect" buzzdetect

    • This installs buzzdetect to the subdirectory “buzzdetect” under your current working directory.

3. Install dependencies

We’ll refer to the buzzdetect folder in your chosen location as the “project directory.”

The project directory contains a file named environment.yml. This file specifies all of buzzdetect’s dependencies. Using Conda, we can install all of the packages buzzdetect requires into one location.

NOTE: if you want to use a GPU to (drastically) speed up processing, you must modify environment.yml. Change the - tensorflow section to read - tensorflow[and-cuda]. E.g., - tensorflow>=2.16 should become - tensorflow[and-cuda]>=2.16. If this is not done, all processing will occur on CPU.

You can install your Conda environment wherever you wish, but these steps will install it in the same directory as the buzzdetect project.

  1. Open a terminal in the the project directory

  2. Run the command: conda env create -f environment.yml -p ./.venv

    • The -p argument is the path where the Conda environment will be created

  3. Conda will create the ./.venv subdirectory and install a virtual environment with all required packages there.

4. Confirm proper installation

  1. Activate the conda environment by running conda activate ./.venv.

  2. Run conda list; you should be able to see all of the specified dependencies (along with many other packages).

5. Test an audio file

buzzdetect comes prepackaged with our latest model (the name will change, but look in the models directory.) and an audio analysis file, testbuzz.mp3. Try running an analysis!

  • If you like buttons, read the GUI documentation.

  • If you’re savvy in the command line, read the command line documentation.

  • If you’re using buzzdetect in another Python project, check out the API documentation.

For further reading, see workflow.

Happy listening!