Wheely creates one zip file (aka wheelhouse) containing all pre-downloaded project dependencies/wheels.

$ cd docker
$ docker build -t datasentics/wheely .

Pass your requirements.txt file to the following command to have your all.wheehouse.zip file generated into given /path/to/output-directory dir on your computer.

$ cat /path/to/requirements.txt | docker run -i --rm -v /path/to/output-directory:/output datasentics/wheely

Upload the wheelhouse to some (sub)folder in dbfs:/mnt using Databricks CLI or Azure Blob Storage/AWS S3 GUI.

Either install the library manually on your cluster (Libraries -> Install Library -> DBFS -> Python Whl) or use dbutils.library.install() in your notebook.