Python spectrogram generation from audio file on google cloud run

Require a python project which takes as parameter an audio file url (file stored on google storage), retrieves it locally, converts it to a log spectrogram using librosa library, and saves it locally in a .png file. Once saved the script should send a AI/ML prediction request to a deployed AutoML model (some code will be provided) & return the result. The project will be run on Google Cloud Run server and should be deployable easily with instructions. Once deployed a sample browser request like below should execute successfully: [login to view URL]