MS3 - Model Serving Flask app (30%)

[TimkLee]

Create a simple Flask app with the following endpoints:

• [ ] /predict - predict the shot’s probability of being a goal given the inputs
• Input: the input features to your model, compatible with model.predict() or model.predict_proba()
• Output: the predictions for all of the features; the output
• [ ] /logs - retrieve logs written from your app
• [ ] /download_registry_model wraps the same function in comet ml API to download a model and update the currently loaded model
• if yes, load that model and write to the log about the model change
• if no, try downloading the model:
  • if it succeeds, load that model and write to the log about the model change.
  • If it fails, write to the log about the failure and keep the currently loaded model

[TimkLee]

Notes: Should not put any task specific functionality into your Flask app; any data preprocessing (such creating all of the features needed for your model) should be done prior to making the request to the serving app.