port 4200 Udp/tcp for Prefect

Download the needed files from:

https://www1.nyc.gov/site/tlc/about/tlc-trip-record-data.page

PreReq

Start MLFLOW

cd ~/github/mlops-zoomcamp/02-experiment-tracking
mlflow ui --backend-store-uri sqlite:///mlflow.db --serve-artifacts --artifacts-destination ./artifacts

Install prefect

pip install prefect==2.0b5

Set dir from previous exercise

ln -s ../02-experiment-tracking/data/
ln -s ../02-experiment-tracking/models/ .
ln -s ../02-experiment-tracking/mlflow.db
ln -s ../02-experiment-tracking/mlruns/ .

Start the server config

cd 03-orchestration/
prefect config set PREFECT_ORION_UI_API_URL="http://localhost:4200/api"
prefect orion start --host 0.0.0.0

On client side

prefect config set PREFECT_API_URL="http://localhost:4200/api"
prefect config view

If PREFECT_API URL config is not set correctly: prefect config unset PREFECT_API_URL
prefect config set PREFECT_API_URL="http://localhost:4200/api" prefect config view prefect storage ls

Set the storage

Select the local storage for the test purpose prefect storage create prefect storage ls

Start a simple flow

python prefect_flow.py

Create the deployment

prefect  deployment create prefect_deploy.py 

Create the Work queue on the GUI

Start the work queue

prefect work-queue preview ba5b0352-75f6-4e91-8781-e6f093b69f46 prefect agent start ba5b0352-75f6-4e91-8781-e6f093b69f46