prefect-example

Build, manage, and monitor your real-time machine learning application using Ensign, Prefect, and River.

This is an example of a real-time sentiment analysis application that ingests sample Yelp review data downloaded from Kaggle and makes real-time predictions using Ensign and River.

You have created the application but how do you manage and monitor it over time? This is where Prefect comes in. It is an extremely simple to use workflow orchestration platform. All you have to do is install it using pip and then you can turn functions in your code to Flows that you can monitor in a dashboard. You can either host a server locally within your own infrastructure or sign up for Prefect Cloud.

To use Ensign, create a free account on rotational.app. You will need to do the following once you create an account:

• Create a project.
• Add the following topics to the project: river_pipeline, river_metrics. Check out this video on how to add a topic. You can choose your own names for the topic but make sure that you update the code accordingly.
• Generate API keys for your project.

You will need to create and source the following environment variables using the API keys you just downloaded:

export ENSIGN_CLIENT_ID="your client id here"
export ENSIGN_CLIENT_SECRET="your client secret here"

The sentiment analysis application consists of three components:

• YelpDataPublisher reads data from the yelp.csv file and publishes to the river_pipeline topic. Note that this csv can easily be replaced by a real-time data source. Check out Rotational Labs Data Playground for examples!
• YelpDataSubscriber listens for new messages on the river_pipeline topic and loads and uses an online model that learns incrementally as it receives new training instances. It also calculates the precision and recall scores which it publishes to the river_metrics topic.
• MetricsSubscriber listens for new messages on the river_metrics topic and checks to see if the precision or recall score are below a pre-specified threshold and prints the values if they fall below the threshold.

Steps to run the application

Create a virtual environment

$ virtualenv venv

Activate the virtual environment

$ source venv/bin/activate

Install the required packages

$ pip install -r requirements.txt

Set up and start the Prefect server

You can either host your own Prefect server on your own infrastructure or use Prefect Cloud. Follow the instructions after signing up to create a Prefect server on the cloud.

If you choose to set up a local server, you simply need to open up a terminal window and run the following command to get started. Make sure that prefect is installed in your virtual environment first.

$ prefect server start

Open three more terminal windows

Run the MetricsSubscriber in the first window (make sure to activate the virtual environment first)

$ source venv/bin/activate

$ python river_sentiment_analysis.py metrics

Run the YelpDataSubscriber in the second window (make sure to activate the virtual environment first)

$ source venv/bin/activate

$ python river_sentiment_analysis.py subscribe

Run the YelpDataPublisher in the third window (make sure to activate the virtual environment first)

$ source venv/bin/activate

$ python river_sentiment_analysis.py publish

You can then take a look at the dashboard to view your jobs. It will look something like this:

Here is a view of the Flow runs:

You can also check the logs for warnings and errors: