Searchable Django Applications with Elastic App Search (PyCon US 2021)

Where are the slides for the Workshop?

Here they are!

What to do before the Workshop?

• Fork / Clone the repository

  • $ git clone https://github.com/elastic/pycon-2021-workshop-app-search
  • $ cd pycon-2021-workshop-app-search

• Create a virtual environment and install all dependencies in requirements.txt:

  • $ python -m venv venv
  • $ source venv/bin/activate
  • $ (venv) python -m pip install -r requirements.txt

• Start an Enterprise Search instance. There are two options here:

  • Create an Elastic Cloud account (comes with a 14 day free trial no credit card required)
  • Install docker-compose and use docker-compose up to start an Enterprise Search instance locally

• Update the url and private_key values in config.yml per the instructions for starting the Enterprise Search instance below.

• Make sure at least examples/example_1_making_requests.py is able to run completely. You should see the message Everything worked! :-) in your terminal. All examples are meant to be run via python examples/example_N_...py or via the "Run" functionality in your IDE.

Starting an Enterprise Search instance

Elastic Cloud

• Create an Elastic Cloud account
• Create a new Deployment using the "Enterprise Search" template
• Select a cloud provider and region and select Create Deployment. The instance will be provisioned and available in a few minutes
• Click Launch or Copy Endpoint in the new deployment overview and copy the URL (minus the /login path). The result should look like approximately this: https://(deployment name)-1a6dc7.ent.us-west1.gcp.cloud.es.io
• Open config.yml and set url field to be "https://(deployment name)-1a6dc7.ent.us-west1.gcp.cloud.es.io" and run the script and ensure it passes without error.
• Select Launch App Search, from here you'll be placed in a getting started screen. You can follow these steps if you'd like otherwise you can select Skip Onboarding in the top.
• That's it, you're now in App Search!

Docker Compose

• Download and install docker-compose
• With an open terminal in the cloned git repository run docker-compose up. Wait for all services to spin up and configure for the first time which may take a minute.
• You should be able to connect to http://localhost:3002 in a browser
• Open config.yml and set url field to be "http://localhost:3002" and run the script and ensure it passes without error.
• The username will be enterprise_search and the password will be changeme.
• Select Launch App Search, from here you'll be placed in a getting started screen. You can follow these steps if you'd like otherwise you can select Skip Onboarding in the top.
• That's it, you're now in App Search!

Finding your Private Key

• When in App Search, navigate to the Credentials tab
• Click the Copy to Clipboard button next to the Private Key
• Paste the value into the private_key field in config.yml

Getting Started with the Django App

Setup

• Ensure your virtualenv is activated and requirements.txt is installed
• Change directory into django/
• Run python manage.py migrations to create db.sqlite3
• Run python manage.py runserver to start the server on http://localhost:8000. Don't use these exact configurations in production!
• Access http://localhost:8000/reset to load data into the Django database

URLs of Interest

• http://localhost:8000/<park-id> Shows a park by ID within the Django database
• http://localhost:8000/search Has the Reference UI configured for App Search
• http://localhost:8000/remove-park Removes park-acadia from the Django database, see how signals keep App Search up to date.
• http://localhost:8000/add-park Adds park-acadia from the Django database, see how signals keep App Search up to date.
• http://localhost:8000/remove-m2m Removes the state Tennessee from park-great-smoky-mountains, see how signals keep App Search up to date.
• http://localhost:8000/add-m2m Adds the state Tennessee to park-great-smoky-mountains, see how signals keep App Search up to date.
• http://localhost:8000/reset Resets the Django database to its original set of data (takes a while because we're not using background tasks)
• http://localhost:8000/run-task Runs the background task to index documents from the database into a new source engine and update a meta engine.

Indexing Models into App Search

See django/parks/models.py for the signals approach to indexing documents in App Search. See django/parks/tasks.py for the tasks approach to indexing documents in App Search.

When doing this in real life I recommend using background tasks for signal handlers to avoid blocking web requests.

Reference UI

The Reference UI has been generated and the production compressed JS and CSS are stored in django/static/. When you're creating your own Reference UI you should not copy these and instead follow the instructions on the Reference UI repository.