Skip to main content

Deploy WebChat Lambda

Prerequisites​

  • For our WebChat to work, we would need to setup few resources on AWS. To do this, we must have Terraform CLI installed and configured with a private token.
  • One of the services we would deploy would be an AWS Lambda, so we need Python and a package manager installed in our machine, preferably Poetry.

Configuring Poetry​

At the time of writing, Poetry's latest stable version is v1.1. But we need at least v1.2 to make the whole setup process easier. To update you poetry version to the required version, run the following commands:

poetry self update --preview
poetry plugin add git+https://github.com/python-poetry/poetry-bundle-plugin#main

This is a one-time setup.

note

If you get an error about Poetry not being installed with the official installer, uninstall Poetry first and use the official installer this time. Once installed, run the above commands again.

Setting up the code for Lambda​

Create a new Poetry project by using the command:

poetry new my-web-chat-lambda
cd my-web-chat-lambda

Now follow these instructions to setup private Neev repository for Poetry. Once added, Poetry would look for Neev packages from this repository and the remaining packages from PyPi.

Let's install web-chat-lambda:

poetry add web-chat-lambda

This package includes a helpful lambda function, which we will use directly afterwards to avoid having to write any code.

Zipping our code​

Our Python code and its dependencies must be zipped in order to use with the Terraform template. Because we are not writing any code, this just means that only our dependencies must be zipped.

To create a zip, follow these instructions:

poetry bundle venv bundle-folder
# replace `python3.9` with the folder you have on your machine
cd bundle-folder/lib/python3.9/site-packages

Now we have to zip all the contents of site-packages/. Let's name it lambda-code.zip:

# while being inside `site-packages/`, zip the contents of this folder
zip -qr lambda-code.zip *

# and move the zip to the project root
mv lambda-code ../../../../

Deploying AWS resources using Terraform​

Follow the instructions in this repository to leverage a Terraform (TF) template. Configure the file along with changing these attributes:

module "lambda_module" {
    # ...
    lambda_function_handler = "web_chat_lambda.lambda_function.lambda_handler"
    lambda_source = "../path/to/lambda-code.zip"
    # ...
}

Given that you have configured Terraform CLI with a private token to retrieve the TF template (if not done, see prerequisites), you can use terraform init followed by terraform apply to create the required resource.

Setting up an API Gateway​

You would have add the deployed Lambda to an API Gateway manually so that the WebChat frontend could use it. This Lambda would be receiving a POST request.