Serverless Framework - Create PDF in AWS Lambda with Chromium and Mustache
Main takeaways:
1 - Serverless Framework Packaging
As described in the Serverless Framework Packaing Guide", Serverless Framework will auto-detect and exclude development dependencies based on the runtime your service is using.", that means, for Node.js projects, everything inside devDependencies will not be included in your bundle.
In our case, we do want to deploy our main packages, we need to declare in our package.json:
...
"dependencies": {
"chrome-aws-lambda": "2.1.1",
"mustache": "4.0.1",
"puppeteer-core": "2.1.1"
}
...As descrived on chrome-aws-lambda repository: "This package is versioned based on the underlying puppeteer minor version...", to avoid headaches we are keeping the versions in sync.
Everything else that you may use for development or testing should be installed in devDependencies
2 - Teach API Gateway to respond in Binary Formats
Using Amazon API Gateway, we need to define each method or resource we want to create and respond to, for example: "on path '/my-path' respond with HTTP method GET accepting formats text/html". We could break it down in at least 3 parts: 1) a path, 2) a method, and 3) a request format.
Serverless Framework takes care of all of it for us and API Gateway by default treats message body as text payloads. To enable any other responses from your AWS Lambda (e.g. JPEG, GZip or XML file), you need to enable binary media types in API Gateway.
Again, Serverless Framework make it easy for us to implement that. In your serverless.yml you can define:
provider:
apiGateway:
binaryMediaTypes:
- '*/*'With that, API Gateway will look at the response Content-Type and request Accept HTTP headers to decide which media types should be converted from text to binary.
We are using */* here, because when you perform a GET request from a browser navigation bar (e.g. in Firefox, Chrome etc) they usually do not send application/pdf, because there is no format extension in the URL. These requests fell under */*.
3 - LAMBDA-PROXY response requires correct Content-Type, body format and encode flag
We are using LAMBDA-PROXY integration and crafting the request response inside our lambda. We want it to return a binary type and to do that, we need to respond with the correct Content-Type header, in our case, it will be application/json.
In this case, we want to return a PDF file as part of the body response (enabling the Browser to render the PDF in a Tab), we will need to convert it to Base64.
The input format of a Lambda function with proxy integration event, includes a flag called isBase64Encoded.
We need to return this flag with a value of true, to tell API Gateway to convert from Base64 to binary.
Lambda communicates with API Gateway via a text base interface and API Gateway communicates with any client converting the response to them.
With that in mind, we need to build our response like:
{
statusCode: 200,
body: pdf.toString("base64"),
headers: {
"Content-Type": "application/pdf",
},
isBase64Encoded: true,
};The pdf.toString("base64") is based on pupperter API, visit their docs for more information.