yake

A Rake-like DSL for writing AWS Lambda handlers

amancevice
168
7
Ruby

λake

gem
rspec
coverage
maintainability

ko-fi

Write your AWS Lambda function handlers using a Rake-like declarative syntax:

# ./lambda_function.rb
require 'yake'

handler :lambda_handler do |event|
  # Your code here
end

# Handler signature: `lambda_function.lambda_handler`

You can even declare Sinatra-like API Gateway routes for a main entrypoint:

# ./lambda_function.rb
require 'yake/api'

header 'content-type' => 'application/json'

get '/fizz' do
  respond 200, { ok: true }.to_json
end

handler :lambda_handler do |event|
  route event
rescue => err
  respond 500, { message: err.message }.to_json
end

# Handler signature: `lambda_function.lambda_handler`

Installation

Add this line to your application’s Gemfile:

gem 'yake'

And then execute:

bundle install

Or install it yourself as:

gem install yake

Why Is It Called “yake”?

“λ” + Rake, but “λ” is hard to type and I think “y” looks like a funny little upside-down-and-backwards Lambda symbol.

Why Use It?

So why use yake for your Lambda functions?

Event Logging

By default, the handler function wraps its block in log lines formatted to match the style of Amazon’s native Lambda logs sent to CloudWatch. Each invocation of the handler will log both the input event and the returned value, prefixed with the ID of the request:

START RequestId: 149c500f-028a-4b57-8977-0ef568cf8caf Version: $LATEST
INFO RequestId: 149c500f-028a-4b57-8977-0ef568cf8caf EVENT { … }
…
INFO RequestId: 149c500f-028a-4b57-8977-0ef568cf8caf RETURN { … }
END RequestId: 149c500f-028a-4b57-8977-0ef568cf8caf
REPORT RequestId: 149c500f-028a-4b57-8977-0ef568cf8caf  Duration: 43.97 ms  Billed Duration: 44 ms  Memory Size: 128 MB  Max Memory Used: 77 MB

Logging the request ID in this way makes gathering logs lines for a particular execution in CloudWatch much easier.

You can customize or disable the logger:

logging :off              # disables logging entirely
logging pretty: false     # Logs event/result in compact JSON
logging :on, MyLogger.new # Use a custom logger

Include Yake::Logger on a class to access this logger:

class Fizz
  include Yake::Logger
end

Fizz.new.logger == Yake.logger
# => true

API Routes

A common use of Lambda functions is as a proxy for API Gateway. Oftentimes users will deploy a single Lambda function to handle all requests coming from API Gateway.

Requiring the yake/api module will add the API-specific DSL into your handler.

Define API routes using Sinatra-like syntax

any '/…' do |event|
  # Handle 'ANY /…' route key events
end

delete '/…' do |event|
  # Handle 'DELETE /…' route key events
end

get '/…' do |event|
  # Handle 'GET /…' route key events
end

head '/…' do |event|
  # Handle 'HEAD /…' route key events
end

options '/…' do |event|
  # Handle 'OPTIONS /…' route key events
end

patch '/…' do |event|
  # Handle 'PATCH /…' route key events
end

post '/…' do |event|
  # Handle 'POST /…' route key events
end

put '/…' do |event|
  # Handle 'PUT /…' route key events
end

Helper methods are also made available to help produce a response for API Gateway:

Set a default header for ALL responses:

header 'content-type' => 'application/json; charset=utf-8'
header 'x-custom-header' => 'fizz'

Produce an API Gateway-style response object:

respond 200, { ok: true }.to_json, 'x-extra-header' => 'buzz'
# {
#   "statusCode" => 200,
#   "body" => '{"ok":true}',
#   "headers" => { "x-extra-header" => "buzz" }
# }

Route an event to one of the declared routes:

handler :lambda_handler do |event|
  route event
rescue Yake::Errors::UndeclaredRoute => err
  respond 404, { message: err.message }.to_json
rescue => err
  respond 500, { message: err.message }.to_json
end

Zero Dependencies

Finally, yake does not depend on any other gems, using the Ruby stdlib only. This helps keep your Lambda packages slim & speedy.

Support Helpers

As of ~> 0.5, yake comes with a support module for common transformations.

Enable the helpers by requiring the support submodule:

require 'yake/support'

Object helpers:

MyObject.new.some_method
# => NoMethodError

MyObject.new.try(:some_method)
# => nil

10.try(:some_method) { |x| x ** 2 }
# => 100

Hash helpers:

{ a: { b: 'c', d: 'e' }, f: 'g' }.deep_keys
# => [:a, :b, :d, :f]

left  = { a: 'b', c: { d: %w[e] } }
right = { a: 'a', c: { d: %w[d] } }
left.deep_merge(right)
# => { :a => "a", :c => { :d => ["e", "d"] } }

{ a: { b: 'c', d: 'e' }, f: 'g' }.deep_transform_keys(&:to_s)
# => { "a" => { "b" => "c", "d" => "e" }, "f" => "g" }

hash = { a: { b: 'c', d: 'e' }, f: 'g' }
hash.deep_transform_keys!(&:to_s)
# => { "a" => { "b" => "c", "d" => "e" }, "f" => "g" }

{ f: 'g', a: { d: 'e', b: 'c' } }.deep_sort
# => { a: { b: 'c', d: 'e' }, f: 'g' }

{ fizz: 'buzz' }.encode64
# => "eyJmaXp6IjoiYnV6eiJ9\n"

{ fizz: 'buzz', jazz: 'fuzz' }.except(:buzz)
# => { :fizz => 'buzz' }

{ fizz: 'buzz' }.strict_encode64
# => "eyJmaXp6IjoiYnV6eiJ9"

{ fizz: { buzz: %w[jazz fuzz] } }.stringify_names
# => { "fizz" => { "buzz" => ["jazz", "fuzz"] } }

{ 'fizz' => { 'buzz' => %w[jazz fuzz] } }.symbolize_names
# => { :fizz => { :buzz => ["jazz", "fuzz"] } }

{ fizz: 'buzz' }.to_form
# => "fizz=buzz"

{ f: 'g', a: { d: 'e', b: 'c' } }.to_json_sorted
# => '{"a":{"b":"c","d":"e"},"f":"g"}'

{ f: 'g', a: { d: 'e', b: 'c' } }.to_struct
# => #<OpenStruct f="g", a={:d=>"e", :b=>"c"}>

{ f: 'g', a: { d: 'e', b: 'c' } }.to_deep_struct
# => #<OpenStruct f="g", a=#<OpenStruct d="e", b="c">>

{ a: { b: 'c', d: 'e' }, f: 'g' }.to_dynamodb
# => { :a => { :M => { :b => { :S => "c" }, :d => { :S => "e" } } }, :f => { :S => "g" } }

{ a: { M: { b: { S: 'c' }, d: { S: 'e' } } }, f: { S: 'g' } }.to_h_from_dynamodb
# => { :a => { :b => "c", :d => "e" }, :f => "g" }

Integer helpers:

7.weeks
# => 4_233_600

7.days
# => 604_800

7.hours
# => 25_200

7.minutes
# => 420

1234567890.utc
# => 2009-02-13 23:31:30 UTC

String helpers:

host = 'https://example.com/'
path = '/path/to/resource'
host / path
# => "https://example.com/path/to/resource"

'snake_case_string'.camel_case
# => "SnakeCaseString"

"Zml6eg==\n".decode64
# => "fizz"

'fizz'.encode64
# => "Zml6eg==\n"

'fizz'.md5sum
# => "b6bfa6c318811be022d4f73070597660"

'fizz'.sha1sum
# => "c25f5985f2ab63baeb2408a2d7dbc79d8f29d02f"

'CamelCaseString'.snake_case
# => "camel_case_string"

'Zml6eg=='.strict_decode64
# => "fizz"

'fizz'.strict_encode64
# => "Zml6eg=="

'{"fizz":"buzz"}'.to_h_from_json
# => { "fizz" => "buzz" }

'fizz=buzz'.to_h_from_form
# => { "fizz" => "buzz" }

'2009-02-13T23:31:30Z'.utc
# => 2009-02-13 23:31:30 UTC

Symbol helpers

:snake_case_symbol.camel_case
# => :SnakeCaseSymbol

:CamelCaseSymbol.snake_case
# => :camel_case_symbol

UTC Time helpers

UTC.at 1234567890
# => 2009-02-13 23:31:30 UTC

UTC.now
# => 2022-02-26 13:57:07.860539 UTC

Datadog Integration

As of v1.0, yake comes with a helper for writing Lambdas that integrate with Datadog’s datadog-lambda gem.

As of ~> 0.8, yake uses the v2 Datadog Lambda gem.

Creating a Lambda handler that wraps the Datadog tooling is easy:

require 'aws-sdk-someservice'
require 'yake/datadog'

# Configure Datadog to use AWS tracing
Datadog::Lambda.configure_apm { |c| c.tracing.instrument :aws }

datadog :handler do |event|
  # …
end

Deployment

After writing your Lambda handler code you can deploy it to AWS using any number of tools. I recommend the following tools:

  • Terraform — my personal favorite Infrastructure-as-Code tool
  • AWS SAM — a great alternative with less configuration than Terraform
  • Serverless — Supposedly the most popular option, though I have not used it

Development

After checking out the repo, run bundle to install dependencies. Then, run rake spec to run the tests.

Contributing

Bug reports and pull requests are welcome on GitHub at amancevice/yake.

License

The gem is available as open source under the terms of the MIT License.