cdk-tweet-queue 1.0.141

Defines an SQS queue with tweet stream from a search

Tweet Queue for AWS CDK

This is an AWS CDK construct library which allows you to get a feed of Twitter search results into an SQS queue. It works by periodically polling the freely available Twitter Standard Search API and sending all new tweets to an SQS queue.

Inspired by @jlhood's aws-serverless-twitter-event-source

Architecture

1. A CloudWatch Event Rule triggers the poller AWS Lambda function periodically
2. The poller reads the last checkpoint from a DynamoDB table (if exists)
3. The poller issues a Twitter search query for all new tweets
4. The poller enqueues all tweets to an SQS queue
5. The poller stores the ID of the last tweet into the DynamoDB checkpoint table.
6. Rinse & repeat.

Twitter API Keys

To issue a Twitter search request, you will need to apply for a Twitter developer account, and obtain API keys through by defining a new application.

The Twitter API keys are read by the poller from an AWS Secrets Manager entry. The entry must contain the following attributes: consumer_key, consumer_secret, access_token_key and access_token_secret (exact names).

1. Create a new AWS Secrets Manager entry for your API keys
2. Fill in the key values as shown below:
3. Store the key
4. Obtain the ARN of the secret (you will need it soon).

Usage

Use npm to install the module in your CDK project. This will also add it to your package.json file.

$ npm install cdk-tweet-queue

Add a TweetQueue to your CDK stack:

# Example automatically generated from non-compiling source. May contain errors.
from cdk_tweet_queue import TweetQueue

queue = TweetQueue(self, "TweetStream",
    # this is the ARN of the secret you stored
    secret_arn="arn:aws:secretsmanager:us-east-1:1234567891234:secret:xxxxxxxxx",

    # twitter search query
    # see https://developer.twitter.com/en/docs/tweets/search/guides/standard-operators
    query="#awscdk",

    # optional properties
    interval_min=60,  # optional: polling interval in minutes
    retention_period_sec=60,  # optional: queue retention period
    visibility_timeout_sec=60
)

Now, queue is an sqs.Queue object and can be used anywhere a queue is accepted. For example, you could process the queue messages using an AWS Lambda function by setting up an SQS event source mapping.

Development

The project is managed by projen and offers the following commands:

• yarn projen - Synthesize the project configuration.
• yarn compile - Compile all source code.
• yarn test - Run all tests.
• yarn build - Complie, test, and package the module.

Integration test

There is also an integration test that can be executed by running the following commands. You will need to set the TWEET_QUEUE_SECRET_ARN environment variable in order for the test to be able to use your Twitter API keys.

$ yarn integ:deploy

Don't forget to destroy:

$ yarn integ:destroy

You can also run any cdk command on the integration test application by running:

yarn integ <command>

License

Apache-2.0