An application posting generated text from posts on Mastodon using the Markov chain.

$ docker compose build app
$ docker compose run --rm app /app/bot run

Run docker compose run --rm app /app/bot run --help to view help. You can also pass arguments as environment variables.

This application requires a configuration file to run. The format of the configuration file is as follows:

input:
 platform: "mastodon"
 origin: ""
 access_token: ""
output:
 platform: "mastodon"
 origin: ""
 access_token: ""
 post_visibility: "unlisted"
fetch_status_count: 100
state_size: 3
min_words_count: 3

See config/bot_config.go for details.

1. Create ECR repository to upload container image and make note of the repository url.
2. Enter to cmd/lambda.
3. Run sam build to build container image.
4. Run sam deploy --guided and pass the repository url to create new lambda function.

To use the function, follow steps below:

• Set event to specify which S3 bucket and key path will be used to place configuration and model file.
  • See PostEvent struct in cmd/lambda/main.go.
• Put configration file on S3.
  • See ConfigFile struct in config/bot_config.go.