The Best Type of Coffee

Get your right type of coffee ☕

Go application that provides information on the best type of coffee based on user input characteristics. It uses two data sources: the Google Gemini API and a database populated by ETL from a dataset. This dataset was gathered from Kaggle (kudos for the author), see there for further details.

Usage

To run the application, you need to have:

• Go installed
• Google Gemini API key, you can create one here
• Postgres URL (optional)
  • You can run a local docker instance by running the command $ make pg-up or pointing to any other instance, like a cloud instance

The application has tree command entry points:

ETL

This entrypoint has the responsibility to get the Coffee CSV dataset, convert the rows and then insert them into the Postgres database.

It will create and populate the second data source to retrieve the best type of coffee for you, which you can get using the other 2 commands, API and CLI.

This is optimal if only want the Gemini API result, so skip it.

You can run the ETL by the command:

go run cmd/etl/main.go --DATABASE_URL <database_url>

API

To get started, set up your local environment by copying .env.example to .env and configuring the necessary variables.

You can choose to run either an HTTP REST API or a gRPC server:

• HTTP REST API: Runs by default unless gRPC is enabled.
• gRPC Server: To enable gRPC, set the GRPC_ENABLED variable to true in your .env file.

Once you've configured your environment, start the server using the following command:

make run

The application will launch the selected server (HTTP or gRPC) on the default port, or on a custom port defined in your .env file.

For API documentation:

• The full OpenAPI/Swagger spec is available for the HTTP API.
• The gRPC service uses the protobuf definition.

Example Endpoint: GET /v1/best-coffees: Retrieves the best type of coffee based on user-provided characteristics.

CLI

To start the command line interface, run the following command:

go run cmd/cli/main.go --CHAT_GPT_KEY <your_chat_gpt_key> --DATABASE_URL <database_url>

This will run the CLI interface, which will output the best coffees for you from the selected characteristics.

Data Sources

It uses two data sources:

Gemini API

The Gemini API is used to generate responses to user feedback. The application sends user feedback to the API and receives a response based on the user's input.

Database

The application also uses a database that is populated by ETL (Extract, Transform, Load) dataset data.

License

The Best Type of Coffee is released under the Apache License. See LICENSE for details.