iromu/graphql-trino repository overview

⁠Trino GraphQL

A Spring Boot application that dynamically generates a GraphQL schema from Trino catalogs, schemas, and tables. It allows users to explore and query Trino data sources using GraphQL without manually writing schema definitions.

⁠🚀 Features

• Dynamic GraphQL Schema: Automatically scans Trino catalogs, schemas, and tables and exposes them as GraphQL queries.
• Auto-detect Columns: Generates GraphQL object types from Trino table column metadata.
• Filter Support: Query data with dynamic filters using GraphQL input arguments.
• Streamed Results: Efficient streaming of query results from Trino using JDBC.
• GraphiQL Interface: Visual GraphQL playground is exposed at the root URL / for easy testing and exploration.
• Export Schema: Exposes a REST endpoint to download the auto-generated GraphQL schema in SDL format.

⁠🧠 How It Works

1. On startup, the app connects to Trino via JDBC.
2. It fetches available catalogs, schemas, tables, and columns.
3. GraphQL schema is dynamically generated using this metadata.
4. Each table is exposed as a top-level query.
5. Query fields support filters using a generic FilterInput object.
6. Queries return streamed Map<String, Object> rows to minimize memory usage.

⁠🔧 Configuration

Set up your connection to Trino in application.yml or application.properties:

spring:
  datasource:
    url: jdbc:trino://localhost:8080
    username: your-trino-user
    driver-class-name: io.trino.jdbc.TrinoDriver

⁠⚙️ Application Configuration (app.*)

The application supports flexible configuration via application.yml or application.properties, using the app prefix. These properties control how GraphQL schemas are generated from Trino catalogs.

⁠Available Configuration Options

⁠Example (application.yml)

app:
  schema-folder: /tmp/schema
  replace-objects-name-characters: true
  ignore-objects-with-wrong-characters: false
  ignore-cache: true
  include-catalogs:
    - my_catalog
  exclude-catalogs:
    - system
  exclude-schemas:
    - information_schema
    - pg_catalog

⁠🐳 Overriding Configuration via Docker Environment Variables

Spring Boot automatically maps environment variables to configuration properties using a relaxed binding system. This means all the app.* properties can be overridden via Docker environment variables using the following rules:

• Dots (.) in property names become underscores (_)
• Everything is uppercased
• Prefix with APP_ for properties under app

⁠🧪 Example Mappings

⁠🐳 Docker environment Example (in docker-compose.yml)

services:
  graphql-trino:
    image: iromu/graphql-trino:latest
    environment:
      - APP_SCHEMA_FOLDER=/tmp/schema
      - APP_REPLACE_OBJECTS_NAME_CHARACTERS=true
      - APP_IGNORE_OBJECTS_WITH_WRONG_CHARACTERS=false
      - APP_IGNORE_CACHE=true
      - APP_INCLUDE_CATALOGS_0=my_catalog
      - APP_EXCLUDE_CATALOGS_0=system
      - APP_EXCLUDE_SCHEMAS_0=information_schema
      - APP_EXCLUDE_SCHEMAS_1=pg_catalog

⁠🧪 Example GraphQL Query

query {
    my_table(limit: 10, filters: [
        { field: "age", operator: "gt", intValue: 30 },
        { field: "name", operator: "like", stringValue: "Ali" }
    ]) {
        id
        name
        age
    }
}