Skip to content
Generate a GraphQL API based on your SQL database structure
Branch: master
Clone or download
Latest commit b159d82 Jan 11, 2019
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
backends Fix lint errors Feb 13, 2017
cli Add prompt for database hostname Oct 17, 2015
steps Fix lint errors Feb 13, 2017
templates Fix bug that occurs if no edge selection is performed on relay connec… Sep 8, 2015
util Fixed a range of issues related to relay/non-relay Aug 21, 2015
.editorconfig
.eslintignore
.gitignore
LICENSE
README.md Update README.md Jan 11, 2019
cli.js Rename filename -> db-filename cli argument Oct 8, 2015
package.json

README.md

sql-to-graphql

Generate GraphQL schemas and server based on SQL table structure.

⚠️ Unmaintained

If you want to help maintain or develop this, let me know!

What?

GraphQL is pretty awesome, but getting started can be difficult - especially if you are unfamiliar with the concepts it introduces.

sql-to-graphql is a command-line utility that can help you get started. You give it the credentials to an SQL database (MySQL, PostgreSQL and SQLite currently) and it will inspect the tables it finds and do the following:

  • Generate GraphQL-types for each table (including resolvers)
  • Generate an HTTP-server based on Hapi that accepts GraphQL queries
  • Sets up a basic web-frontend that lets you query the server

Disclaimer

This utility is intended to help people get started with GraphQL. It is NOT intended to be used in production.

Installation

npm install -g sql-to-graphql

Usage

sql2graphql [options]

Options:

  • --relay, -r - Generate Relay-style schema (boolean [default: false])
  • --output-dir, -o - Directory to use when generating app (string [required])
  • --es6 - Use ES6 for generated code (boolean [default: false])
  • --database, --db - Database name (string [required])
  • --db-filename - Database filename, used for SQLite (string)
  • --host, -h - Hostname of database server (string [default: "localhost"])
  • --port, -P - Port number of database server (number)
  • --user, -u - Username to use when connecting (string [default: "root"])
  • --password, -p - Password to use when connecting (string [default: ""])
  • --table, -t - Tables to generate type schemas for (array [default: "*"])
  • --backend, -b - Type of database (string [default: "mysql"])
  • --strip-suffix - Remove a suffix from table names when generating types (array)
  • --strip-prefix - Remove a prefix from table names when generating types (array)
  • --interactive, -i - Interactive mode (boolean [default: false])
  • --colors, -c - Colorize the code output (boolean [default: false])
  • --use-tabs - Use tabs for indentation (boolean [default: false])
  • --tab-width - Width of tabs (number [default: 2])
  • --quote - Quote style (single/double) (string [default: "single"])
  • --default-description - The description to use for columns without a comment (string [default: "@TODO DESCRIBE ME"])
  • --unaliased-primary-keys Disable aliasing of primary key fields to "id" for each type (boolean [default: false])
  • --help - Show help (boolean)

A note about connections

At the moment, sql-to-graphql tries to guess connections between tables based on naming conventions. This is far from fool-proof, but here's how it works:

Given you have a table called users (or user) and a table called posts (or post) and the posts table have a column called user_id, it will create a connection called user on the Post-type, giving you the User this post belongs to. It will also create a connection on the User-type named posts, which will return all the posts belonging to this User.

How these connections look depends on the options used (namely, --relay).

License

MIT-licensed. See LICENSE.

You can’t perform that action at this time.