GQLFedUtils

Subcommands

purge

This subcommand will remove any type / input / enum (and their fields) not annotated with some pattern.

GraphQL schemas / directives / scalars are always kept

To use it you need a configuration file with two properties: keepPatterns, secondKeepPatterns. Example:

keepPatterns:
  - "@GK"
  - "@GateKeep"
  - "@GKeep"

secondKeepPatterns:
  - ADMIN
  - ADMINISTRATOR
  - ADMINS
  - admin
  - admins

secondKeepPatterns is optional, but it may be useful to add more specificity.

With this configuration, a type / input / enum (and their fields) that is not annotated with first and second patterns will be removed. Example:

"""
This should be removed
"""
type SomeType {
    id: ID!
}

"""
This should not be removed
@GateKeep ADMIN
"""
type AnotherType {
    """
    @GK ADMINISTRATOR
    """
    id: ID!

    """
    This will be removed because it's not annotated
    """
    name: String!
}

If you run the purge command on the input above, you'll get the following output:

"""
This should not be removed
"""
type AnotherType {
  id: ID!
}

Notice that all type / input / enum (and their fields) that weren't annotated were removed, and the annotation was removed from the ones that do were annotated.

That's basically what this subcommand does. Check src/test/resources for more examples. *.graphql are input files, *.expected.graphql are output files (the expected output for purge subcommand)

Options

• -c, --config=<config>: Path to the config file

• -s, --suffix=<suffix>: Output file(s) will contain that suffix. If not given, input file will be overwritten. The suffix is added between the file name and graphql extension (if present), e.g. if input file is file.graphql and _ -purged_ is the suffix, then the output file will be file-purged.graphql

• -e, --exclude=<excludeFiles>: List of files to exclude from processing. Exclusion files have more precedence than input files, i.e. if you provide the same file as input and exclusion, it'll be excluded

• --overwrite: If --suffix is not given, you'll be asked for confirmation to overwrite input files. Set this flag to confirm your decision in advance, so you won't be asked later

Important note: don't annotate Query or Mutation types. This may break the program.

dot

This subcommand will transpile your GraphQL code into dot, so you can get a visual overview of your graph.

For example, let's say you've the following graph

type Query {
  hero(id: ID!): Character
  planets: [Planet]
}

type Mutation {
  updateCharacter(
    id: ID!,
    friends: [Character!],
  ): Characters

  updatePlanet(name: String!, climate: String!): Planet
}

enum PlanetType {
  GAS_GIANT,
  NEPTUNIAN,
  SUPER_EARTH,
  TERRESTRIAL
}

type Character {
  name: String
  friends: [Character]
  homeWorld: Planet
}

type Planet {
  name: String
  climate: String
  type: PlanetType!
}

examples/dot/example.graphql

If you run

java -jar gqlfedutils.jar dot example.graphql --output example.dot --params --enum-values

The GraphQL schema in example.graphql will be converted to dot code in example.dot.

Then, if you run

sfdp -Tsvg example.dot > example.svg

You'll get a nice visual representation of your graph in example.svg

As you can see, edges are added between inputs/types/enums because otherwise the graph would really be a mess.

Actually, it already looks a bit messy, that's why --params and --enum-values options where added

Options

• -o, --output=<outFile>: Output file in which generated dot code will be saved/

• --enum-values: Set this flag to include enum values in the dot output. This may clutter the output graph. Try it, see how it looks like, and decide.

• -p, --params: Set this flag to include parameters in the dot output. This may clutter the output graph. Try it and see how it looks like.

If you run

java -jar gqlfedutils.jar dot example.graphql --output example.clean.dot

You'll get this output

An interesting note:

These tools make extensive use of String#indexOf because it is really efficient. It's even more efficient than the implementation of Rabin-Karp's algorithm and here is why

Dependencies

• PicoCLI: Parse CLI args and colorize output

• Jetbrains Annotations: Better code documentation

• JUnit 5: Test framework

• Snake YAML: parse configuration from YAML file

License