openapi-generator/docs/usage.md
Vincent Devos c941b87067 [ONLINE] Minor links fix (#2644)
- usage : fix link to generator-specified parameters
- usage : fix next page link not found
2019-04-11 23:29:18 +08:00

19 KiB

id title
usage Usage

Options for OpenAPI Generator are the same whether you're using the CLI, Maven/Gradle Plugins, or Online generation options. This page demonstrates navigating the options via CLI. Commands are presented here in a logical progression as a tutorial, but you're welcome to skip directly to the generate command.

help

The help option lists all commands available to the CLI.

openapi-generator help
usage: openapi-generator-cli <command> [<args>]

The most commonly used openapi-generator-cli commands are:
    config-help   Config help for chosen lang
    generate      Generate code with the specified generator.
    help          Display help information
    list          Lists the available generators
    meta          MetaGenerator. Generator for creating a new template set and configuration for Codegen.  The output will be based on the language you specify, and includes default templates to include.
    validate      Validate specification
    version       Show version information

See 'openapi-generator-cli help <command>' for more information on a specific
command.

list

The list command outputs a formatted list of every available generator. Pass the -s/--short option if you would like a CSV output for easy parsing.

openapi-generator help list
NAME
        openapi-generator-cli list - Lists the available generators

SYNOPSIS
        openapi-generator-cli list [(-s | --short)]

OPTIONS
        -s, --short
            shortened output (suitable for scripting)

Example:

openapi-generator list -s | tr ',' '\n'

For the full list of generators, refer to the Generators List.

config-help

The config-help option provides details about

openapi-generator help config-help
NAME
        openapi-generator-cli config-help - Config help for chosen lang

SYNOPSIS
        openapi-generator-cli config-help
                [(-f <output format> | --format <output format>)]
                [(-g <generator name> | --generator-name <generator name>)]
                [--markdown-header] [--named-header]
                [(-o <output location> | --output <output location>)]

OPTIONS
        -f <output format>, --format <output format>
            Write output files in the desired format. Options are 'text',
            'markdown' or 'yamlsample'. Default is 'text'.

        -g <generator name>, --generator-name <generator name>
            generator to get config help for

        --markdown-header
            When format=markdown, include this option to write out markdown
            headers (e.g. for docusaurus).

        --named-header
            Header includes the generator name, for clarity in output

        -o <output location>, --output <output location>
            Optionally write help to this location, otherwise default is
            standard output

The option of note is -g/--generator-name (other options are exposed for tooling).

You may pass any generator name (see list command) to -g, and options specific to that generator will be displayed. Some generators have many options, while others may have only a few.

Example:

openapi-generator config-help -g go

Outputs:

CONFIG OPTIONS
	packageName
	    Go package name (convention: lowercase). (Default: openapi)

	hideGenerationTimestamp
	    Hides the generation timestamp when files are generated. (Default: true)

	packageVersion
	    Go package version. (Default: 1.0.0)

	withGoCodegenComment
	    whether to include Go codegen comment to disable Go Lint and collapse by default GitHub in PRs and diffs (Default: false)

	withXml
	    whether to include support for application/xml content type and include XML annotations in the model (works with libraries that provide support for JSON and XML) (Default: false)

	prependFormOrBodyParameters
	    Add form or body parameters to the beginning of the parameter list. (Default: false)

To pass these go client generator-specific options to the generate command for a go client, use the --additional-properties option. See the generate command section for an example.

meta

The meta command creates a new Java class and template files, used for creating your own custom templates.

openapi-generator help meta
NAME
        openapi-generator-cli meta - MetaGenerator. Generator for creating a new
        template set and configuration for Codegen. The output will be based on
        the language you specify, and includes default templates to include.

SYNOPSIS
        openapi-generator-cli meta [(-n <name> | --name <name>)]
                [(-o <output directory> | --output <output directory>)]
                [(-p <package> | --package <package>)] [(-t <type> | --type <type>)]

OPTIONS
        -n <name>, --name <name>
            the human-readable name of the generator

        -o <output directory>, --output <output directory>
            where to write the generated files (current dir by default)

        -p <package>, --package <package>
            the package to put the main class into (defaults to
            org.openapitools.codegen)

        -t <type>, --type <type>
            the type of generator that is created

For an in-depth example of using the meta command, see Customization.

validate

The validate command allows you to validate an input specification, optionally providing recommendations for error fixes or other improvements (if available).

openapi-generator help validate
NAME
        openapi-generator-cli validate - Validate specification

SYNOPSIS
        openapi-generator-cli validate
                (-i <spec file> | --input-spec <spec file>) [--recommend]

OPTIONS
        -i <spec file>, --input-spec <spec file>
            location of the OpenAPI spec, as URL or file (required)

        --recommend

Valid Spec Example (using petstore-v3.0.yaml)

openapi-generator validate -i petstore-v3.0.yaml
Validating spec (petstore-v3.0.yaml)
No validation issues detected.

Invalid Spec Example (using petstore-v3.0-invalid.yaml):

openapi-generator validate -i petstore-v3.0-invalid.yaml
Validating spec (petstore-v3.0-invalid.yaml)
Errors:
	-attribute info is missing

[error] Spec has 1 errors.

completion

Although not documented in the help output, the CLI offers a completion command, which can be used for auto-completion.

This command takes one or more parameters representing the args list you would otherwise pass to openapi-generator. For example:

openapi-generator completion config-help
--named-header
-o
--output
-g
--generator-name
-l
--lang

An example bash completion script can be found in the repo at scripts/openapi-generator-cli-completion.bash.

generate

The generate command is the workhorse of the generator toolset. As such, it has many more options and the previous commands. The options are abbreviated below, but you may expand the full descriptions.

openapi-generator help generate
NAME
        openapi-generator-cli generate - Generate code with the specified
        generator.

SYNOPSIS
        openapi-generator-cli generate
                [(-a <authorization> | --auth <authorization>)]
                [--additional-properties <additional properties>...]
                [--api-package <api package>] [--artifact-id <artifact id>]
                [--artifact-version <artifact version>]
                [(-c <configuration file> | --config <configuration file>)]
                [-D <system properties>...] [--enable-post-process-file]
                [(-g <generator name> | --generator-name <generator name>)]
                [--git-repo-id <git repo id>] [--git-user-id <git user id>]
                [--group-id <group id>] [--http-user-agent <http user agent>]
                (-i <spec file> | --input-spec <spec file>)
                [--ignore-file-override <ignore file override location>]
                [--import-mappings <import mappings>...]
                [--instantiation-types <instantiation types>...]
                [--invoker-package <invoker package>]
                [--language-specific-primitives <language specific primitives>...]
                [--library <library>] [--log-to-stderr]
                [--model-name-prefix <model name prefix>]
                [--model-name-suffix <model name suffix>]
                [--model-package <model package>]
                [(-o <output directory> | --output <output directory>)]
                [--release-note <release note>] [--remove-operation-id-prefix]
                [--reserved-words-mappings <reserved word mappings>...]
                [(-s | --skip-overwrite)] [--skip-validate-spec]
                [(-t <template directory> | --template-dir <template directory>)]
                [--type-mappings <type mappings>...] [(-v | --verbose)]

generate OPTIONS
OPTIONS
        -a <authorization>, --auth <authorization>
            adds authorization headers when fetching the OpenAPI definitions
            remotely. Pass in a URL-encoded string of name:header with a comma
            separating multiple values

        --additional-properties <additional properties>
            sets additional properties that can be referenced by the mustache
            templates in the format of name=value,name=value. You can also have
            multiple occurrences of this option.

        --api-package <api package>
            package for generated api classes

        --artifact-id <artifact id>
            artifactId in generated pom.xml

        --artifact-version <artifact version>
            artifact version in generated pom.xml

        -c <configuration file>, --config <configuration file>
            Path to json configuration file. File content should be in a json
            format {"optionKey":"optionValue", "optionKey1":"optionValue1"...}
            Supported options can be different for each language. Run
            config-help -g {generator name} command for language specific config
            options.

        -D <system properties>
            sets specified system properties in the format of
            name=value,name=value (or multiple options, each with name=value)

        --enable-post-process-file
            enablePostProcessFile

        -g <generator name>, --generator-name <generator name>
            generator to use (see langs command for list)

        --git-repo-id <git repo id>
            Git repo ID, e.g. openapi-generator.

        --git-user-id <git user id>
            Git user ID, e.g. openapitools.

        --group-id <group id>
            groupId in generated pom.xml

        --http-user-agent <http user agent>
            HTTP user agent, e.g. codegen_csharp_api_client, default to
            'OpenAPI-Generator/{packageVersion}}/{language}'

        -i <spec file>, --input-spec <spec file>
            location of the OpenAPI spec, as URL or file (required)

        --ignore-file-override <ignore file override location>
            Specifies an override location for the .openapi-generator-ignore
            file. Most useful on initial generation.

        --import-mappings <import mappings>
            specifies mappings between a given class and the import that should
            be used for that class in the format of type=import,type=import. You
            can also have multiple occurrences of this option.

        --instantiation-types <instantiation types>
            sets instantiation type mappings in the format of
            type=instantiatedType,type=instantiatedType.For example (in Java):
            array=ArrayList,map=HashMap. In other words array types will get
            instantiated as ArrayList in generated code. You can also have
            multiple occurrences of this option.

        --invoker-package <invoker package>
            root package for generated code

        --language-specific-primitives <language specific primitives>
            specifies additional language specific primitive types in the format
            of type1,type2,type3,type3. For example:
            String,boolean,Boolean,Double. You can also have multiple
            occurrences of this option.

        --library <library>
            library template (sub-template)

        --log-to-stderr
            write all log messages (not just errors) to STDOUT. Useful for
            piping the JSON output of debug options (e.g. `-DdebugOperations`)
            to an external parser directly while testing a generator.

        --model-name-prefix <model name prefix>
            Prefix that will be prepended to all model names. Default is the
            empty string.

        --model-name-suffix <model name suffix>
            Suffix that will be appended to all model names. Default is the
            empty string.

        --model-package <model package>
            package for generated models

        -o <output directory>, --output <output directory>
            where to write the generated files (current dir by default)

        --release-note <release note>
            Release note, default to 'Minor update'.

        --remove-operation-id-prefix
            Remove prefix of operationId, e.g. config_getId => getId

        --reserved-words-mappings <reserved word mappings>
            specifies how a reserved name should be escaped to. Otherwise, the
            default _<name> is used. For example id=identifier. You can also
            have multiple occurrences of this option.

        -s, --skip-overwrite
            specifies if the existing files should be overwritten during the
            generation.

        --skip-validate-spec
            Skips the default behavior of validating an input specification.

        -t <template directory>, --template-dir <template directory>
            folder containing the template files

        --type-mappings <type mappings>
            sets mappings between OpenAPI spec types and generated code types in
            the format of OpenaAPIType=generatedType,OpenAPIType=generatedType.
            For example: array=List,map=Map,string=String. You can also have
            multiple occurrences of this option.

        -v, --verbose
            verbose mode

At a minimum, generate requires:

  • -g to specify the generator
  • -o to specify a meaningful output directory (defaults to the current directory!)
  • -i to specify the input OpenAPI document

Examples

The following examples use petstore.yaml.

Additional Properties

Generator-specific options should be passed as --additional-properties:

openapi-generator generate -g go --additional-properties=prependFormOrBodyParameters=true \
    -o out -i petstore.yaml

To pass more than one generator property, these can be combined via comma:

--additional-properties=key1=value1,key2=value2

For the full list of generator-specified parameters, refer to generators docs.

Type Mappings and Import Mappings

Most generators allow for types bound to the OpenAPI Specification's types to be remapped to a user's desired types. Not all type mappings can be reassigned, as some generators define mappings which are tightly coupled to the built-in templates.

If you're not using your own templates with star/glob package imports, you will most likely need to combine --type-mappings and --import-mappings together.

  • --type-mappings Defines the user's target type
  • --import-mappings Informs the template of the type to be imported

Here's how one might change the kotlin-spring server generator's default of OffsetDateTime to LocalDateTime:

openapi-generator generate \
    -i petstore.yaml \
    -g kotlin-spring \
    -o out \
    --additional-properties=library=spring-boot,beanValidations=true,swaggerAnnotations=true,serviceImplementation=true \
    --import-mappings=DateTime=java.time.LocalDateTime \
    --type-mappings=DateTime=java.time.LocalDateTime

NOTE: mappings are applied to DateTime, as this is the representation of the primitive type. See DefaultCodegen.

Target External Models

Sometimes you don't want the codegen to make a model for you--you might want to just include one that already exists in your codebase. Say you already have a User object and want to reuse that, which has a different model package from the other generated files:

First, indicate that the class is already included by default. This will keep the codegen from trying to generate the class.

--language-specific-primitives=Pet

This command line option will tell the generator to consider Pet a "primitive" type.

Next, if the Pet class is a different package, add an --import-mapping to tell the generator to include that import wherever Pet is used:

--import-mappings=Pet=com.yourpackage.models.Pet

Now the codegen will know what to import from that specific package.

NOTE: import-mappings is assigned a key-value pair in this example, but multiple values can be comma-separate. For instance:

--import-mappings=Pet=com.yourpackage.models.Pet,User=com.yourpackage.models.User

Configuration File

Rather than passing generator options in a CSV of --additional-properties, you may also provide the settings via JSON file or YAML file.

For example, one of our typescript samples has the following configuration file:

{
  "npmName": "@swagger/typescript-fetch-petstore",
  "npmVersion": "1.0.0",
  "npmRepository" : "https://skimdb.npmjs.com/registry",
  "snapshot" : false,
  "supportsES6": true
}

These settings can be passed via -c filename. Here, we've saved the above as config.json:

openapi-generator generate -i petstore.yaml -g typescript-fetch -o out \
    -c config.json

Same configuration file can be passed into YAML format having following equivalent content:

npmName: "@swagger/typescript-fetch-petstore"
npmVersion: "1.0.0"
npmRepository: "https://skimdb.npmjs.com/registry"
snapshot: false
supportsES6: true

The settings are passed exactly the same as for config.json. The most important part is the file extension. Supported values are yml or yaml. The name of the file should be config.yml or config.yaml (in our example it will be config.yaml.

openapi-generator generate -i petstore.yaml -g typescript-fetch -o out \
    -c config.yaml