Skip to main content


Module openapi


Important Capabilities

Platform Instance

CLI based Ingestion

Install the Plugin

The openapi source works out of the box with acryl-datahub.

Starter Recipe

Check out the following recipe to get started with ingestion! See below for full configuration options.

For general pointers on writing and running a recipe, see our main recipe guide.

type: openapi
name: test_endpoint # this name will appear in DatHub
swagger_file: classicapi/doc/swagger.json # where to search for the OpenApi definitions
get_token: # optional, if you need to get an authentication token beforehand
request_type: get
url: api/authentication/login?username={username}&password={password}
username: your_username # optional
password: your_password # optional
forced_examples: # optionals
/accounts/groupname/{name}: ['test']
/accounts/username/{name}: ['test']
ignore_endpoints: [/ignore/this, /ignore/that, /also/that_other] # optional, the endpoints to ignore

type: "datahub-rest"
server: 'http://localhost:8080'

Config Details

Note that a . is used to denote nested fields in the YAML recipe.

View All Configuration Options
ignore_endpointsArray of object[]

The dataset metadata should be defined directly in the Swagger file, section ["example"]. If this is not true, the following procedures will take place.


The plugin read the swagger file where the endopints are defined and searches for the ones which accept a GET call: those are the ones supposed to give back the datasets.

For every selected endpoint defined in the paths section, the tool searches whether the medatada are already defined in there. As example, if in your swagger file there is the /api/users/ defined as follows:

tags: [ "Users" ]
operationID: GetUsers
description: Retrieve users data
description: Return the list of users
{"user": "username", "name": "Full Name", "job": "any", "is_active": True}

then this plugin has all the information needed to create the dataset in DataHub.

In case there is no example defined, the plugin will try to get the metadata directly from the endpoint. So, if in your swagger file you have

tags: [ "Colors" ]
operationID: GetDefinedColors
description: Retrieve colors
description: Return the list of colors

the tool will make a GET call to https:/// and parse the response obtained.

Automatically recorded examples

Sometimes you can have an endpoint which wants a parameter to work, like{color}.

Since in the OpenApi specifications the listing endpoints are specified just before the detailed ones, in the list of the paths, you will find


defined before{color}

This plugin is set to automatically keep an example of the data given by the first URL, which with some probability will include an example of attribute needed by the second.

So, if by calling GET to the first URL you get as response:

{"pantone code": 100,
"color": "yellow",

the "color": "yellow" part will be used to complete the second link, which will become:

and this last URL will be called to get back the needed metadata.

Automatic guessing of IDs

If no useful example is found, a second procedure will try to guess a numerical ID. So if we have:


and there is no colorID example already found by the plugin, it will try to put a number one (1) at the parameter place

and this URL will be called to get back the needed metadata.

Config details

Token authentication

If this tool needs to get an access token to interrogate the endpoints, this can be requested. Two methods are available at the moment:

  • 'get' : this requires username/password combination to be present in the url. Note that {username} and {password} are mandatory placeholders. They will be replaced with the true credentials at runtime. Note that username and password will be sent in the request address, so it's unsecure. If your provider allows for the other method, please go for it.
  • 'post' : username and password will be inserted in the body of the POST request

In both cases, username and password are the ones defined in the configuration file.

Getting dataset metadata from forced_example

Suppose you have an endpoint defined in the swagger file, but without example given, and the tool is unable to guess the URL. In such cases you can still manually specify it in the forced_examples part of the configuration file.

As example, if in your swagger file you have

tags: [ "Groups" ]
operationID: GetGroup
description: Retrieve group data
description: Return details about the group

and the plugin did not found an example in its previous calls, so the tool have no idea about what substitute to the {name} part.

By specifying in the configuration file

    forced_examples:  # optionals
/accounts/groupname/{name}: ['test']

the plugin is able to build a correct URL, as follows:

Code Coordinates

  • Class Name: datahub.ingestion.source.openapi.OpenApiSource
  • Browse on GitHub


If you've got any questions on configuring ingestion for OpenAPI, feel free to ping us on our Slack