Skip to content

WEBKNOSSOS REST API

The following HTTP requests may be used to interact with the backend of WEBKNOSSOS. Please keep the server load in mind and avoid sending excessive amounts of requests or huge individual ones.

Authentication

All API routes expect the HTTP header X-Auth-Token to be set to your personal WEBKNOSSOS token. You can find this token in the WEBKNOSSOS menu, directly above “logout”.

Versioning

The API is subject to frequent changes. However, older versions will be supported for some time via prefixed routes. - To access the latest version, call the routes below with only the /api prefix. - To access a specific version, call the routes below with a prefix in the form of /api/v1.

New versions will be documented here, detailing the changes. Note, however, that some changes are not considered to be breaking the API and will not lead to new versions. Such changes include new optional parameters as well as new fields in the responses. The same goes for error message wording.

Current api version is v5

  • Changed in all versions: GET /datasets no longer respects the isEditable parameter.

  • New in v5:

  • The annotation json no longer contains skeletonTracingId and volumeTracingId, but instead a list of annotationLayers, each containing tracingId: String, typ: AnnotationLayerType, name: Option[String].

  • createExplorational now expects a list of layer parameters (typ: AnnotationLayerType, fallbackLayerName: Option[String], resolutionRestrictions: Option[ResolutionRestrictions], name: Option[String])

  • New in v4: /projects routes no longer expect name but now id. The same goes for POST /tasks/list when filtering by project.

  • New in v3: the info and finish requests of annotations now expect an additional timestamp GET parameter that should be set to the time the request is sent (e.g. Date.now()).

  • New in v2: in comparison to v1 the annotation isPublic flag was replaced by a visibility field. This field can have the following values: Public, Internal, Private.

Routes

GET /buildinfo

Returns

JSON object containing information about the version of WEBKNOSSOS

GET /users

List all users for which you have read access

Expects

  • Optional GET parameter isEditable=[BOOLEAN]
  • If true: list only users for which you can change team memberships
  • If false: list only users for which you cannot change team memberships
  • Optional GET parameter isAdmin=[BOOLEAN]
  • If true: list only admins
  • If false: list only non-admins

Returns

  • JSON list of objects containing user information, including team memberships and experiences

GET /user

Returns

  • JSON object containing user information about yourself, including team memberships and experiences

GET /user/tasks

List your own task annotations

Expects

  • Optional GET parameter isFinished=[BOOLEAN]
  • if true: list only finished task annotations
  • if false: list only unfinished task annotations
  • Optional GET parameter limit=[INT]
  • return only the first x results (defaults to 1000)
  • Optional GET parameter pageNumber=[INT]
  • return the results starting at offset limit * pageNumber (defaults to 0)
  • Optional GET parameter includeTotalCount=[BOOLEAN]
    • if true returns the total count of entries (defaults to false)

Returns

  • JSON list of objects containing annotation information about your own task annotations, also including task and task type information
  • total count of task annotations in the HTTP header X-Total-Count if parameter is set

Changes Introduced in v2

  • The annotation objects in the returned JSON contain the visibility field instead of the old isPublic field

GET /user/annotations

List your own explorative annotations

Expects

  • Optional GET parameter isFinished=[BOOLEAN]
  • if true: list only finished (=archived) explorative annotations
  • if false: list only unfinished (=non-archived) explorative annotations
  • Optional GET parameter limit=[INT]
  • return only the first x results (defaults to 1000)
  • Optional GET parameter pageNumber=[INT]
  • return the results starting at offset limit * pageNumber (defaults to 0)
  • Optional GET parameter includeTotalCount=[BOOLEAN]
    • if true returns the total count of entries (defaults to false)

Returns

  • JSON list of objects containing annotation information about your own explorative annotations
  • total count of explorative annotations in the HTTP header X-Total-Count if parameter is set

Changes Introduced in v2

  • The annotation objects in the returned JSON contain the visibility field instead of the old isPublic field

GET /users/:id

Expects

  • In the url: :id id of a user

Returns

  • JSON object containing user information about the selected user, including team memberships and experiences

GET /users/:id/tasks

List the task annotations of a user

Expects

  • In the url: :id id of a user
  • Optional GET parameter isFinished=[BOOLEAN]
  • if true: list only finished task annotations
  • if false: list only unfinished task annotations
  • Optional GET parameter limit=[INT]
  • return only the first x results (defaults to 1000)
  • Optional GET parameter pageNumber=[INT]
  • return the results starting at offset limit * pageNumber (defaults to 0)
  • Optional GET parameter includeTotalCount=[BOOLEAN]
    • if true returns the total count of entries (defaults to false)

Returns

  • JSON list of objects containing annotation information about the task annotations of the user, also including task and task type information
  • total count of task annotations in the HTTP header X-Total-Count if parameter is set
  • total count of task annotations in the HTTP header X-Total-Count if parameter is set

Changes Introduced in v2

  • The annotation objects in the returned JSON contain the visibility field instead of the old isPublic field

GET /users/:id/annotations

List the explorative annotations of a user

Expects

  • In the url: :id id of a user
  • Optional GET parameter isFinished=[BOOLEAN]
  • If true: list only finished (=archived) explorative annotations
  • If false: list only unfinished (=non-archived) explorative annotations
  • Optional GET parameter limit=[INT]
  • return only the first x results (defaults to 1000)
  • Optional GET parameter pageNumber=[INT]
  • return the results starting at offset limit * pageNumber (defaults to 0)
  • Optional GET parameter includeTotalCount=[BOOLEAN]
    • if true returns the total count of entries (defaults to false)

Returns

  • JSON list of objects containing annotation information about the explorative annotations of the user
  • total count of explorative annotations in the HTTP header X-Total-Count if parameter is set

Changes Introduced in v2

  • The annotation objects in the returned JSON contain the visibility field instead of the old isPublic field

GET /teams

List all teams that you can manage

Returns

  • JSON list of objects containing team information

GET /datasets

List all datasets for which you have read access

Expects

  • Optional GET parameter isActive=[BOOLEAN]
  • If true: list only datasets that are active (=imported)
  • If false: list only datasets that are inactive (=non-imported)

Returns

  • JSON list of objects containing dataset information

Notes

  • The list of resolutions for each data layer is always empty in this list, for performance reasons. To get the actual resolutions, please use GET /datasets/:organizationName/:dataSetName (see below).
  • This route used to accept an optional filter parameter isEditable. This is no longer supported. This change affects all api versions.

GET /datasets/:organizationName/:dataSetName

Expects

  • In the url: :organizationName the url-safe name of your organization, e.g. sample_organization or edbdcad7d2033380
  • In the url: :dataSetName the name of the dataset

Returns

  • JSON object containing dataset information

POST /datasets/:organizationName/:dataSetName/createExplorational

Expects

  • In the url: :organizationName the url-safe name of your organization, e.g. sample_organization or edbdcad7d2033380
  • In the url: :dataSetName the name of the dataset
  • In the JSON body: A list of layer parameter objects (typ: AnnotationLayerType, fallbackLayerName: Option[String], resolutionRestrictions: Option[ResolutionRestrictions], name: Option[String]) where the type ResolutionRestrictions is an object with min: Option[Int], max: Option[Int] and AnnotationLayerType is a string with possible values Skeleton, Volume.

Returns

  • JSON object containing annotation information about the newly created annotation, including the assigned id

Changes Introduced in v5

  • Now expects a List of layer parameters, rather than the old format (a single object containing typ: String, fallbackLayerName: Option[String], resolutionRestrictions: Option[ResolutionRestrictions])

GET /datastores

Lists all available datastores

Returns

  • JSON list of objects containing datastore information

POST /datastores

Create a new datastore

Expects

  • JSON object
    • "name" [STRING] name of the datastore
    • "url" [STRING] url from the datastore, used for communication with wk
    • "publicUrl" [STRING] publicly accessible url from the datastore, used for user facing links
    • "key" [STRING] key used to identify the datastore
    • "isScratch" [BOOLEAN] (optional, default: false) whether or not the datastore is hosted on a scratch/experimental environment
    • "isForeign" [BOOLEAN] (optional, default: false) whether or not the datastore belongs to this wk instance or belongs to a foreign wk instance
    • "isConnector" [BOOLEAN] (optional, default: false) whether or not the datastore is a wk-connect instance
    • "allowsUpload" [BOOLEAN] (optional, default: true) whether or not the datastore supports dataset upload via browser

Returns

  • JSON object containing information about the newly created datastore

Note

  • This route is only accessible for administrators.

DELETE /datastores/:name

Deletes a datastore from the wk database

Expects

  • In the url: :name - the name of the datastore which should be deleted

Note

  • This route is only accessible for administrators.

PUT /datastores/:name

Update an existing datastore

Expects

  • In the url: :name - the name of the datastore which should be deleted
  • JSON object
    • "name" [STRING] name of the datastore
    • "url" [STRING] url from the datastore, used for communication with wk
    • "publicUrl" [STRING] publicly accessible url from the datastore, used for user facing links
    • "isScratch" [BOOLEAN] (optional, default: false) whether or not the datastore is hosted on a scratch/experimental environment
    • "isForeign" [BOOLEAN] (optional, default: false) whether or not the datastore belongs to this wk instance or belongs to a foreign wk instance
    • "isConnector" [BOOLEAN] (optional, default: false) whether or not the datastore is a wk-connect instance
    • "allowsUpload" [BOOLEAN] (optional, default: true) whether or not the datastore supports dataset upload via browser

Returns

  • JSON object containing information about the updated datastore

Note

  • This route is only accessible for administrators.

PUT /tracingstores/:name

Update an existing tracingstore

Expects

  • In the url: :name - the name of the tracingstore which should be deleted
  • JSON object
    • "name" [STRING] name of the tracingstore
    • "url" [STRING] url from the tracingstore, used for communication with wk
    • "publicUrl" [STRING] publicly accessible url from the tracingstore, used for user facing links

Returns

  • JSON object containing information about the updated tracingstore

Note

  • This route is only accessible for administrators.

GET /annotations/:typ/:id/info

Expects

  • In the url: :typ – one of Task, Explorational, CompoundTask, CompoundProject, CompoundTaskType
  • In the url: id
  • for Task and Explorational annotations, :id is an annotation id
  • for CompoundTask :id is a task id
  • for CompoundProject :id is a project id
  • for CompoundTaskType :id is a task type id
  • GET parameter timestamp=[INT] timestamp in milliseconds (time the request is sent)

Returns

  • JSON object containing annotation information about the selected annotation

Note

The compound annotations are created as merged from the finished annotations associated with the Task/Project/TaskType. This merging is performed before this info request is answered and can be slow for large numbers of annotations. The merged annotations are then stored in a cache for a few minutes, but not on disk. If requested again within this time, the request will be answered more quickly, but newly finished annotations will not be included yet. This cache is shared between this info request and the download request.

Changes Introduced in v3

  • Expects additional GET parameter timestamp=[INT] timestamp in milliseconds (time the request is sent)

Changes Introduced in v2

  • The returned JSON contains the visibility field instead of the isPublic flag

GET /annotations/:typ/:id/download

Download an annotation as NML/ZIP

Expects

  • In the url: :typ and :id as described above under GET /annotations/:typ/:id/info

Returns

  • As chunked file stream:
  • In case of a single skeleton-only annotation:
    • A single NML file
  • In case of a single annotation with volume layers
    • A ZIP file containing both an NML file and data zip files with the volume data for each volume layer. The data zip file names are referenced in the <volume> tags in the NML file.
  • In case of compound downloads (CompoundTask/CompoundProject/CompoundTaskType):
    • A ZIP file containing individual NML files for all associated annotations

Note

  • The volume data zip file name used to always be data.zip, when wk supported only a single volume layer per annotation. Please use the file names referenced in the <volume> tags in the NML file instead.

POST /annotations/upload

Upload NML(s) or ZIP(s) of NML(s) to create a new explorative annotation

Expects

  • As file attachment: any number of NML files or ZIP files containing NMLs, optionally with at most one volume data ZIP referenced from an NML in a ZIP
  • As form parameter: createGroupForEachFile [String] should be one of "true" or "false"
  • If "true": in merged annotation, create tree group wrapping the trees of each file
  • If "false": in merged annotation, rename trees with the respective file name as prefix

Returns

  • JSON object containing annotation information about the newly created annotation, including the assigned id

POST /annotations/:typ/:id/duplicate

Duplicate an annotation (“copy to my account”)

Expects

  • In the url: :typ and :id as described above under GET /annotations/:typ/:id/info

Returns

  • JSON object containing annotation information about the newly created (duplicated) annotation, including the assigned id

Changes Introduced in v2

  • The returned JSON contains the visibility field instead of the isPublic flag

PATCH /annotations/:typ/:id/edit

Edit metadata of an annotation

Expects

  • In the url: :typ – one of Task, Explorational
  • In the url: :id an annotation id
  • JSON object with optional fields
  • "name" [STRING] new name for the annotation
  • "description" [STRING] new description for the annotation
  • "visibility" ["Public" | "Internal" | "Private"] new visibility for the annotation
  • "tags" [JSON LIST OF STRINGS] list of tags for the annotation

Returns

  • JSON object containing annotation information about the edited annotation

Changes Introduced in v2

  • The request and returned JSON contain the isPublic [BOOLEAN] flag instead of the visibility field

PATCH /annotations/:typ/:id/finish

Expects

  • In the url: :typ – one of Task, Explorational
  • In the url: :id an annotation id
  • GET parameter timestamp=[INT] timestamp in milliseconds (time the request is sent)

Returns

  • JSON object containing annotation information about the finished annotation

Changes Introduced in v3

  • Expects additional GET parameter timestamp=[INT] timestamp in milliseconds (time the request is sent)

Changes Introduced in v2

  • The returned JSON contains the visibility field instead of the isPublic flag

PATCH /annotations/:typ/:id/reopen

Expects

  • In the url: :typ – one of Task, Explorational
  • In the url: :id an annotation id

Returns

  • JSON object containing annotation information about the reopened annotation

Changes Introduced in v2

  • The returned JSON contains the visibility field instead of the isPublic flag

PUT /annotations/Task/:id/reset

Reset a task annotation to its base state

Expects

  • In the url: :id an annotation id

Returns

  • JSON object containing annotation information about the reset task annotation

Changes Introduced in v2

  • The returned JSON contains the visibility field instead of the isPublic flag

POST /annotations/:typ/:id/merge/:mergedTyp/:mergedId

Merge two annotations, creating a new explorative.

Expects

  • In the url: :typ and :id as described above under GET /annotations/:typ/:id/info
  • In the url: :mergedTyp and :mergedId as described above under GET /annotations/:typ/:id/info

Returns

  • JSON object containing annotation information about the merged annotation

Changes Introduced in v2

  • The returned JSON contains the visibility field instead of the isPublic flag

POST /tasks

Create tasks without attached NML files

Expects

  • JSON list of objects. Each has these fields:
  • "taskTypeId" [STRING]
  • "neededExperience" [JSON OBJECT] with these fields:
    • "domain" [STRING]
    • "value" [INT]
  • "openInstances" [INT]
  • "projectName" [STRING]
  • "scriptId" (optional) [STRING]
  • "boundingBox" (optional) [JSON OBJECT] with these fields:
    • "topLeft" [JSON LIST OF THREE INTS]
    • "width" [INT]
    • "height" [INT]
    • "depth" [INT]
  • "dataSet" [STRING] name of the dataset
  • "editPosition" [JSON LIST OF THREE INTS] starting coordinates
  • "editRotation" [JSON LIST OF THREE INTS] starting rotation
  • "creationInfo" (optional) [STRING] identifier to match with task creation result report
  • "description" (optional) [STRING]

Returns

  • JSON list with status reports about the created tasks, including their ids

Note

This request will also return status code 200 if some tasks could not be created. Failed tasks are instead encoded in the JSON response.

POST /tasks/createFromFiles

Create tasks with attached NML files

Expects

  • Form data field "formJSON"
  • JSON object with these fields:
    • ""
    • "taskTypeId" [STRING]
    • "neededExperience" [JSON OBJECT] with these fields:
    • "domain" [STRING]
    • "value" [INT]
    • "openInstances" [INT]
    • "projectName" [STRING]
    • "scriptId" (optional) [STRING]
    • "boundingBox" (optional) [JSON OBJECT] with these fields:
    • "topLeft" [JSON LIST OF THREE INTS]
    • "width" [INT]
    • "height" [INT]
    • "depth" [INT]
  • Attached NML files

Returns

  • JSON list with status reports about the created tasks, including their ids

Note

For each attached NML file, one task is created. This request will also return status code 200 if some tasks could not be created. Failed tasks are instead encoded in the JSON response

POST /tasks/list

List tasks matching search criteria

Expects

  • JSON object with four optional fields:
  • "user" [STRING] show only tasks on which the user with this id has worked
  • "project" [STRING] show only tasks of the project with this id
  • "ids" [JSON LIST OF STRINGS] show only tasks with these ids
  • "tasktype" [STRING] show only tasks matching the task type with this id
  • "random" [BOOLEAN] if true, return a randomized subset of the results, rather than the first 1000 matches in the database

Returns

  • JSON list of objects containing task information
  • Note that a maximum of 1000 results is returned

Changes Introduced in v4

  • The "project" field in the JSON object is no longer its name but instead its id.

GET /tasks/:id

Expects

  • In the url: :id id of a task

Returns

  • JSON object containing task information

DELETE /tasks/:id

Delete one task and all its annotations

Expects

  • In the url: :id id of a task

PUT /tasks/:id

Update the number of open instances for a task

Expects

  • In the url: :id id of a task
  • JSON object like the ones in the list in POST /tasks (only "openInstances" is changed, though)

Returns

  • JSON object containing task information on the updated task

GET /tasks/:id/annotations

List annotations of a task

Expects

  • In the url: :id a task id

Returns

  • JSON list of objects containing annotation information on the annotations of the task
  • Cancelled annotations are not returned

Changes Introduced in v2

  • The annotation objects in the returned JSON contain the visibility flag instead of the isPublic field

GET /projects

List all projects for which you have read access

Returns

JSON list of objects containing project information

GET /projects/assignments

List all projects for which you have read access, annotated with the number of open task instances

Returns

JSON list of objects containing project information, with additional field numberOfOpenAssignments

POST /projects

Create a new project

Expects

  • JSON object with the following fields:
  • "name" [STRING]
  • "team" [STRING] id of a team
  • "priority" [INT]
  • "paused" [BOOLEAN]
  • "expectedTime" [INT] time limit
  • "owner" [STRING] id of a user
  • "isBlacklistedFromReport" [BOOLEAN]

Returns

JSON object containing project information about the newly created project, including the assigned id

GET /projects/:id

Expects

  • In the url: :id id of a project

Returns

  • JSON object containing project information about the selected project

Changes Introduced in v4

  • The request no longer expects name in the url, but instead id

DELETE /projects/:id

Delete a project and all its tasks and annotations

Expects

  • In the url: :id id of a project

Changes Introduced in v4

  • The request no longer expects name in the url, but instead id

PUT /projects/:id

Update a project

Expects

  • JSON object just like in POST /projects

Returns

  • JSON object containing project information about the updated project

Changes Introduced in v4

  • The request no longer expects name in the url, but instead id

GET /projects/:id/tasks

List all tasks of a project

Expects

  • In the url: :id id of a project
  • Optional GET parameter limit=[INT]
  • return only the first x results (defaults to infinity)
  • Optional GET parameter pageNumber=[INT]
  • return the results starting at offset limit * pageNumber (defaults to 0)
  • Optional GET parameter includeTotalCount=[BOOLEAN]
    • if true returns the total count of entries (defaults to false)

Returns

  • JSON list of objects containing task information
  • total count of tasks in the HTTP header X-Total-Count if parameter is set

Note

  • For smoother backwards compatibility, the limit defaults to infinity. However, to ease server load and improve response time, we suggest using a limit of 1000

Changes Introduced in v4

  • The request no longer expects name in the url, but instead id

PATCH /projects/:id/incrementEachTasksInstances

Increment the open instances for each task of a project.

Expects

  • In the url: id [STRING] id of a project
  • Optional GET parameter delta=[INT] number of additional instances for each task (defaults to 1)

Returns

  • JSON object containing project information about the updated project, with additional field numberOfOpenAssignments

Changes Introduced in v4

  • The request no longer expects name in the url, but instead id

PATCH /projects/:id/pause

Pause a project (no tasks will be assigned until resumed)

Expects

  • In the url: :id [STRING] id of a project

Returns

  • JSON object containing project information about the updated project

Changes Introduced in v4

  • The request no longer expects name in the url, but instead id

PATCH /projects/:id/resume

Resume a paused project

Expects

  • In the url: :id [STRING] id of a project

Returns

  • JSON object containing project information about the updated project

Changes Introduced in v4

  • The request no longer expects name in the url, but instead id