voplica.api.ai.dataset

The protobuf package voplica.api.ai.dataset: 4 services, 144 messages and 2 enums, defined in voplica/api/ai/dataset/dataset.proto, voplica/api/ai/dataset/dataset_generation.proto, voplica/api/ai/dataset/dataset_preparation.proto, voplica/api/ai/dataset/dataset_subject.proto.

Services

DatasetService

Each method is called as POST https://api.voplica.com/voplica.api.ai.dataset.DatasetService/<Method> with its request message as JSON and your API key (Authentication); a method marked Anonymous allowed also answers without one. See the JSON API.

Method	Request	Response	Description
CreateDataset	CreateDatasetRequest	CreateDatasetResponse
GetDataset	GetDatasetRequest	GetDatasetResponse	Anonymous allowed.
GetMultiDataset	GetMultiDatasetRequest	GetMultiDatasetResponse	Anonymous allowed. Batch lookup (FE name hydration, e.g. the creator earnings pages). Visibility-filtered like GetDataset; missing/inaccessible ids are silently omitted rather than erroring.
GetDatasetMetadata	GetDatasetMetadataRequest	GetDatasetMetadataResponse	Returns owner-only metadata (preparation lifecycle hints, description, etc.). Access restricted to the dataset owner.
GetDatasetLikenessDisclosure	GetDatasetLikenessDisclosureRequest	GetDatasetLikenessDisclosureResponse	Anonymous allowed.
UpdateDataset	UpdateDatasetRequest	UpdateDatasetResponse
DeleteDataset	DeleteDatasetRequest	DeleteDatasetResponse
CloneDataset	CloneDatasetRequest	CloneDatasetResponse	Clone-to-edit for public datasets (and duplication for owners). See CloneDatasetRequest.
AttachAssetToDataset	AttachAssetToDatasetRequest	AttachAssetToDatasetResponse	Attaches a source asset to the dataset.
UpdateDatasetAssetAutoCaptionHint	UpdateDatasetAssetAutoCaptionHintRequest	UpdateDatasetAssetAutoCaptionHintResponse	Updates the per-asset auto-caption hint on an existing dataset_assets membership row.
DetachAssetFromDataset	DetachAssetFromDatasetRequest	DetachAssetFromDatasetResponse
GetAttachedAssetsToDataset	GetAttachedAssetsToDatasetRequest	GetAttachedAssetsToDatasetResponse	Anonymous allowed. Works only when the accessing user has dataset download permission (owner OR public + download_allowed).
AddDatasetPreviewAsset	AddDatasetPreviewAssetRequest	AddDatasetPreviewAssetResponse
DeleteDatasetPreviewAsset	DeleteDatasetPreviewAssetRequest	DeleteDatasetPreviewAssetResponse
ListDatasets	ListDatasetsRequest	ListDatasetsResponse	Anonymous allowed.
ListDatasetAssets	ListDatasetAssetsRequest	ListDatasetAssetsResponse	Anonymous allowed. Lists source dataset assets. Prepared artifacts are exposed via the dataset preparation service.
GetDatasetPreviewAssets	GetDatasetPreviewAssetsRequest	GetDatasetPreviewAssetsResponse	Anonymous allowed.

DatasetGenerationService

Each method is called as POST https://api.voplica.com/voplica.api.ai.dataset.DatasetGenerationService/<Method> with its request message as JSON and your API key (Authentication); a method marked Anonymous allowed also answers without one. See the JSON API.

Method	Request	Response	Description
GenerateDatasetImages	GenerateDatasetImagesRequest	GenerateDatasetImagesResponse	Validates the target dataset (owned by the caller, `IMAGE_ASSET`) or creates it from the inline spec, validates character access and explicit look existence, persists the run, and executes wrapper `workflow.ai.dataset.generation.v1` AS THE USER (so it is billed like every other ai workflow). A trigger failure marks the run FAILED and is reported to the caller.
GetDatasetGenerationRun	GetDatasetGenerationRunRequest	GetDatasetGenerationRunResponse
ListDatasetGenerationRuns	ListDatasetGenerationRunsRequest	ListDatasetGenerationRunsResponse
CancelDatasetGenerationRun	CancelDatasetGenerationRunRequest	CancelDatasetGenerationRunResponse	Cancels the workflow for an in-flight generation run (engine `stopExecution` as the user). Restricted to the dataset owner OR the user who started the run.

DatasetPreparationService

Each method is called as POST https://api.voplica.com/voplica.api.ai.dataset.DatasetPreparationService/<Method> with its request message as JSON and your API key (Authentication); a method marked Anonymous allowed also answers without one. See the JSON API.

Method	Request	Response	Description
CreateDatasetPreparationRecipe	CreateDatasetPreparationRecipeRequest	CreateDatasetPreparationRecipeResponse
GetDatasetPreparationRecipe	GetDatasetPreparationRecipeRequest	GetDatasetPreparationRecipeResponse	Anonymous allowed.
UpdateDatasetPreparationRecipe	UpdateDatasetPreparationRecipeRequest	UpdateDatasetPreparationRecipeResponse
DeleteDatasetPreparationRecipe	DeleteDatasetPreparationRecipeRequest	DeleteDatasetPreparationRecipeResponse
ListDatasetPreparationRecipes	ListDatasetPreparationRecipesRequest	ListDatasetPreparationRecipesResponse	Anonymous allowed.
PlanDatasetPreparation	PlanDatasetPreparationRequest	PlanDatasetPreparationResponse	Returns what would be reused, prepared, and captioned before starting preparation. Read-only. Restricted to the dataset owner.
PrepareDataset	PrepareDatasetRequest	PrepareDatasetResponse	Plans preparation, find-or-creates the dataset’s `DatasetPreparation` for the recipe, persists a run row (with its frozen work items), and starts the workflow when work is needed. Restricted to the dataset owner.
CancelDatasetPreparation	CancelDatasetPreparationRequest	CancelDatasetPreparationResponse	Cancels the workflow for an in-flight preparation run. Restricted to the dataset owner OR the user who started the run.
GetDatasetPreparation	GetDatasetPreparationRequest	GetDatasetPreparationResponse	Anonymous allowed. Visibility follows source dataset visibility plus the dataset’s download_allowed flag.
ListDatasetPreparations	ListDatasetPreparationsRequest	ListDatasetPreparationsResponse	Anonymous allowed.
UpdateDatasetPreparationConfig	UpdateDatasetPreparationConfigRequest	UpdateDatasetPreparationConfigResponse
DeleteDatasetPreparation	DeleteDatasetPreparationRequest	DeleteDatasetPreparationResponse
GetPreparedItem	GetPreparedItemRequest	GetPreparedItemResponse	Anonymous allowed.
ListPreparedItems	ListPreparedItemsRequest	ListPreparedItemsResponse	Anonymous allowed.
OverridePreparedItemFile	OverridePreparedItemFileRequest	OverridePreparedItemFileResponse
ResetPreparedItemFile	ResetPreparedItemFileRequest	ResetPreparedItemFileResponse
UpsertPreparedItemCaption	UpsertPreparedItemCaptionRequest	UpsertPreparedItemCaptionResponse
DeletePreparedItemCaption	DeletePreparedItemCaptionRequest	DeletePreparedItemCaptionResponse
ResetPreparedItemCaption	ResetPreparedItemCaptionRequest	ResetPreparedItemCaptionResponse
SetPreparedItemExcluded	SetPreparedItemExcludedRequest	SetPreparedItemExcludedResponse
GenerateDatasetItemCaptions	GenerateDatasetItemCaptionsRequest	GenerateDatasetItemCaptionsResponse
GetDatasetPreparationSnapshot	GetDatasetPreparationSnapshotRequest	GetDatasetPreparationSnapshotResponse	Anonymous allowed.
ListDatasetPreparationSnapshots	ListDatasetPreparationSnapshotsRequest	ListDatasetPreparationSnapshotsResponse	Anonymous allowed.
ListDatasetPreparationSnapshotItems	ListDatasetPreparationSnapshotItemsRequest	ListDatasetPreparationSnapshotItemsResponse	Anonymous allowed. The trainer’s one-stream read of complete frozen training examples.
GetDatasetPreparationRun	GetDatasetPreparationRunRequest	GetDatasetPreparationRunResponse
ListDatasetPreparationRuns	ListDatasetPreparationRunsRequest	ListDatasetPreparationRunsResponse
ListDatasetPreparationRunItems	ListDatasetPreparationRunItemsRequest	ListDatasetPreparationRunItemsResponse

DatasetSubjectService

Each method is called as POST https://api.voplica.com/voplica.api.ai.dataset.DatasetSubjectService/<Method> with its request message as JSON and your API key (Authentication); a method marked Anonymous allowed also answers without one. See the JSON API.

Method	Request	Response	Description
DiscoverDatasetSubjects	DiscoverDatasetSubjectsRequest	DiscoverDatasetSubjectsResponse	Freezes a run + work items and starts `ai.dataset_subject_discovery_workflow` when work exists (no-work -> run returned already COMPLETED). Dataset owner only.
GetDatasetSubjectDiscoveryRun	GetDatasetSubjectDiscoveryRunRequest	GetDatasetSubjectDiscoveryRunResponse
ListDatasetSubjectDiscoveryRuns	ListDatasetSubjectDiscoveryRunsRequest	ListDatasetSubjectDiscoveryRunsResponse
ListDatasetSubjectDiscoveryRunItems	ListDatasetSubjectDiscoveryRunItemsRequest	ListDatasetSubjectDiscoveryRunItemsResponse	Same audience as the run: its owner.
CancelDatasetSubjectDiscovery	CancelDatasetSubjectDiscoveryRequest	CancelDatasetSubjectDiscoveryResponse
CreateDatasetSubject	CreateDatasetSubjectRequest	CreateDatasetSubjectResponse
GetDatasetSubject	GetDatasetSubjectRequest	GetDatasetSubjectResponse	Anonymous allowed.
UpdateDatasetSubject	UpdateDatasetSubjectRequest	UpdateDatasetSubjectResponse
MergeDatasetSubjects	MergeDatasetSubjectsRequest	MergeDatasetSubjectsResponse
DeleteDatasetSubject	DeleteDatasetSubjectRequest	DeleteDatasetSubjectResponse
ListDatasetSubjects	ListDatasetSubjectsRequest	ListDatasetSubjectsResponse	Anonymous allowed.
AssignDatasetSubjectInstances	AssignDatasetSubjectInstancesRequest	AssignDatasetSubjectInstancesResponse
ListDatasetSubjectInstances	ListDatasetSubjectInstancesRequest	ListDatasetSubjectInstancesResponse	Anonymous allowed.

Messages

AddDatasetPreviewAssetRequest

Field	Number	Type	Label	Description
dataset_id	1	`string`
preview_asset_id	2	`string`

AddDatasetPreviewAssetResponse

Field	Number	Type	Label	Description
created_asset_id	1	`string`

AssetFilesData

Field	Number	Type	Label	Description
source_asset_groups	1	voplica.types.file.AssetFileVariantGroups
auto_caption_hint	2	`string`		Auto-caption hint stored on the dataset_assets membership row, if any (feeds auto-caption generation; not a training caption).

AttachAssetToDatasetRequest

Field	Number	Type	Description
dataset_id	1	`string`
asset_id	2	`string`
auto_caption_hint	3	`string`	Optional per-asset hint that feeds AUTO-CAPTION generation (e.g. “John, a 30yo man with brown hair”). It is NOT itself a training caption and does NOT affect the visual `source_item_fingerprint` or preparation reuse - training captions live in the caption layer (used when auto-generating item captions). Leave empty to attach without a hint.

AttachAssetToDatasetResponse

No fields.

CloneDatasetRequest

Clones a dataset into the CALLER’S space: memberships + auto_caption_hints, subjects + instances (+ embeddings), preparations + items + captions - all as REFERENCES to the same content-addressed files (no file bytes are copied). Process history (runs, snapshots, training lineage) is NOT cloned. Allowed for the owner, or for any user when the dataset is PUBLIC with download_allowed = true - this is THE modification path for non-owners (the identity/prepared layers of someone else’s dataset are read-only).

Field	Number	Type	Label	Description
source_dataset_id	1	`string`
name	2	`string`	optional	Name of the new dataset. Defaults to the source name + “ (copy)“.

CloneDatasetResponse

Field	Number	Type	Label	Description
dataset_id	1	`string`		The new dataset owned by the caller.
workflow_execution_id	2	`string`	optional	Populated when the clone runs asynchronously (large datasets); poll the dataset’s metadata for readiness. Empty when the clone completed synchronously.

CreateDatasetRequest

Field	Number	Type	Label	Description
name	1	`string`
asset_type	2	voplica.types.base.AssetType		Asset type of the assets that can be attached to the dataset. Immutable after creation.
description	3	`string`	optional
search_and_use_visibility_type	4	voplica.types.base.VisibilityType		If `false` - only the owner can see or use the dataset. If `true` - anyone can use the dataset to train their models. The dataset is searchable and exposes basic info and previews. Files can only be downloaded when `download_allowed` is `true`.
download_allowed	5	`bool`		If `true` anyone who can use the dataset will also be able to download dataset files. If `false` nobody (except the owner) can download dataset files.
revenue_share	6	`bool`		When enabled, the dataset becomes eligible for revenue sharing. The owner earns part of the commission Voplica earns from model trainings that used the dataset. Trainers are charged the same regardless of revenue-share status.
best_for_category	7	voplica.types.ai.common.SubjectCategory
mpa_rating	8	voplica.types.base.MPARating
based_on_likeness_category	9	voplica.types.base.LikenessCategory		Required when the dataset contains a likeness of a real person or copyright-protected content.
likeness_granted_permission_verification	10	voplica.types.base.LikenessGrantedPermissionVerification

CreateDatasetResponse

Field	Number	Type	Label	Description
dataset_id	1	`string`

DatasetAssetsFetchingOptions

Field	Number	Type	Label	Description
source_asset_retriever	1	voplica.types.file.FilesRetriever	optional

DatasetFieldsFetchingOptions

Field	Number	Type	Label	Description
default_preview_retriever	1	voplica.types.file.FilesRetriever	optional	If provided, returns the file variant groups for the dataset’s default preview asset.

DeleteDatasetPreviewAssetRequest

Field	Number	Type	Label	Description
dataset_id	1	`string`
preview_asset_id	2	`string`

DeleteDatasetPreviewAssetResponse

No fields.

DeleteDatasetRequest

Field	Number	Type	Label	Description
dataset_id	1	`string`

DeleteDatasetResponse

No fields.

DetachAssetFromDatasetRequest

Field	Number	Type	Label	Description
dataset_id	1	`string`
asset_id	2	`string`

DetachAssetFromDatasetResponse

No fields.

GetAttachedAssetsToDatasetRequest

Field	Number	Type	Label	Description
dataset_id	1	`string`
asset_ids	2	`string`	repeated	Up to 100 asset ids. Ids not part of the dataset are silently dropped from the response. Useful as a membership check.
fields_fetching_options	3	DatasetAssetsFetchingOptions

GetAttachedAssetsToDatasetResponse

Field	Number	Type	Label	Description
dataset_assets	1	AssetFilesData	repeated

GetDatasetLikenessDisclosureRequest

Field	Number	Type	Label	Description
dataset_id	1	`string`

GetDatasetLikenessDisclosureResponse

Field	Number	Type	Label	Description
likeness_disclosure	1	voplica.types.base.LikenessDisclosure

GetDatasetMetadataRequest

Field	Number	Type	Label	Description
dataset_id	1	`string`

GetDatasetMetadataResponse

Field	Number	Type	Label	Description
dataset_metadata	1	voplica.types.ai.dataset.DatasetMetadata

GetDatasetPreviewAssetsRequest

Field	Number	Type	Label	Description
dataset_id	1	`string`
asset_id_filter	2	`string`	optional	Optionally filter by asset id. If unset, all preview assets are returned.
previews_retriever	3	voplica.types.file.FilesRetriever	optional

GetDatasetPreviewAssetsResponse

Field	Number	Type	Label	Description
all_previews	1	voplica.types.file.AssetFileVariantGroups	repeated	0 to 20 preview assets. `file_variant_groups` are returned only if requested explicitly.

GetDatasetRequest

Field	Number	Type	Label	Description
dataset_id	1	`string`
fields_fetching_options	2	DatasetFieldsFetchingOptions

GetDatasetResponse

Field	Number	Type	Label	Description
dataset	1	voplica.types.ai.dataset.Dataset

GetMultiDatasetRequest

Field	Number	Type	Label	Description
ids	1	`string`	repeated	<=100

GetMultiDatasetResponse

Field	Number	Type	Label	Description
datasets	1	voplica.types.ai.dataset.Dataset	repeated	Visibility-filtered: missing/inaccessible ids are silently omitted.

ListDatasetAssetsFilters

Field	Number	Type	Label	Description
dataset_id	1	`string`

ListDatasetAssetsRequest

Field	Number	Type	Label
filters	1	ListDatasetAssetsFilters	optional
limit	2	`int64`
sort	3	ListDatasetAssetsSorting
fields_fetching_options	4	DatasetAssetsFetchingOptions
cursor	10	`string`	optional
before	11	`bool`
check_more_available	12	`bool`

ListDatasetAssetsResponse

Field	Number	Type	Label
assets_with_files	1	AssetFilesData	repeated
next_token	2	`string`	optional
prev_token	3	`string`	optional
more_available	4	`bool`	optional

ListDatasetAssetsSorting

Field	Number	Type	Label	Description
sort_key	1	DatasetAssetSortKey
asc_sort	2	`bool`

ListDatasetsFilters

Field	Number	Type	Label	Description
owner_id	1	`string`	optional
asset_type	2	voplica.types.base.AssetType	optional
visibility	3	voplica.types.base.VisibilityType	optional
category	4	voplica.types.ai.common.SubjectCategory	optional
mpa_filter	5	MPAFilter	optional
download_allowed	6	`bool`	optional
query	7	`string`	optional	Case-insensitive dataset NAME substring search (‘%term%’), composed with every other filter above. Server-side guards: blank after trim = no filter; 1 char after trim = INVALID_ARGUMENT (min 2 chars); capped at 100 chars. When set, results are ordered by created_at (keyset) regardless of the requested trending/last-file sort key — mirrors the Voices/Characters `query` relation semantics.

ListDatasetsRequest

Field	Number	Type	Label	Description
filters	1	ListDatasetsFilters	optional
limit	2	`int64`
sort	3	ListDatasetsSorting
fields_fetching_options	4	DatasetFieldsFetchingOptions
cursor	10	`string`	optional	To continue pagination use `next_token` with `before = false` to fetch the next page or `prev_token` with `before = true` to fetch the previous page.
before	11	`bool`
check_more_available	12	`bool`		When true, the response indicates whether more items are available (uses `limit + 1`).

ListDatasetsResponse

Field	Number	Type	Label
datasets	1	voplica.types.ai.dataset.Dataset	repeated
next_token	2	`string`	optional
prev_token	3	`string`	optional
more_available	4	`bool`	optional

ListDatasetsSorting

Field	Number	Type	Label	Description
sort_key	1	DatasetSortKey
asc_sort	2	`bool`

MPAFilter

Field	Number	Type	Label	Description
max_mpa_rating	1	voplica.types.base.MPARating	oneof `filter`	Return datasets with this mpa rating or below.
nsfw	2	`bool`	oneof `filter`	`true` means NSFW only (X or XXX). `false` means SFW only (G/PG/PG_13/R).

UpdateDatasetAssetAutoCaptionHintRequest

Field	Number	Type	Description
dataset_id	1	`string`
asset_id	2	`string`
auto_caption_hint	3	`string`	New auto-caption hint. Empty clears any existing hint. This only influences future auto-caption generation; it does NOT change the `source_item_fingerprint` and never invalidates prepared visual artifacts.

UpdateDatasetAssetAutoCaptionHintResponse

No fields.

UpdateDatasetRequest

Field	Number	Type	Label	Description
dataset_id	1	`string`
name	2	`string`	optional
description	3	`string`	optional
visibility	4	voplica.types.base.VisibilityType	optional
download_allowed	5	`bool`	optional
revenue_share	6	`bool`	optional
best_for_category	7	voplica.types.ai.common.SubjectCategory	optional
default_preview	8	voplica.types.file.AssetSelectionStrategy	optional
mpa_rating	9	voplica.types.base.MPARating	optional	The owner may raise or lower the rating; UNRECOGNIZED is refused; an explicit MPA_RATING_UNSPECIFIED on an already-RATED dataset is refused too (a no-op on an unrated one).

UpdateDatasetResponse

Field	Number	Type	Label	Description
dataset	1	voplica.types.ai.dataset.Dataset

CancelDatasetGenerationRunRequest

Field	Number	Type	Label	Description
run_id	1	`string`

CancelDatasetGenerationRunResponse

Field	Number	Type	Label	Description
job_status	1	voplica.types.base.JobStatus		Will typically be CANCELLING immediately, transitioning to CANCELLED once the worker acknowledges the cancellation. A run whose engine execution is already gone (a zombie) is marked CANCELLED directly.

GenerateDatasetImagesRequest

Starts a synthetic dataset generation run. Restricted to the dataset owner (an existing dataset_id must be owned by the caller AND be an IMAGE_ASSET dataset) and to a character the caller may use.

Field	Number	Type	Label	Description
dataset_id	1	`string`	oneof `target`	An EXISTING dataset. Must be owned by the caller and be of type `IMAGE_ASSET`.
inline_new_dataset	2	NewDatasetSpec	oneof `target`	Create a new `IMAGE_ASSET` dataset owned by the caller and generate into it. The created dataset id is returned on the run (`DatasetGenerationRun.dataset_id`).
character_id	3	`string`		The character every shot is identity-conditioned on.
look_ids	4	`string`	repeated
params	5	voplica.types.ai.dataset.DatasetGenerationParams		Run tuning. NOTE `params.refine` DEFAULTS TO TRUE when absent - see `voplica.types.ai.dataset.DatasetGenerationParams.refine`; it must be read presence-aware.
privacy_mode	6	voplica.types.workflow.WorkflowPrivacyMode

GenerateDatasetImagesResponse

Field	Number	Type	Label	Description
run	1	voplica.types.ai.dataset.DatasetGenerationRun		The persisted run, already PENDING (or FAILED when the workflow trigger itself failed). Its `dataset_id` is the created dataset for the inline-new-dataset path, and its `root_workflow_execution_id` is the execution the user is billed for.

GetDatasetGenerationRunRequest

Field	Number	Type	Label	Description
run_id	1	`string`

GetDatasetGenerationRunResponse

Field	Number	Type	Label	Description
run	1	voplica.types.ai.dataset.DatasetGenerationRun

ListDatasetGenerationRunsRequest

Field	Number	Type	Label	Description
dataset_id	1	`string`		Runs of ONE dataset - the Dataset page’s generation surface.
limit	2	`int64`		Max amount of elements per page
cursor	10	`string`	optional	to continue pagination use `next_token` with `before = false` to fetch the next page or `prev_token` with `before = true` to fetch the previous page.
before	11	`bool`		Used only with `cursor`. If `false` it means fetch data after cursor. If `true` it means fetch data before cursor.
check_more_available	12	`bool`		Check limit + 1 to know if there is more data available. By default it is `false`. Meaning that `more_available` flag won’t be returned.

ListDatasetGenerationRunsResponse

Field	Number	Type	Label	Description
runs	1	voplica.types.ai.dataset.DatasetGenerationRun	repeated	Ordered by `created_at` (newest first).
next_token	2	`string`	optional	Token to pass in the next query’s Pagination directive to go to the next page It will be missing if this request didn’t return any data (i.e. no more data available)
prev_token	3	`string`	optional	Token to pass in the next query’s Pagination directive to go to the previous page It will be missing if this request didn’t return any data (i.e. no more data available)
more_available	4	`bool`	optional	Returns if there is more data available (i.e. checks `limit + 1`). Returned only if `check_more_available` is set to `true`

NewDatasetSpec

Inline spec for creating the target dataset as part of GenerateDatasetImages - the “generate into a NEW dataset” path of the dialog, so the user does not have to create an empty dataset first.

Deliberately BOUNDED and deliberately PARTIAL: - asset_type is NOT a field. It is forced to IMAGE_ASSET; a generation run can only ever write images, and the same rule is validated for an existing dataset_id.

the likeness fields (based_on_likeness_category, likeness_granted_permission_verification) are NOT here. A caller that must declare a real-person likeness calls DatasetService.CreateDataset first and passes the resulting dataset_id on the target oneof instead. Keeping them off the inline path means the inline path can never under-declare them by omission. - revenue_share (CreateDatasetRequest field 6) is NOT here either, and for the same shape of reason: it is a MONETARY opt-in, and defaulting it from an inline convenience path would enrol a dataset in revenue sharing as a side effect of generating images - a commitment the caller never expressed. Omitted means it takes CreateDatasetRequest’s own default (false); a caller who wants revenue sharing creates the dataset explicitly via DatasetService.CreateDataset and passes dataset_id on the target oneof, exactly as with the likeness fields above. Everything else defaults exactly as CreateDatasetRequest does.

Field	Number	Type	Label	Description
name	1	`string`
description	2	`string`	optional
search_and_use_visibility_type	3	voplica.types.base.VisibilityType		If `false` - only the owner can see or use the dataset.
download_allowed	4	`bool`		If `true` anyone who can use the dataset will also be able to download its files.
mpa_rating	5	voplica.types.base.MPARating
best_for_category	6	voplica.types.ai.common.SubjectCategory		Typically CHARACTER_GENERIC / CHARACTER_FACE for a generated character dataset.

CancelDatasetPreparationRequest

Field	Number	Type	Label	Description
preparation_run_id	1	`string`	oneof `target`
root_workflow_execution_id	2	`string`	oneof `target`	Root workflow execution id returned by `PrepareDataset` (or the combined prepare-and-train parent) - cancels the in-flight run(s) it owns.

CancelDatasetPreparationResponse

Field	Number	Type	Label	Description
job_status	1	voplica.types.base.JobStatus		Will typically be CANCELLING immediately, transitioning to CANCELLED once the worker acknowledges the cancellation.

CreateDatasetPreparationRecipeRequest

Field	Number	Type	Label	Description
recipe	1	voplica.types.ai.dataset.DatasetPreparationRecipe		Recipe payload. `id`, `owner_id`, `created_at`, and `recipe_fingerprint` are ignored on the request side - the service assigns them.

CreateDatasetPreparationRecipeResponse

Field	Number	Type	Label	Description
recipe	1	voplica.types.ai.dataset.DatasetPreparationRecipe

DatasetPreparationByRecipe

Field	Number	Type	Label	Description
source_dataset_id	1	`string`
recipe_fingerprint	2	`string`

DatasetPreparationRecipeRef

Field	Number	Type	Label	Description
recipe_id	1	`string`	oneof `ref`	Use a persistent recipe by id.
recipe	2	voplica.types.ai.dataset.DatasetPreparationRecipe	oneof `ref`	Use an ad-hoc recipe payload. Not persisted as a row in `dataset_preparation_recipes`, but the full payload is stored on the `DatasetPreparation` so it remains retrievable later via `GetDatasetPreparation(include_full_recipe = true)`.

DeleteDatasetPreparationRecipeRequest

Field	Number	Type	Label	Description
recipe_id	1	`string`

DeleteDatasetPreparationRecipeResponse

No fields.

DeleteDatasetPreparationRequest

Field	Number	Type	Label	Description
preparation_id	1	`string`

DeleteDatasetPreparationResponse

No fields.

DeletePreparedItemCaptionRequest

Field	Number	Type	Label	Description
item	1	PreparedItemRef
format	2	voplica.types.ai.dataset.DatasetCaptionFormat

DeletePreparedItemCaptionResponse

Field	Number	Type	Label	Description
item	1	voplica.types.ai.dataset.PreparedItem

GenerateDatasetItemCaptionsRequest

Triggers ai.dataset_item_captioning_workflow over the preparation’s items (see voplica/types/ai/dataset/dataset_item_captioning_workflow.proto). Formats default to the preparation’s captioning.formats when empty.

Field	Number	Type	Label	Description
preparation_id	1	`string`
formats	2	voplica.types.ai.dataset.DatasetCaptionFormat	repeated
scope	3	voplica.types.ai.dataset.DatasetItemCaptioningScope
generation_config	4	voplica.types.ai.dataset.DatasetCaptionGenerationConfig	optional	Overrides the preparation’s `captioning.generation_config` for this trigger when set.
idempotency_key	5	`string`	optional	Optional idempotency key. Duplicate calls with the same key return the existing execution.
privacy_mode	6	voplica.types.workflow.WorkflowPrivacyMode

GenerateDatasetItemCaptionsResponse

Field	Number	Type	Label	Description
workflow_execution_id	1	`string`
estimated_items_count	2	`uint64`

GetDatasetPreparationRecipeRequest

Field	Number	Type	Label	Description
recipe_id	1	`string`	oneof `selector`
recipe_fingerprint	2	`string`	oneof `selector`

GetDatasetPreparationRecipeResponse

Field	Number	Type	Label	Description
recipe	1	voplica.types.ai.dataset.DatasetPreparationRecipe

GetDatasetPreparationRequest

Field	Number	Type	Label	Description
preparation_id	1	`string`	oneof `selector`
by_recipe	2	DatasetPreparationByRecipe	oneof `selector`
include_full_recipe	3	`bool`		When true and the preparation was created from an AD-HOC recipe, populates `DatasetPreparation.embedded_recipe` with the persisted recipe payload.

GetDatasetPreparationResponse

Field	Number	Type	Label	Description
preparation	1	voplica.types.ai.dataset.DatasetPreparation

GetDatasetPreparationRunRequest

Field	Number	Type	Label	Description
preparation_run_id	1	`string`	oneof `selector`
workflow_execution_id	2	`string`	oneof `selector`	Inner workflow execution id (the one stored in `DatasetPreparationRun.workflow_execution_id`). Returns nothing for runs whose inner execution has not been launched yet.
root_workflow_execution_id	3	`string`	oneof `selector`	Root workflow execution id (the one returned to the caller by `PrepareDataset` or by `StartModelTrainingJob` for combined runs). Use this when the caller only has the user-facing execution id.

GetDatasetPreparationRunResponse

Field	Number	Type	Label	Description
run	1	voplica.types.ai.dataset.DatasetPreparationRun

GetDatasetPreparationSnapshotRequest

Field	Number	Type	Label	Description
preparation_snapshot_id	1	`string`
fields_fetching_options	2	voplica.types.ai.dataset.DatasetPreparationSnapshotFetchingOptions

GetDatasetPreparationSnapshotResponse

Field	Number	Type	Label	Description
snapshot	1	voplica.types.ai.dataset.DatasetPreparationSnapshot

GetPreparedItemRequest

Field	Number	Type	Label	Description
item	1	PreparedItemRef
fields_fetching_options	2	voplica.types.ai.dataset.PreparedItemFetchingOptions	optional

GetPreparedItemResponse

Field	Number	Type	Label	Description
item	1	voplica.types.ai.dataset.PreparedItem

ListDatasetPreparationRecipesFilters

Field	Number	Type	Label	Description
owner_id	1	`string`	optional
asset_type	2	voplica.types.base.AssetType	optional
category	3	voplica.types.ai.common.SubjectCategory	optional	Matches recipes having AT LEAST ONE target with this category (category is a per-target property; a character recipe typically carries CHARACTER_GENERIC + CHARACTER_FACE).
visibility	4	voplica.types.base.VisibilityType	optional

ListDatasetPreparationRecipesRequest

Field	Number	Type	Label
filters	1	ListDatasetPreparationRecipesFilters	optional
limit	2	`int64`
cursor	10	`string`	optional
before	11	`bool`
check_more_available	12	`bool`

ListDatasetPreparationRecipesResponse

Field	Number	Type	Label
recipes	1	voplica.types.ai.dataset.DatasetPreparationRecipe	repeated
next_token	2	`string`	optional
prev_token	3	`string`	optional
more_available	4	`bool`	optional

ListDatasetPreparationRunItemsRequest

Field	Number	Type	Label
preparation_run_id	1	`string`
source_asset_ids	2	`string`	repeated
limit	3	`int64`
cursor	10	`string`	optional
before	11	`bool`
check_more_available	12	`bool`

ListDatasetPreparationRunItemsResponse

Field	Number	Type	Label
items	1	voplica.types.ai.dataset.DatasetPreparationRunItem	repeated
next_token	2	`string`	optional
prev_token	3	`string`	optional
more_available	4	`bool`	optional

ListDatasetPreparationRunsFilters

Field	Number	Type	Label
source_dataset_id	1	`string`	optional
preparation_id	2	`string`	optional
status	3	voplica.types.base.JobStatus	optional
owner_id	4	`string`	optional

ListDatasetPreparationRunsRequest

Field	Number	Type	Label
filters	1	ListDatasetPreparationRunsFilters	optional
limit	2	`int64`
cursor	10	`string`	optional
before	11	`bool`
check_more_available	12	`bool`

ListDatasetPreparationRunsResponse

Field	Number	Type	Label
runs	1	voplica.types.ai.dataset.DatasetPreparationRun	repeated
next_token	2	`string`	optional
prev_token	3	`string`	optional
more_available	4	`bool`	optional

ListDatasetPreparationSnapshotItemsFilters

Field	Number	Type	Label	Description
source_asset_ids	1	`string`	repeated
subject_ids	2	`string`	repeated
target_keys	3	`string`	repeated
categories	4	voplica.types.ai.common.SubjectCategory	repeated
caption_format_present	5	voplica.types.ai.dataset.DatasetCaptionFormat	optional	Only items carrying a caption of this format.

ListDatasetPreparationSnapshotItemsRequest

The trainer’s one-stream read: each row is a complete frozen training example (sample file [, mask/alpha], captions per format).

Field	Number	Type	Label
preparation_snapshot_id	1	`string`
filters	2	ListDatasetPreparationSnapshotItemsFilters	optional
fields_fetching_options	3	voplica.types.ai.dataset.PreparedItemFetchingOptions	optional
limit	4	`int64`
cursor	10	`string`	optional
before	11	`bool`
check_more_available	12	`bool`

ListDatasetPreparationSnapshotItemsResponse

Field	Number	Type	Label
items	1	voplica.types.ai.dataset.DatasetPreparationSnapshotItem	repeated
next_token	2	`string`	optional
prev_token	3	`string`	optional
more_available	4	`bool`	optional

ListDatasetPreparationSnapshotsFilters

Field	Number	Type	Label
source_dataset_id	1	`string`	optional
preparation_id	2	`string`	optional
recipe_fingerprint	3	`string`	optional
owner_id	4	`string`	optional

ListDatasetPreparationSnapshotsRequest

Field	Number	Type	Label
filters	1	ListDatasetPreparationSnapshotsFilters	optional
limit	2	`int64`
cursor	10	`string`	optional
before	11	`bool`
check_more_available	12	`bool`

ListDatasetPreparationSnapshotsResponse

Field	Number	Type	Label
snapshots	1	voplica.types.ai.dataset.DatasetPreparationSnapshot	repeated
next_token	2	`string`	optional
prev_token	3	`string`	optional
more_available	4	`bool`	optional

ListDatasetPreparationsRequest

Field	Number	Type	Label
source_dataset_id	1	`string`	optional
recipe_fingerprint	2	`string`	optional
owner_id	3	`string`	optional
limit	4	`int64`
cursor	10	`string`	optional
before	11	`bool`
check_more_available	12	`bool`

ListDatasetPreparationsResponse

Field	Number	Type	Label
preparations	1	voplica.types.ai.dataset.DatasetPreparation	repeated
next_token	2	`string`	optional
prev_token	3	`string`	optional
more_available	4	`bool`	optional

ListPreparedItemsFilters

Field	Number	Type	Label	Description
source_asset_ids	1	`string`	repeated
subject_ids	2	`string`	repeated
subject_instance_ids	3	`string`	repeated
target_keys	4	`string`	repeated
categories	5	voplica.types.ai.common.SubjectCategory	repeated
excluded	6	`bool`	optional	Unset = non-excluded only (the training-ready view). true = excluded only. false = all.
missing_caption_formats	7	voplica.types.ai.dataset.DatasetCaptionFormat	repeated	Items missing a caption in ANY of these formats (captioning work list). Multiple formats = one paging pass for a multi-format captioning job (OR semantics; the item’s missing formats are visible on the item itself).
missing_file_kind	8	voplica.types.ai.dataset.PreparedFileKind	optional	Items missing a file of this kind (visual work list).

ListPreparedItemsRequest

Field	Number	Type	Label
preparation_id	1	`string`
filters	2	ListPreparedItemsFilters	optional
fields_fetching_options	3	voplica.types.ai.dataset.PreparedItemFetchingOptions	optional
limit	4	`int64`
cursor	10	`string`	optional
before	11	`bool`
check_more_available	12	`bool`

ListPreparedItemsResponse

Field	Number	Type	Label
items	1	voplica.types.ai.dataset.PreparedItem	repeated
next_token	2	`string`	optional
prev_token	3	`string`	optional
more_available	4	`bool`	optional

OverridePreparedItemFileRequest

Replaces (or creates) ONE file of an item with a user-uploaded file (own crop, tweaked mask). FIND-OR-CREATE on the natural key: provenance USER_EDITED when replacing an auto-generated file, USER_AUTHORED when the slot (or the whole item) did not exist. Bumps the preparation version; snapshots that froze the previous file are unaffected. Auto-preparation and re-runs never overwrite USER_* files.

Field	Number	Type
item	1	PreparedItemRef
kind	2	voplica.types.ai.dataset.PreparedFileKind
file	3	voplica.types.file.FileInfo

OverridePreparedItemFileResponse

Field	Number	Type	Label	Description
item	1	voplica.types.ai.dataset.PreparedItem

PlanDatasetPreparationRequest

Field	Number	Type	Label	Description
source_dataset_id	1	`string`
recipe_ref	2	DatasetPreparationRecipeRef
reuse_policy	3	voplica.types.ai.dataset.DatasetPreparationReusePolicy
source_asset_ids	4	`string`	repeated	Optional subset. Empty means all source assets currently attached to the dataset.
subject_scope	5	voplica.types.ai.dataset.DatasetPreparationSubjectScope	optional	Optional config overrides to plan against (defaults: the existing preparation’s config, or empty scope/captioning when no preparation exists yet). Lets the caller preview the cost of a scope/captioning change before applying it.
captioning	6	voplica.types.ai.dataset.DatasetPreparationCaptioningConfig	optional

PlanDatasetPreparationResponse

Field	Number	Type	Label	Description
plan	1	voplica.types.ai.dataset.DatasetPreparationPlan
source_snapshot	2	voplica.types.ai.dataset.DatasetPreparationSourceSnapshot		The frozen source snapshot the plan was built from. Provided so callers can detect dataset drift between plan and subsequent `PrepareDataset` calls by comparing the `source_snapshot_fingerprint`.

PrepareDatasetRequest

Field	Number	Type	Label	Description
source_dataset_id	1	`string`
recipe_ref	2	DatasetPreparationRecipeRef
reuse_policy	3	voplica.types.ai.dataset.DatasetPreparationReusePolicy
source_asset_ids	4	`string`	repeated	Optional subset. Empty means all source assets currently attached to the dataset.
idempotency_key	5	`string`	optional	Optional client-provided idempotency key. If a run with the same idempotency key already exists for the dataset, the existing run is returned without starting a new workflow.
expected_source_snapshot_fingerprint	6	`string`	optional	Optional “if-match” guard. When set, `PrepareDataset` fails with FAILED_PRECONDITION if the current source dataset snapshot does not match this fingerprint. Use the `source_snapshot.source_snapshot_fingerprint` returned by `PlanDatasetPreparation` to guarantee the run uses the inputs the user reviewed.
subject_scope	7	voplica.types.ai.dataset.DatasetPreparationSubjectScope	optional
captioning	8	voplica.types.ai.dataset.DatasetPreparationCaptioningConfig	optional
preparation_name	9	`string`	optional	Optional display name applied to the found-or-created preparation when it does not have a name yet (an existing non-empty name is never overwritten by this request - rename explicitly via `UpdateDatasetPreparationConfig`). Display-only, never fingerprinted.
privacy_mode	10	voplica.types.workflow.WorkflowPrivacyMode

PrepareDatasetResponse

Field	Number	Type	Label	Description
preparation_run_id	1	`string`		Run row created for this preparation (always created; it records the frozen inputs and the plan even when no work was needed). When the plan required no work the run is returned already COMPLETED and no workflow is started.
preparation_id	2	`string`		The mutable preparation this run writes into (find-or-created for `(source_dataset_id, recipe_fingerprint)`).
workflow_execution_id	3	`string`	optional	Workflow execution started for the run. Empty when no work was needed.
plan	4	voplica.types.ai.dataset.DatasetPreparationPlan

PreparedItemNaturalKey

Natural key of an item slot. Used by curation RPCs with FIND-OR-CREATE semantics so a user can supply their OWN crop/mask/caption for a slot no worker ever filled.

Field	Number	Type	Label	Description
preparation_id	1	`string`
source_asset_id	2	`string`
subject_instance_id	3	`string`	optional	Empty for WHOLE_IMAGE targets.
target_key	4	`string`

PreparedItemRef

Selects an existing item by id, or an item SLOT by natural key (which may not exist yet - creating RPCs will find-or-create it; read RPCs return NOT_FOUND).

Field	Number	Type	Label	Description
prepared_item_id	1	`string`	oneof `ref`
natural_key	2	PreparedItemNaturalKey	oneof `ref`

ResetPreparedItemCaptionRequest

Reverts one caption to AUTO_GENERATED (regenerating via the captioning worker if needed).

Field	Number	Type	Label	Description
item	1	PreparedItemRef
format	2	voplica.types.ai.dataset.DatasetCaptionFormat

ResetPreparedItemCaptionResponse

Field	Number	Type	Label	Description
item	1	voplica.types.ai.dataset.PreparedItem

ResetPreparedItemFileRequest

Reverts one USER_* file back to the auto-generated one (re-preparing if needed) and bumps the preparation version.

Field	Number	Type	Label	Description
item	1	PreparedItemRef
kind	2	voplica.types.ai.dataset.PreparedFileKind

ResetPreparedItemFileResponse

Field	Number	Type	Label	Description
item	1	voplica.types.ai.dataset.PreparedItem

SetPreparedItemExcludedRequest

Curation: excluded items are omitted from future snapshots (and therefore from training). Bumps the preparation version.

Field	Number	Type	Label
item	1	PreparedItemRef
excluded	2	`bool`
reason	3	`string`	optional

SetPreparedItemExcludedResponse

Field	Number	Type	Label	Description
item	1	voplica.types.ai.dataset.PreparedItem

UpdateDatasetPreparationConfigRequest

Updates the preparation’s MUTABLE, dataset-local configuration. Does NOT bump version (config changes alter future planned work, not frozen content) and does not start any workflow - follow with PlanDatasetPreparation/PrepareDataset.

Field	Number	Type	Label	Description
preparation_id	1	`string`
subject_scope	2	voplica.types.ai.dataset.DatasetPreparationSubjectScope	optional
captioning	3	voplica.types.ai.dataset.DatasetPreparationCaptioningConfig	optional
name	4	`string`	optional	Display name (rename). Absent = unchanged; empty string = clear the name. Display-only: never contributes to fingerprints or preparation reuse.

UpdateDatasetPreparationConfigResponse

Field	Number	Type	Label	Description
preparation	1	voplica.types.ai.dataset.DatasetPreparation

UpdateDatasetPreparationRecipeRequest

Field	Number	Type	Label	Description
recipe_id	1	`string`
name	2	`string`	optional	Only display fields are mutable. Behavior-affecting fields require creating a new recipe (which yields a new `recipe_fingerprint`).
description	3	`string`	optional
visibility	4	voplica.types.base.VisibilityType	optional

UpdateDatasetPreparationRecipeResponse

Field	Number	Type	Label	Description
recipe	1	voplica.types.ai.dataset.DatasetPreparationRecipe

UpsertPreparedItemCaptionRequest

Creates or replaces ONE caption (per format) of an item. FIND-OR-CREATE on the natural key (an item slot with a user caption but no files yet is valid - files can arrive later). Provenance USER_AUTHORED for a new caption, USER_EDITED when overwriting an auto-generated one. Bumps the preparation version. The text must NOT contain trigger words.

Field	Number	Type
item	1	PreparedItemRef
format	2	voplica.types.ai.dataset.DatasetCaptionFormat
text	3	`string`

UpsertPreparedItemCaptionResponse

Field	Number	Type	Label	Description
item	1	voplica.types.ai.dataset.PreparedItem

AssignDatasetSubjectInstancesRequest

Batch curation. Side effect: every DatasetPreparation of the dataset holding items that reference a touched instance gets those items’ denormalized subject_id updated and its version bumped (subject labels are frozen snapshot content).

Field	Number	Type	Label	Description
source_dataset_id	1	`string`
assignments	2	DatasetSubjectInstanceAssignment	repeated

AssignDatasetSubjectInstancesResponse

Field	Number	Type	Label	Description
updated_count	1	`uint32`

CancelDatasetSubjectDiscoveryRequest

Field	Number	Type	Label	Description
discovery_run_id	1	`string`	oneof `target`
root_workflow_execution_id	2	`string`	oneof `target`

CancelDatasetSubjectDiscoveryResponse

Field	Number	Type	Label	Description
job_status	1	voplica.types.base.JobStatus

CreateDatasetSubjectRequest

Field	Number	Type	Label	Description
source_dataset_id	1	`string`
name	2	`string`
description	3	`string`	optional
category	4	voplica.types.ai.common.SubjectCategory		Localizable categories only (CHARACTER_GENERIC, OBJECT, VEHICLE, CLOTHES, …).
references	5	voplica.types.ai.dataset.DatasetSubjectReference	repeated	Optional exemplars: a face photo pins WHICH person is the character; an object image (+ description) is the defining input for reference-guided object detection.

CreateDatasetSubjectResponse

Field	Number	Type	Label	Description
subject	1	voplica.types.ai.dataset.DatasetSubject

DatasetSubjectFetchingOptions

Field	Number	Type	Label	Description
previews_retriever	1	voplica.types.file.FilesRetriever	optional	Hydrate `short_preview.default_preview` (the subject avatar Asset variants) for the curation UI. When unset the subject is returned without a hydrated preview (id/source only).

DatasetSubjectInstanceAssignment

Field	Number	Type	Label	Description
subject_instance_id	1	`string`
subject_id	2	`string`	oneof `to`	Assign to this identity (state becomes USER_ASSIGNED).
unassign	3	`bool`	oneof `to`	Detach from any identity (state becomes UNASSIGNED).
reject	4	`bool`	oneof `to`	Mark as not-a-subject (reflection, poster, false positive; state USER_REJECTED).

DatasetSubjectInstanceFetchingOptions

Field	Number	Type	Label	Description
preview_files_retriever	1	voplica.types.file.FilesRetriever	optional	Hydrate `preview_file` variants for the curation UI.

DeleteDatasetSubjectRequest

Deletes the identity. Member instances become UNASSIGNED. Frozen preparation snapshots keep their denormalized subject ids (lineage); live PER_SUBJECT filtering simply matches nothing.

Field	Number	Type	Label	Description
subject_id	1	`string`

DeleteDatasetSubjectResponse

No fields.

DiscoverDatasetSubjectsRequest

Field	Number	Type	Label	Description
source_dataset_id	1	`string`
policy	2	voplica.types.ai.dataset.DatasetSubjectDiscoveryPolicy
source_asset_ids	3	`string`	repeated	Optional subset. Empty means all source assets currently attached to the dataset.
idempotency_key	4	`string`	optional	Optional client-provided idempotency key: a replay with the same key returns the existing run without starting a new workflow.
expected_source_snapshot_fingerprint	5	`string`	optional	Optional “if-match” drift guard (FAILED_PRECONDITION on mismatch), same semantics as `PrepareDatasetRequest.expected_source_snapshot_fingerprint`.
privacy_mode	6	voplica.types.workflow.WorkflowPrivacyMode

DiscoverDatasetSubjectsResponse

Field	Number	Type	Label	Description
discovery_run_id	1	`string`		Always created (records the frozen inputs even when no work was needed - the run is then returned already COMPLETED and no workflow is started).
workflow_execution_id	2	`string`	optional

GetDatasetSubjectDiscoveryRunRequest

Field	Number	Type	Label
discovery_run_id	1	`string`	oneof `selector`
workflow_execution_id	2	`string`	oneof `selector`
root_workflow_execution_id	3	`string`	oneof `selector`

GetDatasetSubjectDiscoveryRunResponse

Field	Number	Type	Label	Description
run	1	voplica.types.ai.dataset.DatasetSubjectDiscoveryRun

GetDatasetSubjectRequest

Field	Number	Type	Label	Description
subject_id	1	`string`
fields_fetching_options	2	DatasetSubjectFetchingOptions	optional

GetDatasetSubjectResponse

Field	Number	Type	Label	Description
subject	1	voplica.types.ai.dataset.DatasetSubject

ListDatasetSubjectDiscoveryRunItemsRequest

Field	Number	Type	Label
discovery_run_id	1	`string`
source_asset_ids	2	`string`	repeated
limit	3	`int64`
cursor	10	`string`	optional
before	11	`bool`
check_more_available	12	`bool`

ListDatasetSubjectDiscoveryRunItemsResponse

Field	Number	Type	Label
items	1	voplica.types.ai.dataset.DatasetSubjectDiscoveryRunItem	repeated
next_token	2	`string`	optional
prev_token	3	`string`	optional
more_available	4	`bool`	optional

ListDatasetSubjectDiscoveryRunsRequest

Field	Number	Type	Label
source_dataset_id	1	`string`
status	2	voplica.types.base.JobStatus	optional
limit	3	`int64`
cursor	10	`string`	optional
before	11	`bool`
check_more_available	12	`bool`

ListDatasetSubjectDiscoveryRunsResponse

Field	Number	Type	Label
runs	1	voplica.types.ai.dataset.DatasetSubjectDiscoveryRun	repeated
next_token	2	`string`	optional
prev_token	3	`string`	optional
more_available	4	`bool`	optional

ListDatasetSubjectInstancesFilters

Field	Number	Type	Label
subject_id	1	`string`	optional
unassigned_only	2	`bool`	optional
include_rejected	3	`bool`	optional
source_asset_ids	4	`string`	repeated
min_quality	5	`float`	optional

ListDatasetSubjectInstancesRequest

Field	Number	Type	Label
source_dataset_id	1	`string`
filters	2	ListDatasetSubjectInstancesFilters	optional
fields_fetching_options	3	DatasetSubjectInstanceFetchingOptions	optional
limit	4	`int64`
cursor	10	`string`	optional
before	11	`bool`
check_more_available	12	`bool`

ListDatasetSubjectInstancesResponse

Field	Number	Type	Label
instances	1	voplica.types.ai.dataset.DatasetSubjectInstance	repeated
next_token	2	`string`	optional
prev_token	3	`string`	optional
more_available	4	`bool`	optional

ListDatasetSubjectsRequest

Field	Number	Type	Label
source_dataset_id	1	`string`
limit	2	`int64`
fields_fetching_options	3	DatasetSubjectFetchingOptions	optional
cursor	10	`string`	optional
before	11	`bool`
check_more_available	12	`bool`

ListDatasetSubjectsResponse

Field	Number	Type	Label
subjects	1	voplica.types.ai.dataset.DatasetSubject	repeated
next_token	2	`string`	optional
prev_token	3	`string`	optional
more_available	4	`bool`	optional

MergeDatasetSubjectsRequest

Merges duplicate identities (e.g. clustering split one person into two clusters). All instances of from_subject_ids become AUTO_ASSIGNED members of into_subject_id (USER_ASSIGNED stay USER_ASSIGNED); the source subjects are deleted; the target centroid is recomputed.

Field	Number	Type	Label
source_dataset_id	1	`string`
from_subject_ids	2	`string`	repeated
into_subject_id	3	`string`

MergeDatasetSubjectsResponse

Field	Number	Type	Label	Description
subject	1	voplica.types.ai.dataset.DatasetSubject

UpdateDatasetSubjectRequest

Field	Number	Type	Label	Description
subject_id	1	`string`
name	2	`string`	optional
description	3	`string`	optional
references	5	voplica.types.ai.dataset.DatasetSubjectReference	repeated	Full replacement when set (repeated fields cannot be optional; absent = keep).
replace_references	6	`bool`
preview_instance_id	7	`string`	oneof `preview_source`
preview_reference_id	8	`string`	oneof `preview_source`

UpdateDatasetSubjectResponse

Field	Number	Type	Label	Description
subject	1	voplica.types.ai.dataset.DatasetSubject

Enums

DatasetAssetSortKey

Name	Number	Description
DATASET_ASSET_SORT_KEY_UNSPECIFIED	0
ADDED_AT	1

DatasetSortKey

Name	Number	Description
DATASET_SORT_KEY_UNSPECIFIED	0
CREATED_AT	1
LAST_FILE_ADDED_AT	2
TRENDING_SCORE	3