Breakthrough Insights
The kind of Breakthrough output an insight represents.
Attributes:
| Name | Type | Description |
|---|---|---|
OPTIMIZER |
str
|
A standard optimizer result. |
CUSTOM_OPTIMIZER |
str
|
A custom-configured optimizer result. |
IMPACT_CHART |
str
|
An impact chart showing how inputs affect an outcome. |
MOLECULE |
str
|
A molecule-related insight. |
SMART_DOE |
str
|
A smart design-of-experiments result. |
GENERATE |
str
|
A generative result (e.g. generated candidates). |
Progress state of a Breakthrough insight computation.
Attributes:
| Name | Type | Description |
|---|---|---|
QUEUED |
str
|
The computation has been queued but has not started. |
BUILDING_MODELS |
str
|
Underlying models are being trained. |
GENERATING_CANDIDATES |
str
|
Candidate solutions are being generated. |
COMPLETE |
str
|
The computation has finished successfully. |
ERROR |
str
|
The computation failed. |
Which system produced an insight's payload.
Attributes:
| Name | Type | Description |
|---|---|---|
BREAKTHROUGH |
str
|
Payload produced by Breakthrough. |
ALBERTO |
str
|
Legacy payload configuration retained for backwards compatibility with insights created in 2024 that predate the current session schema; slated for deprecation. |
Bases: BaseAlbertModel
Result metadata recorded for a Breakthrough insight.
Show JSON schema:
{
"description": "Result metadata recorded for a Breakthrough insight.",
"properties": {
"BuildLogs": {
"anyOf": [
{
"additionalProperties": true,
"type": "object"
},
{
"type": "null"
}
],
"default": null,
"description": "Free-form logs captured while the insight was computed.",
"title": "Buildlogs"
},
"Metrics": {
"anyOf": [
{
"additionalProperties": true,
"type": "object"
},
{
"type": "null"
}
],
"default": null,
"description": "Free-form performance metrics recorded for the insight.",
"title": "Metrics"
},
"Settings": {
"anyOf": [
{
"additionalProperties": true,
"type": "object"
},
{
"type": "null"
}
],
"default": null,
"description": "Free-form settings used to produce the insight.",
"title": "Settings"
}
},
"title": "BTInsightRegistry",
"type": "object"
}
Fields:
Free-form logs captured while the insight was computed.
Free-form performance metrics recorded for the insight.
Bases: BaseResource
An output or insight produced by Breakthrough.
Albert Breakthrough is Albert's inverse-design / ML optimization capability. An insight captures a
Breakthrough result, such as an optimizer run, impact chart, or set of
generated candidates. It can trace back to the dataset, model session, and
model it came from via dataset_id, model_session_id, and model_id
(see BTDataset and
BTModel). Insights are managed through
BTInsightCollection.
Example
Show JSON schema:
{
"$defs": {
"AuditFields": {
"description": "The audit fields for a resource",
"properties": {
"by": {
"default": null,
"title": "By",
"type": "string"
},
"byName": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"title": "Byname"
},
"at": {
"anyOf": [
{
"format": "date-time",
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"title": "At"
}
},
"title": "AuditFields",
"type": "object"
},
"BTInsightCategory": {
"description": "The kind of Breakthrough output an insight represents.\n\nAttributes\n----------\nOPTIMIZER : str\n A standard optimizer result.\nCUSTOM_OPTIMIZER : str\n A custom-configured optimizer result.\nIMPACT_CHART : str\n An impact chart showing how inputs affect an outcome.\nMOLECULE : str\n A molecule-related insight.\nSMART_DOE : str\n A smart design-of-experiments result.\nGENERATE : str\n A generative result (e.g. generated candidates).",
"enum": [
"Optimizer",
"Custom Optimizer",
"Impact Chart",
"Molecule",
"Smart DOE",
"Generate"
],
"title": "BTInsightCategory",
"type": "string"
},
"BTInsightPayloadType": {
"description": "Which system produced an insight's payload.\n\nAttributes\n----------\nBREAKTHROUGH : str\n Payload produced by Breakthrough.\nALBERTO : str\n Legacy payload configuration retained for backwards compatibility with\n insights created in 2024 that predate the current session schema; slated\n for deprecation.",
"enum": [
"Breakthrough",
"Alberto"
],
"title": "BTInsightPayloadType",
"type": "string"
},
"BTInsightRegistry": {
"description": "Result metadata recorded for a Breakthrough insight.",
"properties": {
"BuildLogs": {
"anyOf": [
{
"additionalProperties": true,
"type": "object"
},
{
"type": "null"
}
],
"default": null,
"description": "Free-form logs captured while the insight was computed.",
"title": "Buildlogs"
},
"Metrics": {
"anyOf": [
{
"additionalProperties": true,
"type": "object"
},
{
"type": "null"
}
],
"default": null,
"description": "Free-form performance metrics recorded for the insight.",
"title": "Metrics"
},
"Settings": {
"anyOf": [
{
"additionalProperties": true,
"type": "object"
},
{
"type": "null"
}
],
"default": null,
"description": "Free-form settings used to produce the insight.",
"title": "Settings"
}
},
"title": "BTInsightRegistry",
"type": "object"
},
"BTInsightState": {
"description": "Progress state of a Breakthrough insight computation.\n\nAttributes\n----------\nQUEUED : str\n The computation has been queued but has not started.\nBUILDING_MODELS : str\n Underlying models are being trained.\nGENERATING_CANDIDATES : str\n Candidate solutions are being generated.\nCOMPLETE : str\n The computation has finished successfully.\nERROR : str\n The computation failed.",
"enum": [
"Queued",
"Building Models",
"Generating Candidates",
"Complete",
"Error"
],
"title": "BTInsightState",
"type": "string"
},
"Status": {
"description": "The status of a resource.\n\nAttributes\n----------\nACTIVE : str\n The resource is fully operational and visible in normal operations.\nINACTIVE : str\n The resource is hidden from normal operations and disabled from use.",
"enum": [
"active",
"inactive"
],
"title": "Status",
"type": "string"
}
},
"description": "An output or insight produced by Breakthrough.\n\nAlbert Breakthrough is Albert's inverse-design / ML optimization capability. An insight captures a\nBreakthrough result, such as an optimizer run, impact chart, or set of\ngenerated candidates. It can trace back to the dataset, model session, and\nmodel it came from via ``dataset_id``, ``model_session_id``, and ``model_id``\n(see [`BTDataset`][albert.resources.btdataset.BTDataset] and\n[`BTModel`][albert.resources.btmodel.BTModel]). Insights are managed through\n[`BTInsightCollection`][albert.collections.btinsight.BTInsightCollection].\n\n!!! example\n ```python\n from albert.resources.btinsight import BTInsight, BTInsightCategory\n\n insight = BTInsight(\n name=\"Cost optimizer run\",\n category=BTInsightCategory.OPTIMIZER,\n )\n ```",
"properties": {
"status": {
"anyOf": [
{
"$ref": "#/$defs/Status"
},
{
"type": "null"
}
],
"default": null,
"description": "The status of the resource, optional."
},
"Created": {
"anyOf": [
{
"$ref": "#/$defs/AuditFields"
},
{
"type": "null"
}
],
"default": null,
"description": "Audit fields for the creation of the resource, optional."
},
"Updated": {
"anyOf": [
{
"$ref": "#/$defs/AuditFields"
},
{
"type": "null"
}
],
"default": null,
"description": "Audit fields for the update of the resource, optional."
},
"name": {
"description": "Human-readable name of the insight.",
"title": "Name",
"type": "string"
},
"category": {
"$ref": "#/$defs/BTInsightCategory",
"description": "The kind of Breakthrough output this insight represents."
},
"Metadata": {
"anyOf": [
{
"additionalProperties": true,
"type": "object"
},
{
"type": "null"
}
],
"default": null,
"description": "Free-form metadata associated with the insight.",
"title": "Metadata"
},
"state": {
"anyOf": [
{
"$ref": "#/$defs/BTInsightState"
},
{
"type": "null"
}
],
"default": null,
"description": "Current progress state of the insight computation."
},
"albertId": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Unique identifier of the insight (format ``INS...``). Assigned by Albert on creation.",
"title": "Albertid"
},
"parentId": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Identifier of the project the insight belongs to (format ``PRO...``).",
"title": "Parentid"
},
"datasetId": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Identifier of the dataset the insight is based on (format ``DST...``).",
"title": "Datasetid"
},
"modelSessionId": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Identifier of the model session the insight came from (format ``MDS...``).",
"title": "Modelsessionid"
},
"modelId": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Identifier of the model the insight came from (format ``MDL...``).",
"title": "Modelid"
},
"outputKey": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Storage key for the insight's output artifact, if applicable.",
"title": "Outputkey"
},
"startTime": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "When the insight computation started, if recorded.",
"title": "Starttime"
},
"endTime": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "When the insight computation finished, if recorded.",
"title": "Endtime"
},
"totalTime": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Total time taken to compute the insight, if recorded.",
"title": "Totaltime"
},
"RawPayload": {
"anyOf": [
{
"additionalProperties": true,
"type": "object"
},
{
"type": "null"
}
],
"default": null,
"description": "The raw result payload produced for the insight.",
"title": "Rawpayload"
},
"payloadType": {
"anyOf": [
{
"$ref": "#/$defs/BTInsightPayloadType"
},
{
"type": "null"
}
],
"default": null,
"description": "Which system produced the payload."
},
"Registry": {
"anyOf": [
{
"$ref": "#/$defs/BTInsightRegistry"
},
{
"type": "null"
}
],
"default": null,
"description": "Result metadata (logs, metrics, settings) recorded for the insight."
},
"contentEdited": {
"anyOf": [
{
"type": "boolean"
},
{
"type": "null"
}
],
"default": null,
"description": "Whether the insight's content has been manually edited.",
"title": "Contentedited"
}
},
"required": [
"name",
"category"
],
"title": "BTInsight",
"type": "object"
}
Fields:
-
status(Status | None) -
created(AuditFields | None) -
updated(AuditFields | None) -
name(str) -
category(BTInsightCategory) -
metadata(dict[str, Any] | None) -
state(BTInsightState | None) -
id(BTInsightId | None) -
parent_id(ProjectId | None) -
dataset_id(BTDatasetId | None) -
model_session_id(BTModelSessionId | None) -
model_id(BTModelId | None) -
output_key(str | None) -
start_time(str | None) -
end_time(str | None) -
total_time(str | None) -
raw_payload(dict[str, Any] | None) -
payload_type(BTInsightPayloadType | None) -
registry(BTInsightRegistry | None) -
content_edited(bool | None)
id: BTInsightId | None = None
Unique identifier of the insight (format INS...). Assigned by Albert on creation.
parent_id: ProjectId | None = None
Identifier of the project the insight belongs to (format PRO...).
dataset_id: BTDatasetId | None = None
Identifier of the dataset the insight is based on (format DST...).
model_session_id: BTModelSessionId | None = None
Identifier of the model session the insight came from (format MDS...).
model_id: BTModelId | None = None
Identifier of the model the insight came from (format MDL...).
output_key: str | None = None
Storage key for the insight's output artifact, if applicable.
The raw result payload produced for the insight.
registry: BTInsightRegistry | None = None
Result metadata (logs, metrics, settings) recorded for the insight.
content_edited: bool | None = None
Whether the insight's content has been manually edited.