Implementation:Evidentlyai Evidently SDK Models

Knowledge Sources	Evidentlyai_Evidently
Domains	SDK, Data Models
Last Updated	2026-02-14 12:00 GMT

Overview

Defines core Pydantic data models for the Evidently SDK including project, dashboard, panel, snapshot, and metric configuration models.

Description

The SDK Models module provides the foundational data models used across the Evidently SDK for representing projects, dashboards, panels, metrics, and snapshots. These models are used for API communication with the Evidently workspace and for configuring dashboard visualizations.

Key models:

• DashboardTabModel -- Represents a tab in a dashboard. Contains an auto-generated id (UUID7), an optional title, and a list of panels (panel IDs) that belong to the tab.

• PanelMetric -- Configures a metric to display in a dashboard panel. Includes legend text, tags, metadata, a metric identifier string (automatically prefixed with "evidently:metric_v2:" if not already), metric_labels for display customization, and view_params for rendering configuration. The metric_is_alias validator ensures metric identifiers are in the canonical format.

• DashboardPanelPlot -- Represents a single visualization panel in a dashboard. Contains an auto-generated id, title, optional subtitle and size, a list of PanelMetric values, and plot_params for configuring the visualization type (text, counter, line, bar, pie).

• DashboardModel -- The complete dashboard configuration containing lists of tabs and panels.

• ProjectModel -- Represents an Evidently project with id, name, optional description, optional org_id (for cloud workspaces), and a version string (default: "2").

• SnapshotLink -- Associates a snapshot with a dataset, specifying the snapshot_id, dataset_type (e.g., "reference", "current"), and dataset_subtype (e.g., "production", "training").

• SnapshotMetadataModel -- Metadata for a snapshot (evaluation run) including id, name, metadata, tags, timestamp, and links to associated datasets.

Usage

Use these models when programmatically creating or reading project configurations, dashboard layouts, and snapshot metadata through the Evidently SDK. The panel and dashboard models are used with the panel factory functions in evidently.sdk.panels to build dashboard configurations.

Code Reference

Source Location

• Repository: Evidentlyai_Evidently
• File: src/evidently/sdk/models.py

Signature

class DashboardTabModel(BaseModel):
    id: TabID = Field(default_factory=uuid6.uuid7)
    title: Optional[str]
    panels: List[PanelID]

class PanelMetric(BaseModel):
    legend: Optional[str] = None
    tags: List[str] = Field(default_factory=list)
    metadata: Dict[str, str] = Field(default_factory=dict)
    metric: str  # auto-prefixed with "evidently:metric_v2:"
    metric_labels: Dict[str, str] = Field(default_factory=dict)
    view_params: Dict[str, Any] = Field(default_factory=dict)

class DashboardPanelPlot(BaseModel):
    id: PanelID = Field(default_factory=uuid6.uuid7)
    title: str
    subtitle: Optional[str]
    size: Optional[str]
    values: List[PanelMetric]
    plot_params: Dict[str, Any] = Field(default_factory=dict)

class DashboardModel(BaseModel):
    tabs: List[DashboardTabModel]
    panels: List[DashboardPanelPlot]

class ProjectModel(BaseModel):
    id: ProjectID = Field(default_factory=new_id)
    name: str
    description: Optional[str] = None
    org_id: Optional[OrgID] = None
    version: str = "2"

class SnapshotLink(BaseModel):
    snapshot_id: SnapshotID
    dataset_type: str
    dataset_subtype: str

class SnapshotMetadataModel(BaseModel):
    id: SnapshotID
    name: Optional[str]
    metadata: Dict[str, str]
    tags: List[str]
    timestamp: datetime
    links: SnapshotLinks

Import

from evidently.sdk.models import (
    DashboardTabModel,
    PanelMetric,
    DashboardPanelPlot,
    DashboardModel,
    ProjectModel,
    SnapshotLink,
    SnapshotMetadataModel,
)

I/O Contract

Inputs

Name	Type	Required	Description
title	str	Yes (for DashboardPanelPlot, DashboardTabModel)	Title of the panel or tab
name	str	Yes (for ProjectModel)	Name of the project
metric	str	Yes (for PanelMetric)	Metric identifier; auto-prefixed with "evidently:metric_v2:" if not already present
values	List[PanelMetric]	Yes (for DashboardPanelPlot)	List of metrics to display in the panel
snapshot_id	SnapshotID	Yes (for SnapshotLink)	ID of the snapshot to link
dataset_type	str	Yes (for SnapshotLink)	Type of dataset (e.g., "reference", "current")
dataset_subtype	str	Yes (for SnapshotLink)	Subtype of dataset (e.g., "production", "training")

Outputs

Name	Type	Description
DashboardTabModel	BaseModel	A tab configuration with auto-generated ID
PanelMetric	BaseModel	A validated metric configuration with canonical metric identifier
DashboardPanelPlot	BaseModel	A panel configuration with auto-generated ID
ProjectModel	BaseModel	A project model with auto-generated ID
SnapshotMetadataModel	BaseModel	Snapshot metadata with timestamp and dataset links

Usage Examples

from evidently.sdk.models import (
    ProjectModel,
    DashboardModel,
    DashboardTabModel,
    DashboardPanelPlot,
    PanelMetric,
    SnapshotLink,
)

# Create a project model
project = ProjectModel(name="My ML Project", description="Production monitoring")

# Create a panel metric (auto-prefixed)
metric = PanelMetric(
    metric="ColumnSummary",  # becomes "evidently:metric_v2:ColumnSummary"
    legend="Mean Value",
    metric_labels={"column": "prediction"},
)

# Create a panel
panel = DashboardPanelPlot(
    title="Prediction Distribution",
    subtitle="Over time",
    size="full",
    values=[metric],
    plot_params={"plot_type": "line"},
)

# Create a dashboard
dashboard = DashboardModel(
    tabs=[DashboardTabModel(title="Overview", panels=[panel.id])],
    panels=[panel],
)

# Create a snapshot link
link = SnapshotLink(
    snapshot_id="snapshot-uuid",
    dataset_type="current",
    dataset_subtype="production",
)

Related Pages

• Environment:Evidentlyai_Evidently_Python_Core_Environment