Schemas¶

Pydantic data models used across the AXIS backend.

Core Schemas¶

Classes¶

`DataFormat` ¶

Bases: StrEnum

Supported evaluation data formats.

`UploadResponse` ¶

Bases: BaseModel

Response after uploading evaluation data.

`DataFormatResponse` ¶

Bases: BaseModel

Response for data format detection.

`DataPreviewResponse` ¶

Bases: BaseModel

Preview of uploaded data with column info.

`MetricSummary` ¶

Bases: BaseModel

Summary statistics for a single metric.

`SummaryResponse` ¶

Bases: BaseModel

Aggregated summary of all metrics.

`DistributionStats` ¶

Bases: BaseModel

Descriptive statistics for a metric distribution.

`Histogram` ¶

Bases: BaseModel

Histogram bin counts and edges.

`DistributionResponse` ¶

Bases: BaseModel

Full distribution data for a single metric.

`EvaluationRecord` ¶

Bases: BaseModel

A single evaluation record with input/output fields.

Classes¶

`Config` ¶

Pydantic config allowing extra fields.

`TreeMetric` ¶

Bases: BaseModel

A metric in the tree-format evaluation structure.

`Annotation` ¶

Bases: BaseModel

A human annotation on an evaluation record.

`Persona` ¶

Bases: BaseModel

A simulation persona with traits.

`SimulationConfig` ¶

Bases: BaseModel

Configuration for running a simulation.

`SimulationResult` ¶

Bases: BaseModel

Result of a single simulation run.

`CalibrationResult` ¶

Bases: BaseModel

Inter-annotator agreement calibration results.

`ChatMessage` ¶

Bases: BaseModel

A single chat message.

`AnalysisInsight` ¶

Bases: BaseModel

An automated insight from data analysis.

Database Schemas¶

Classes¶

`SSLMode` ¶

Bases: StrEnum

Database SSL connection modes.

`DatabaseConnectionRequest` ¶

Bases: BaseModel

Request to connect to a database.

Functions¶

`validate_host(v)` `classmethod` ¶

Validate host is not empty.

Source code in backend/app/models/database_schemas.py

@field_validator("host")
@classmethod
def validate_host(cls, v: str) -> str:
    """Validate host is not empty."""
    if not v or not v.strip():
        raise ValueError("Host cannot be empty")
    return v.strip()

`validate_database(v)` `classmethod` ¶

Validate database name is not empty.

Source code in backend/app/models/database_schemas.py

@field_validator("database")
@classmethod
def validate_database(cls, v: str) -> str:
    """Validate database name is not empty."""
    if not v or not v.strip():
        raise ValueError("Database name cannot be empty")
    return v.strip()

`validate_username(v)` `classmethod` ¶

Validate username is not empty.

Source code in backend/app/models/database_schemas.py

@field_validator("username")
@classmethod
def validate_username(cls, v: str) -> str:
    """Validate username is not empty."""
    if not v or not v.strip():
        raise ValueError("Username cannot be empty")
    return v.strip()

`validate_port(v)` `classmethod` ¶

Validate port is in valid range.

Source code in backend/app/models/database_schemas.py

@field_validator("port")
@classmethod
def validate_port(cls, v: int) -> int:
    """Validate port is in valid range."""
    if v < 1 or v > 65535:
        raise ValueError("Port must be between 1 and 65535")
    return v

`ConnectResponse` ¶

Bases: BaseModel

Response from a successful database connection.

`TableIdentifier` ¶

Bases: BaseModel

Identifier for a database table.

Functions¶

`validate_name(v)` `classmethod` ¶

Validate table name is not empty.

Source code in backend/app/models/database_schemas.py

@field_validator("name")
@classmethod
def validate_name(cls, v: str) -> str:
    """Validate table name is not empty."""
    if not v or not v.strip():
        raise ValueError("Table name cannot be empty")
    return v.strip()

`TableInfo` ¶

Bases: BaseModel

Information about a database table.

`TablesListResponse` ¶

Bases: BaseModel

Response containing list of tables.

`ColumnInfo` ¶

Bases: BaseModel

Information about a table column.

`TableSchemaResponse` ¶

Bases: BaseModel

Response containing table schema information.

`ColumnMapping` ¶

Bases: BaseModel

Mapping from source DB column to target AXIS column.

Supported target columns: - dataset_id, query, actual_output, expected_output - retrieved_content, conversation, additional_input, document_text - actual_reference, expected_reference - tools_called, expected_tools, acceptance_criteria - latency, trace_id, observation_id

`FilterCondition` ¶

Bases: BaseModel

Simple equality filter condition.

`DatabaseImportRequest` ¶

Bases: BaseModel

Request to import data from database.

Functions¶

`validate_limit(v)` `classmethod` ¶

Validate limit is within bounds.

Source code in backend/app/models/database_schemas.py

@field_validator("limit")
@classmethod
def validate_limit(cls, v: int) -> int:
    """Validate limit is within bounds."""
    if v < 1:
        raise ValueError("Limit must be at least 1")
    if v > 10000:
        raise ValueError("Limit cannot exceed 10000 rows")
    return v

`PreviewRequest` ¶

Bases: BaseModel

Request to preview data from database.

`PreviewResponse` ¶

Bases: BaseModel

Response containing preview data.

`DistinctValuesRequest` ¶

Bases: BaseModel

Request to get distinct values for a column.

`DistinctValuesResponse` ¶

Bases: BaseModel

Response containing distinct values.

`DatabaseDefaults` ¶

Bases: BaseModel

Default database connection values from config.

`QueryPreviewRequest` ¶

Bases: BaseModel

Request to preview results of a SQL query.

Functions¶

`validate_query(v)` `classmethod` ¶

Validate query is safe.

Source code in backend/app/models/database_schemas.py

@field_validator("query")
@classmethod
def validate_query(cls, v: str) -> str:
    """Validate query is safe."""
    return _validate_sql_query(v)

`validate_limit(v)` `classmethod` ¶

Validate limit is within bounds.

Source code in backend/app/models/database_schemas.py

@field_validator("limit")
@classmethod
def validate_limit(cls, v: int) -> int:
    """Validate limit is within bounds."""
    if v < 1:
        raise ValueError("Limit must be at least 1")
    if v > 100:
        raise ValueError("Preview limit cannot exceed 100 rows")
    return v

`QueryImportRequest` ¶

Bases: BaseModel

Request to import data from a SQL query.

Functions¶

`validate_query(v)` `classmethod` ¶

Validate query is safe.

Source code in backend/app/models/database_schemas.py

@field_validator("query")
@classmethod
def validate_query(cls, v: str) -> str:
    """Validate query is safe."""
    return _validate_sql_query(v)

`validate_limit(v)` `classmethod` ¶

Validate limit is within bounds.

Source code in backend/app/models/database_schemas.py

@field_validator("limit")
@classmethod
def validate_limit(cls, v: int) -> int:
    """Validate limit is within bounds."""
    if v < 1:
        raise ValueError("Limit must be at least 1")
    if v > 50000:
        raise ValueError("Import limit cannot exceed 50000 rows")
    return v

Evaluation Runner Schemas¶

Classes¶

`AgentType` ¶

Bases: StrEnum

Type of agent connection for generating outputs.

`LLMProvider` ¶

Bases: StrEnum

Supported LLM providers for evaluation metrics.

`MetricInfo` ¶

Bases: BaseModel

Information about an available evaluation metric.

`MetricsResponse` ¶

Bases: BaseModel

Response with available metrics.

`ColumnMapping` ¶

Bases: BaseModel

Mapping of canonical field names to actual column names in the dataset.

`DatasetInfo` ¶

Bases: BaseModel

Information about an uploaded dataset.

`UploadResponse` ¶

Bases: BaseModel

Response from dataset upload.

`AgentAPIConfig` ¶

Bases: BaseModel

Configuration for Agent API connection.

`PromptTemplateConfig` ¶

Bases: BaseModel

Configuration for LLM prompt template.

`AgentConfig` ¶

Bases: BaseModel

Configuration for agent connection.

`TestConnectionRequest` ¶

Bases: BaseModel

Request to test agent connection.

`TestConnectionResponse` ¶

Bases: BaseModel

Response from testing agent connection.

`DatasetConfig` ¶

Bases: BaseModel

Dataset configuration including column mapping and data.

`EvaluationRunRequest` ¶

Bases: BaseModel

Request to run an evaluation.

`MetricResult` ¶

Bases: BaseModel

Result for a single metric across all items.

`ItemResult` ¶

Bases: BaseModel

Result for a single evaluation item.

`EvaluationSummary` ¶

Bases: BaseModel

Summary of evaluation results.

`EvaluationResultResponse` ¶

Bases: BaseModel

Full response from completed evaluation.

`ProgressEvent` ¶

Bases: BaseModel

Progress update during evaluation.

`LogEvent` ¶

Bases: BaseModel

Log message during evaluation.

`CompleteEvent` ¶

Bases: BaseModel

Completion event with results.

`ErrorEvent` ¶

Bases: BaseModel

Error event during evaluation.

Memory Schemas¶

Classes¶

`RuleRecord` ¶

Bases: BaseModel

A single extracted rule. Only id is typed; all other role-keyed fields are extras.

`ActionCount` ¶

Bases: BaseModel

Count of rules per action type.

`ProductCount` ¶

Bases: BaseModel

Count of rules per product type.

`SummaryResponse` ¶

Bases: BaseModel

Summary statistics for the rule memory.

`RulesResponse` ¶

Bases: BaseModel

Paginated list of rules with available filters.

`QualityResponse` ¶

Bases: BaseModel

Decision quality breakdown into aligned, divergent, and partial.

`SoftThresholdsResponse` ¶

Bases: BaseModel

Rules with soft thresholds.

`HardStopsResponse` ¶

Bases: BaseModel

Rules classified as hard stops.

`BatchInfo` ¶

Bases: BaseModel

Metadata for a single extraction batch.

`BatchesResponse` ¶

Bases: BaseModel

List of extraction batches.

`TraceResponse` ¶

Bases: BaseModel

Trace detail for a single rule — all fields are dynamic.

`ConflictInfo` ¶

Bases: BaseModel

Information about conflicting rules for a risk factor.

`ConflictsResponse` ¶

Bases: BaseModel

Response containing detected rule conflicts.

`StatusCountsResponse` ¶

Bases: BaseModel

Counts of rules grouped by ingestion status.

`RuleCreateRequest` ¶

Bases: BaseModel

Create a new rule — accepts any role-keyed fields.

`RuleUpdateRequest` ¶

Bases: BaseModel

Partial update for a rule — only provided fields are changed.

`RuleUpdateResponse` ¶

Bases: BaseModel

Response after creating or updating a rule.

`RuleDeleteResponse` ¶

Bases: BaseModel

Response after deleting a rule.

`MemoryUploadResponse` ¶

Bases: BaseModel

Response after uploading memory data.

Graph Schemas¶

Classes¶

`GraphNode` ¶

Bases: BaseModel

A node in the knowledge graph.

`GraphEdge` ¶

Bases: BaseModel

An edge in the knowledge graph.

`GraphData` ¶

Bases: BaseModel

Full graph payload with nodes and edges.

`GraphResponse` ¶

Bases: BaseModel

API response wrapping graph data.

`GraphSearchResult` ¶

Bases: BaseModel

A single search result from the graph.

`GraphSearchResponse` ¶

Bases: BaseModel

API response for graph search.

`GraphNeighborhoodResponse` ¶

Bases: BaseModel

API response for a node's neighborhood subgraph.

`GraphSummaryResponse` ¶

Bases: BaseModel

API response for graph summary statistics.

`GraphStatusResponse` ¶

Bases: BaseModel

API response for graph connection health check.

Alignment Schemas¶

Classes¶

`LLMProvider` ¶

Bases: StrEnum

Supported LLM providers.

`FewShotExample` ¶

Bases: BaseModel

A few-shot example for the judge prompt.

`JudgeConfig` ¶

Bases: BaseModel

Configuration for the LLM judge.

`AlignmentMetrics` ¶

Bases: BaseModel

Computed alignment metrics between human and LLM scores.

`AlignmentResult` ¶

Bases: BaseModel

Result for a single evaluated record.

`EvaluationRequest` ¶

Bases: BaseModel

Request to run LLM evaluation on a dataset.

`EvaluationResponse` ¶

Bases: BaseModel

Response from running LLM evaluation.

`MetricsRequest` ¶

Bases: BaseModel

Request to calculate alignment metrics.

`MetricsResponse` ¶

Bases: BaseModel

Response with computed alignment metrics.

`MisalignmentPattern` ¶

Bases: BaseModel

A pattern identified in misaligned cases.

`MisalignmentAnalysis` ¶

Bases: BaseModel

Analysis of misalignment patterns.

`MisalignmentAnalysisRequest` ¶

Bases: BaseModel

Request to analyze misalignment patterns.

`MisalignmentAnalysisResponse` ¶

Bases: BaseModel

Response with misalignment analysis.

`PromptSuggestion` ¶

Bases: BaseModel

A suggested improvement to the judge prompt.

`OptimizedPrompt` ¶

Bases: BaseModel

Optimized prompt with suggestions.

`OptimizePromptRequest` ¶

Bases: BaseModel

Request to optimize the judge prompt.

`OptimizePromptResponse` ¶

Bases: BaseModel

Response with optimized prompt.

`ExampleSelectionStrategy` ¶

Bases: StrEnum

Strategy for selecting few-shot examples.

`SuggestExamplesRequest` ¶

Bases: BaseModel

Request to suggest few-shot examples.

`SuggestExamplesResponse` ¶

Bases: BaseModel

Response with suggested few-shot examples.

`ModelInfo` ¶

Bases: BaseModel

Information about an available LLM model.

`ModelsResponse` ¶

Bases: BaseModel

Response with available models.

`SavedConfig` ¶

Bases: BaseModel

A saved judge configuration.

`SaveConfigRequest` ¶

Bases: BaseModel

Request to save a judge configuration.

`SaveConfigResponse` ¶

Bases: BaseModel

Response after saving configuration.

`ConfigsListResponse` ¶

Bases: BaseModel

Response with list of saved configurations.

`ClusteringMethod` ¶

Bases: StrEnum

Method for clustering annotation patterns.

`AnnotationWithNotes` ¶

Bases: BaseModel

Annotation with optional notes for Truesight pattern discovery.

`ErrorPattern` ¶

Bases: BaseModel

AI-clustered error pattern from annotation notes.

`LearningArtifactSchema` ¶

Bases: BaseModel

A distilled learning artifact from the EvidencePipeline.

`PipelineResultSchema` ¶

Bases: BaseModel

Metadata from the EvidencePipeline run.

`ClusterPatternsRequest` ¶

Bases: BaseModel

Request to cluster annotation notes into patterns.

`ClusterPatternsResponse` ¶

Bases: BaseModel

Response with clustered patterns.

Report Schemas¶

Classes¶

`InsightPatternSchema` ¶

Bases: BaseModel

A cross-metric pattern discovered from evaluation issues.

`InsightResultSchema` ¶

Bases: BaseModel

Structured insight result from InsightExtractor analysis.

Schemas¶

Core Schemas¶

Classes¶

DataFormat ¶

UploadResponse ¶

DataFormatResponse ¶

DataPreviewResponse ¶

MetricSummary ¶

SummaryResponse ¶

DistributionStats ¶

Histogram ¶

DistributionResponse ¶

EvaluationRecord ¶

Classes¶

Config ¶

TreeMetric ¶

Annotation ¶

Persona ¶

SimulationConfig ¶

SimulationResult ¶

CalibrationResult ¶

ChatMessage ¶

AnalysisInsight ¶

Database Schemas¶

Classes¶

SSLMode ¶

DatabaseConnectionRequest ¶

Functions¶

validate_host(v) classmethod ¶

validate_database(v) classmethod ¶

validate_username(v) classmethod ¶

validate_port(v) classmethod ¶

ConnectResponse ¶

TableIdentifier ¶

Functions¶

validate_name(v) classmethod ¶

TableInfo ¶

TablesListResponse ¶

ColumnInfo ¶

TableSchemaResponse ¶

ColumnMapping ¶

FilterCondition ¶

DatabaseImportRequest ¶

Functions¶

validate_limit(v) classmethod ¶

PreviewRequest ¶

PreviewResponse ¶

DistinctValuesRequest ¶

DistinctValuesResponse ¶

DatabaseDefaults ¶

QueryPreviewRequest ¶

Functions¶

validate_query(v) classmethod ¶

validate_limit(v) classmethod ¶

QueryImportRequest ¶

Functions¶

validate_query(v) classmethod ¶

validate_limit(v) classmethod ¶

Evaluation Runner Schemas¶

Classes¶

AgentType ¶

LLMProvider ¶

MetricInfo ¶

MetricsResponse ¶

ColumnMapping ¶

DatasetInfo ¶

UploadResponse ¶

AgentAPIConfig ¶

PromptTemplateConfig ¶

AgentConfig ¶

TestConnectionRequest ¶

TestConnectionResponse ¶

DatasetConfig ¶

EvaluationRunRequest ¶

MetricResult ¶

ItemResult ¶

EvaluationSummary ¶

EvaluationResultResponse ¶

ProgressEvent ¶

LogEvent ¶

`DataFormat` ¶

`UploadResponse` ¶

`DataFormatResponse` ¶

`DataPreviewResponse` ¶

`MetricSummary` ¶

`SummaryResponse` ¶

`DistributionStats` ¶

`Histogram` ¶

`DistributionResponse` ¶

`EvaluationRecord` ¶

`Config` ¶

`TreeMetric` ¶

`Annotation` ¶

`Persona` ¶

`SimulationConfig` ¶

`SimulationResult` ¶

`CalibrationResult` ¶

`ChatMessage` ¶

`AnalysisInsight` ¶

`SSLMode` ¶

`DatabaseConnectionRequest` ¶

`validate_host(v)` `classmethod` ¶

`validate_database(v)` `classmethod` ¶

`validate_username(v)` `classmethod` ¶

`validate_port(v)` `classmethod` ¶

`ConnectResponse` ¶

`TableIdentifier` ¶

`validate_name(v)` `classmethod` ¶

`TableInfo` ¶

`TablesListResponse` ¶

`ColumnInfo` ¶

`TableSchemaResponse` ¶

`ColumnMapping` ¶

`FilterCondition` ¶

`DatabaseImportRequest` ¶

`validate_limit(v)` `classmethod` ¶

`PreviewRequest` ¶

`PreviewResponse` ¶

`DistinctValuesRequest` ¶

`DistinctValuesResponse` ¶

`DatabaseDefaults` ¶

`QueryPreviewRequest` ¶

`validate_query(v)` `classmethod` ¶

`validate_limit(v)` `classmethod` ¶

`QueryImportRequest` ¶

`validate_query(v)` `classmethod` ¶

`validate_limit(v)` `classmethod` ¶

`AgentType` ¶

`LLMProvider` ¶

`MetricInfo` ¶

`MetricsResponse` ¶

`ColumnMapping` ¶

`DatasetInfo` ¶

`UploadResponse` ¶

`AgentAPIConfig` ¶

`PromptTemplateConfig` ¶

`AgentConfig` ¶

`TestConnectionRequest` ¶

`TestConnectionResponse` ¶

`DatasetConfig` ¶

`EvaluationRunRequest` ¶

`MetricResult` ¶

`ItemResult` ¶

`EvaluationSummary` ¶

`EvaluationResultResponse` ¶

`ProgressEvent` ¶

`LogEvent` ¶

`CompleteEvent` ¶

`ErrorEvent` ¶

`RuleRecord` ¶

`ActionCount` ¶

`ProductCount` ¶

`SummaryResponse` ¶

`RulesResponse` ¶

`QualityResponse` ¶

`SoftThresholdsResponse` ¶

`HardStopsResponse` ¶

`BatchInfo` ¶

`BatchesResponse` ¶

`TraceResponse` ¶