Exceptions¶

Exceptions raised by the rivers runtime. All are re-exported under rivers.exceptions — catch them from there rather than reaching into _core.

import rivers as rs
from rivers.exceptions import GraphValidationError, ConfigurationError

try:
    repo = rs.CodeRepository(assets=[bad_asset])
except GraphValidationError as exc:
    print("Graph could not be validated:", exc)

Every class extends Python's built-in Exception.

Class	Raised when
`AssetDefinitionError`	An `@Asset` definition is invalid (bad signature, conflicting flags, …).
`AssetNotFoundError`	A referenced asset key does not exist in the repository.
`AssetOutputValidationError`	An asset output fails validation (type, schema, metadata).
`ConfigurationError`	User configuration (resources, run config) cannot be resolved.
`ExecutionError`	An asset/task body raised during execution.
`GraphValidationError`	The asset graph fails validation (cycles, dangling deps).
`InvalidMetadataError`	A metadata value cannot be encoded as a `MetadataValue`.
`NodeNotFoundError`	A referenced node (asset/task) is missing from the graph.
`PartitionDefinitionError`	A `PartitionsDefinition` is constructed with invalid arguments.
`PartitionValidationError`	A `PartitionKey` is not valid for the asset's partitions def.
`ResultDefinitionError`	An asset's `Output` / `Observation` is malformed.
`ScheduleDefinitionError`	A `Schedule` is constructed with invalid arguments.
`SensorDefinitionError`	A `Sensor` is constructed with invalid arguments.
`StorageError`	A storage backend failure (connection, transaction, serialization).
`TaskDefinitionError`	A `Task` / `BashTask` definition is invalid.

When materialize(raise_on_error=True) (the default) is used, an ExecutionError is raised on the first failed step. With raise_on_error=False, failures are collected into RunResult.failed_assets instead.