@Scout can I access asset metadata programmatically from an instance of dagster._core.definitions.asset_graph.InternalAssetGraph

The documentation does not provide a direct example or method for accessing asset metadata programmatically from an instance of

dagster._core.definitions.asset_graph.InternalAssetGraph

. However, it does mention that the context object passed to functions decorated with

@asset

contains an instance attribute that can query the Dagster event log for metadata related to assets. For example, if you want to access metadata from a historical materialization of one asset from inside another asset, you can use the

context.instance

to get the latest materialization event and then access the metadata from that event. Here's a snippet from the documentation illustrating this approach:

python @asset def asset1():     return Output(value=..., metadata={"num_rows": 5})  @asset  # or op def asset_1_num_rows(context):     instance = context.instance     materialization = instance.get_latest_materialization_event(AssetKey(["asset1"])).asset_materialization     return materialization.metadata["num_rows"]

This example shows how to retrieve the metadata for the number of rows from the latest materialization of

asset1

within another asset. If you are looking to interact with the

InternalAssetGraph

directly, you would typically be working with the internal APIs of Dagster, which are not covered in the user-facing documentation. The

InternalAssetGraph

is part of Dagster's core implementation and is not intended for direct manipulation by end-users. Instead, the recommended approach is to use the public APIs and the context objects provided by Dagster's decorators like

@asset

. If you need to perform operations that are not covered by the public APIs, you might need to explore the source code or reach out to the Dagster community for more specific guidance.