Matrix Views¶
A factory class responsible for creating virtual xarray Datasets from Consist's metadata.
This factory allows users to query the Consist metadata catalog for multi-dimensional
data (matrices) and lazily load the corresponding Zarr stores from disk. It consolidates
data across different runs into a single, unified xarray Dataset, indexed by run_id,
year, and iteration, facilitating comparative analysis and exploration.
Attributes:
| Name | Type | Description |
|---|---|---|
tracker |
Tracker
|
An instance of the Consist |
load_matrix_view(concept_key, variables=None, *, run_ids=None, parent_id=None, model=None, status=None)
¶
Returns a lazy xarray Dataset containing all runs that match the concept_key.
This method queries the Consist database for all matrix-type artifacts
associated with the given concept_key. It then lazily opens each corresponding
Zarr store using xarray and concatenates them into a single xarray.Dataset
along a new run_id dimension. This allows for convenient analysis of
multi-dimensional data across different experimental runs.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
concept_key
|
str
|
The semantic key (e.g., "model_output_grid", "simulation_results") identifying the collection of matrix artifacts to load. |
required |
variables
|
Optional[List[str]]
|
A list of variable names to load from each Zarr store. If |
None
|
run_ids
|
Optional[List[str]]
|
Optional list of run IDs to include in the view. |
None
|
parent_id
|
Optional[str]
|
Optional scenario/parent run ID to filter by. |
None
|
model
|
Optional[str]
|
Optional model name to filter by. |
None
|
status
|
Optional[str]
|
Optional run status to filter by (e.g., "completed"). |
None
|
Returns:
| Type | Description |
|---|---|
Dataset
|
A lazy-loaded |
Raises:
| Type | Description |
|---|---|
ImportError
|
If the |
RuntimeError
|
If the |