PatchAttrs
PatchAttrs stores metadata about a Patch. Some attributes describe identity or provenance, some summarize coordinates, and some describe the data array itself.
This note explains how DASCore interprets a few important attributes, and its internal policies around these attributes.
data_type
data_type is an optional label for the kind of data contained in a patch. It is useful for display defaults, plotting choices, grouping, and quick inspection, but it is not the canonical source of physical meaning, rather the data and coordinate units, as well as the patch history serve this purpose.
A stale or misleading data_type is much worse than an empty one.
| Situation | data_type behavior |
|---|---|
| Output is still the same measured quantity, just filtered/resampled/selected/reordered | Preserve existing data_type. |
| Output is a known derived product with a stable meaning | Set a specific snake_case data_type. |
| Output changes physical meaning but no stable label is appropriate | Clear data_type to "". |
DASCore-assigned data_type values should be snake_case and listed in VALID_DATA_TYPES in dascore.constants. Correctness-critical code should prefer units, coordinates, and explicit validation.
Patch functions
patch_function can manage output data_type for patch methods.
| Decorator value | Behavior |
|---|---|
data_type=None |
Preserve the returned patch’s data_type. This is the default for backward compatibility. |
data_type="" |
Clear the returned patch’s data_type. |
data_type="some_value" |
Set the returned patch’s data_type to that value. |
Functions may still require a specific input label with required_attrs, for example required_attrs={"data_type": "velocity"}. This should only be used when the function’s assumptions truly depend on that label and are documented.