Work with processing metadata¶

Run the tutorial first. The examples below continue with its imports, Item, and timestamps, in order.

Update or remove a field¶

processing.level = "L3"
processing.lineage = None
assert "processing:lineage" not in item.properties

ProcessingExtension.ext(item) requires extension membership. Pass add_if_missing=True to add it. Use individual setters to preserve other values; apply() clears omitted fields.

Attach Asset fields¶

asset = pystac.Asset(href="https://example.org/product.tif", media_type="image/tiff")
item.add_asset("data", asset)
asset_processing = ProcessingExtension.ext(asset)
assert asset_processing.level == "L3"
asset_processing.level = "L2"
assert asset.extra_fields["processing:level"] == "L2"
assert processing.level == "L3"
asset_processing.level = None
assert asset_processing.level == "L3"

Attach Assets to their owner before wrapping them. Missing Asset fields fall back to an Item owner's properties; writes affect only the Asset. Extension membership is checked or added on the owner.

Manage Collection summaries¶

collection = pystac.Collection(
    id="example-collection",
    description="Example processed products",
    extent=pystac.Extent(
        pystac.SpatialExtent([[-180.0, -90.0, 180.0, 90.0]]),
        pystac.TemporalExtent([[acquired_at, None]]),
    ),
    license="proprietary",
)
summaries = ProcessingExtension.summaries(collection, add_if_missing=True)
summaries.level = ["L2", "L3"]
summaries.processing_datetime = ["2024-01-02T12:00:00Z"]
summaries.software = {
    "type": "object",
    "properties": {"example-processor": {"type": "string"}},
}
assert summaries.level == ["L2", "L3"]
assert summaries.processing_datetime == ["2024-01-02T12:00:00Z"]
summaries.software = None
assert summaries.software is None

Summary accessors read and replace whole lists or schema dictionaries. They do not aggregate Items or offer apply(). Passing a Collection to ProcessingExtension.ext() raises pystac.ExtensionTypeError.

Use Collection item asset definitions¶

collection.item_assets = {
    "data": pystac.ItemAssetDefinition({"type": "image/tiff", "roles": ["data"]}),
}
definition = collection.item_assets["data"]
definition_processing = ProcessingExtension.ext(definition, add_if_missing=True)
definition_processing.level = "L2"
assert definition.properties["processing:level"] == "L2"

Retrieve definitions through collection.item_assets so PySTAC attaches their Collection owner.

Describe a processing Provider¶

provider = pystac.Provider(
    name="Example processing centre",
    roles=[pystac.ProviderRole.PROCESSOR],
)
provider_processing = ProcessingExtension.provider(provider)
provider_processing.apply(level="L2", software={"example-processor": "1.0.0"})
collection.providers = [provider]
ProcessingExtension.add_to(collection)
assert provider.extra_fields["processing:level"] == "L2"

The Provider wrapper writes extra_fields. It does not check roles or manage extension membership; declare the extension on the Collection separately.

Link provenance resources¶

Use ordinary PySTAC links for the specification's relation types: derived_from for inputs, processing-expression for a script, processing-execution for a run, processing-software for dependencies, and processing-validation for validation results.

item.add_link(pystac.Link(
    rel="derived_from",
    target="https://example.org/source-item.json",
    media_type="application/geo+json",
))
item.add_link(pystac.Link(
    rel="processing-expression",
    target="https://example.org/process.py",
    media_type="text/plain",
))