Scope, architecture, and validation¶

Package structure¶

The distribution is pystac-ext-processing; its import path is pystac.extensions.processing. The wheel excludes the shared namespace initializer owned by PySTAC.

ProcessingExtension.ext() selects a wrapper for an Item, Asset, or ItemAssetDefinition. Wrappers modify the object's dictionaries directly. PySTAC's normal serialization includes those fields. The package stores metadata and does not execute processing expressions.

Asset wrappers capture an Item owner's property dictionary as a fallback when constructed. Provider wrappers are independent of extension membership. Collection summaries use a separate helper with list and schema-dictionary accessors.

Specification scope¶

The Processing specification places Collection processing metadata on producer or processor Providers, summaries, Assets, or item asset definitions, rather than directly on the Collection root. Item fields belong in properties and may also appear on Assets.

processing:version identifies the main processing chain; processing:software records individual tools. These differ from metadata versioning. Processing time also differs from an Item's metadata creation time. See the specification for the detailed conventions.

Validation boundaries¶

The implementation declares the v1.2.0 schema. That schema requires at least one processing field in Item properties; Asset-only fields do not satisfy its Item requirement. A Collection needs a processing field in an allowed location. Provider-based satisfaction requires a producer or processor role. Summary values are not validated by the extension schema.

The wrappers do not perform these schema checks. Setters store values without enforcing all annotated types or expression formats. Datetime accessors use PySTAC's conversion utilities; summary datetime accessors leave list values unchanged. ProcessingExpression checks required members when read, not when constructed.

apply() updates fields sequentially and clears omitted values. It is not transactional: if an assignment fails, earlier changes remain. to_dict() serializes without schema validation.

For full validation, install pystac[validation] and call item.validate() or collection.validate(). Remote schema access or an application-configured validator is needed. Validate summary contents separately where your application requires it.

Migration hooks¶

PROCESSING_EXTENSION_HOOKS exposes the schema URI, the legacy identifier processing, and the supported Item/Collection object types. This module does not automatically register its hook object with PySTAC. Importing the wrapper alone is not a migration workflow.