Create workflow and experiment records¶

Use OGCRecord for workflow and experiment discovery metadata, and PySTAC links to connect related resources. The relationship names in this example are application choices.

import pystac
from pystac.extensions.ogc_record import OGCRecord

workflow = OGCRecord(
    id="ocean-workflow",
    properties={
        "type": "workflow",
        "title": "Ocean temperature workflow",
        "description": "Processing steps for sea surface temperature",
    },
)
workflow.add_link(pystac.Link(
    rel="related", target="https://example.org/projects/ocean-science.json",
))
workflow.add_link(pystac.Link(
    rel="child", target="https://example.org/experiments/experiment-42.json",
))

experiment = OGCRecord(
    id="experiment-42",
    properties={
        "type": "experiment",
        "title": "Ocean workflow execution 42",
        "description": "An execution producing the ocean-temperature product",
    },
)
experiment.add_link(pystac.Link(
    rel="related", target="https://example.org/workflows/ocean-workflow.json",
))
experiment.add_link(pystac.Link(
    rel="child", target="https://example.org/products/ocean-temperature.json",
))
experiment.add_link(pystac.Link(
    rel="environment", target="https://example.org/experiments/environment.yml",
))
experiment.add_link(pystac.Link(
    rel="input", target="https://example.org/experiments/parameters.json",
))

experiment.keywords = ["ocean", "temperature"]
experiment.language = {"code": "en"}
document = experiment.to_record_dict()
assert document["type"] == "Feature"
assert document["properties"]["type"] == "experiment"
assert "stac_version" not in document
restored = OGCRecord.from_dict(document)
assert restored.title == experiment.title
assert restored.get_single_link("related").target == "https://example.org/workflows/ocean-workflow.json"

The workflow and experiment resource type strings describe the application. Constructing these records does not establish full OGC conformance.

Read, write, and copy¶

Use OGCRecord.from_dict(document) on decoded JSON. Use json.dumps(record.to_dict()) to serialize it. Pass href= to from_dict() to set the self link, replacing any incoming self link. Metadata is copied even when preserve_dict=False; migrate is accepted for PySTAC compatibility but has no effect.

record.clone() copies metadata and clones links and assets. record.record_metadata is a live view of record.properties; its to_dict() returns a deep copy. Assigning None to a common metadata accessor removes that property.

Export a STAC Item¶

from datetime import datetime, timezone

record = OGCRecord(
    id="observation",
    datetime=datetime(2026, 1, 1, tzinfo=timezone.utc),
    properties={"type": "dataset", "title": "Observation"},
)
item = record.to_stac_item()
assert "stac_version" in item.to_dict()

STAC export requires datetime or both start_datetime and end_datetime. An OGC time member alone does not satisfy this requirement. Export does not validate the resulting Item. See validation boundaries.