https://github.com/azavea/augdiff-pipeline/pull/17 contains a Node app that converts Overpass augmented diffs into GeoJSON (with before/after geometries). Sequences are written to S3. Empty sequences are omitted, so the current way to retrieve the "latest" available sequence is to list the bucket/prefix (which gets slower as more data is added).
state.yaml should be created/updated after each sequence is written (or omitting) so that consumers can poll a single object to determine the current, definitive state of the stream.
https://github.com/azavea/augdiff-pipeline/pull/17 contains a Node app that converts Overpass augmented diffs into GeoJSON (with before/after geometries). Sequences are written to S3. Empty sequences are omitted, so the current way to retrieve the "latest" available sequence is to list the bucket/prefix (which gets slower as more data is added).
state.yaml
should be created/updated after each sequence is written (or omitting) so that consumers can poll a single object to determine the current, definitive state of the stream.