v0.4.5

[v0.4.5] - 2021-07-12

Fixed

• Don't log S3 credentials

v0.4.4

Added

• Ability to query catalog for item(s) based on property values, use either get_items_by_properties or get_item_by_properties methods

v0.4.3

[v0.4.3] - 2021-03-30

Fixed

• Race condition when setting processing of new catalogs
• Pagination of items in Cirrus API

v0.4.2

[v0.4.2] - 2021-01-12

Added

• Add support for sorting of queries based on updated column

v0.4.1

[v0.4.1] - 2020-11-16

Fixed

• Bug preventing rerun of inputs

v0.4.0

[v0.4.0] - 2020-11-13

Added

• Expanded unit tests
• status attribute to published SNS attributes, either created or updated

Changed

• DynamoDB state database schema changed:
  • input_collections -> collections_workflow, combines collections string and workflow name
  • id -> itemids, IDs of all input STAC Items, no longer prefaced with workflow (moved to collections_workflow)
  • output_collections field removed
  • current_state -> state_updated, same contents
  • updated field added containing just the updated datetime
  • created_at -> created
  • output_urls -> outputs, still a List of canonical STAC Item URLs
  • error_message -> last_error, contains the last execution error if input has ever failed
  • execution -> executions, now a list of all executions for this input catalog.
• Cirrus State Item changed:
  • input_collections -> collections
  • created_at -> created
  • input_catalog -> catalog
  • output_urls -> outputs
  • error_message -> last_error, now stores last execution error
  • execution -> executions, now a list of all executions for this input catalog
  • output_collections removed
  • updated added

v0.3.3

[v0.3.3] - 2020-10-27

Fixed

• Catalog logging when updating

v0.3.2

[v0.3.2] - 2020-10-25

Added

• cirruslib.logging module configuring structured (JSON) logging and get_task_logger for logging from tasks
• cirruslib.stac module for working with the Cirrus static STAC catalog on s3, uses PySTAC
• utils.dict_merged function for doing recursive merges

Changed

• Parsing payload for a task should now use Catalog.from_payload instead of Catalogs.from_payload, which returns a Catalog instead of an array of Catalog objects that always had a length of one
• Claned up logging across all modules

Removed

• Catalogs.from_payload, replaced by Catalog.from_payload
• QUEUED as potential processing state

v0.3.1

[v0.3.1] - 2020-09-27

Changed

• output_options->collections is now optional, if not provided than item collections are not updated

Fixed

• process function will reraise any error occuring while setting up processing so can be retried with redrivepolicy

v0.3.0

[v0.3.0] - 2020-09-02

Changed

• Catalog.from_payload will get output payload from Batch via a separate file "_out.json" rather than the original, as of Cirrus 0.2.0 Batch processes will write output to this new file rather than overwriting the input file