cedar-datacube 0.0.4

Fixed bug in cedar clean command

The cedar convert program will now copy “pre-ARD” image metadata JSON files to the destination directory alongside the “pre-ARD” images (GeoTIFFs converted to NetCDF files). To prevent this behavior, pass --skip-metadata to the program.

Changes were made in where exceptions are handled for the case where a user requests a “pre-ARD” image that returns 0 search results (e.g., wrong time period for sensor, bad historic coverage, etc). By default, Order.add will not raise an EmptyCollectionError when you try to add a “pre-ARD” image that had 0 image results. You can force an error to be raised by setting error_if_empty=True in either Order.add or Tracker.submit.

BREAKING CHANGE - Require image collection filters to be specified in the configuration file according to the image collection. This change has been made to prepare for using image collections that do not have the same filters (e.g., more than just LANDSAT/*/C01/T1_SR data)

Add a check to make sure that Order.name_template creates unique names for pre-ARD before continuing. If there are duplicate names, raises ValueError

Fix warning when calculating mean/std runtime of tasks that haven’t been updated (RuntimeWarning: Mean of empty slice…)

Provide better info in cedar status list when nothing is tracked because the tracking folder doesn’t exist

Fix bug when ordering but not using filters that resulted in 0 images being found

Validate image collection names using click argument callback in cedar submit CLI

Improve error propagation inside cedar.ordering.Order and cedar.tracker.Tracker. Order.add no longer catches and swallows EmptyCollectionError, which now is handled in Tracker.submit

Fix bug in jsonschema validation by allowing tuples & lists to count as ‘array’

Added cedar status cancel command to cancel orders

Refactor internals to use TrackingMetadata model

Fix packaging issue (missing package data)

Fix packaging issue

First public release