Add string representation of CRS as a new metadata field in C3 products

[robbibt]

It would be valuable for users to be able to query GA Collection 3 data by a dataset's native CRS. For example, this could be used as a filter to load only EPSG:32656 data along a UTM boundary where both EPSG:32655 and EPSG:32656 CRSs exist (e.g. to ensure all data is loaded in native CRS with no resampling).

This could be resolved by:

• Adding a new metadata field to GA Collection 3 products called crs_raw or crs_str that contains a string representation of the dataset's EPSG code (e.g. epsg:32656). This would allow allow filtering based on native CRS when loading data using dc.load:

dc.load(product='ga_ls8c_ard_3',
        crs_str='epsg:32656',
        **query)

This would be required for the following GA Collection 3 products only (with the longer term aim to also include this metadata for Sentinel-2 too if those products are upgraded to become part of Collection 3 in the future):

• ga_ls5t_ard_3
• ga_ls7e_ard_3
• ga_ls8c_ard_3
• ga_ls_fc_3
• ga_ls_wo_3

This change would also need to be implemented on both the NCI and Sandbox/AWS environments so that data can be loaded consistently across DEA environments.

[jeremyh]

• Added to the definitions in #281

• Deployed to NCI datacube:

  
  ❯ datacube metadata update digitalearthau/config/eo3/eo3*.odc-type.yaml
  Updated "eo3_landsat_ard"
  Updated "eo3_landsat_l1"
  Updated "eo3"
  ❯ date
  Thu Mar  4 14:56:09 AEDT 2021

❯ # Can we see it in the fields? Yes! ❯ datacube dataset search product=ga_ls8c_ard_3 --limit 1 id: 9781d762-6f48-4da5-b1a4-3657bb5abdfc product: ga_ls8c_ard_3 status: active locations:

• file:///g/data/xu18/ga/ga_ls8c_ard_3/106/068/2019/12/16/ga_ls8c_ard_3-2-0_106068_2019-12-16_final.odc-metadata.yaml fields: cloud_cover: 9.841478561653918 creation_time: 2021-02-22 06:16:16.113579+00:00 crs_raw: epsg:32652 dataset_maturity: final ... etc

  
  ---

@whatnick Who should we assign to run the same update on AWS+Sandbox datacubes?

[alexgleith]

Probably needed to have assigned me, @jeremyh!

[jeremyh]

Deployed by @omad to the AWS sandbox and dev databases.

Closing issue (reopen if you find problems)