DM-29940: Add TAP table_index for DP0.1

[hsinfang]

No description provided.

[gpdf]

This looks right to me and reflects the order documented on DM-29763 by @MelissaGraham et al.

Added @brianv0 to the review to make sure we're using tap:table_index correctly. Perhaps he could try running Felis-to-TAP_SCHEMA conversion on it to make sure the TAP_SCHEMA.tables table comes out as expected?

[gpdf]

I found something we'd all overlooked in previous reviews. It should say "Data Preview 0.1" in the top-level description. Fixed.

[brianv0]

Looks good to me, I ran and checked output:

INSERT INTO tables (schema_name, table_name, table_type, utype, description, table_index) VALUES ('dp01_dc2_catalogs', 'dp01_dc2_catalogs.position', 'table', NULL, 'Select astrometry-related parameters for objects detected in the coadded images, such as coordinates, footprints, patch/tract information, and deblending parameters. (29 columns)', 2);
INSERT INTO tables (schema_name, table_name, table_type, utype, description, table_index) VALUES ('dp01_dc2_catalogs', 'dp01_dc2_catalogs.reference', 'table', NULL, 'Measurements for objects detected in the coadded images, including photometry, astrometry, shape, deblending, model fits, and related background and flag parameters. This table joined with the position table is very similar to the object table, but with additional columns. (236 columns)', 3);
INSERT INTO tables (schema_name, table_name, table_type, utype, description, table_index) VALUES ('dp01_dc2_catalogs', 'dp01_dc2_catalogs.forced_photometry', 'table', NULL, 'Forced photometry measurements for objects detected in the coadded images, at the locations defined by the position table. (747 columns)', 4);
INSERT INTO tables (schema_name, table_name, table_type, utype, description, table_index) VALUES ('dp01_dc2_catalogs', 'dp01_dc2_catalogs.object', 'table', NULL, 'The object table from the DESC DC2 simulated sky survey as described in arXiv:2101.04855. Includes astrometric and photometric parameters for objects detected in coadded images. (137 columns)', 0);
INSERT INTO tables (schema_name, table_name, table_type, utype, description, table_index) VALUES ('dp01_dc2_catalogs', 'dp01_dc2_catalogs.truth_match', 'table', NULL, 'The truth-match table for the DESC DC2''s object table as described in arXiv:2101.04855. Includes the noiseless astrometric and photometric parameters and the best matches to the object table. (30 columns)', 1);

[hsinfang]

Thanks @gpdf for fixing the DP0 -> DP0.1 in the description. I'll create a PR to have that fix in https://github.com/lsst-sqre/lsst-tap-service

Since we are on this, comparing the felis outputs:
INSERT INTO tables (schema_name, table_name, table_type, utype, description, table_index) VALUES ('dp01_dc2_catalogs', 'dp01_dc2_catalogs.position', 'table', NULL, 'Select astrometry-related parameters for objects detected in the coadded images, such as coordinates, footprints, patch/tract information, and deblending parameters. (29 columns)', 2);

versus what is in https://github.com/lsst-sqre/lsst-tap-service/blob/master/docker/tap-schema/1015_dp01_dc2_tap_schema.sql#L2
INSERT INTO tap_schema.tables11 (schema_name, table_name, table_type, utype, description, table_index) VALUES ('dp01_dc2_catalogs', 'dp01_dc2_catalogs.position', 'table', null, 'Select astrometry-related parameters for objects detected in the coadded images, such as coordinates, footprints, patch/tract information, and deblending parameters. (29 columns)', 2);

Note the difference in the beginning:
INSERT INTO tables versus INSERT INTO tap_schema.tables11

I've been fixing this manually. May this be absorbed into felis so no manual fixup is involved?

[hsinfang]

lsst-sqre/lsst-tap-service#55

[gpdf]

Note the difference in the beginning:
INSERT INTO tables versus INSERT INTO tap_schema.tables11

I've been fixing this manually. May this be absorbed into felis so no manual fixup is involved?

"Manually" meaning, there isn't even a script checked in somewhere to do that transformation?

I'm not sure what the "11" suffix is about on the actual TAP service. If we could get a simple script written and stored (in lsst-tap-service?) for this, I think that's OK for now. I would hold off on making changes to Felis itself until the larger questions about how we are going to manage TAP_SCHEMA better are resolved.

[hsinfang]

It's a simple sed command.

Sounds good to me to hold it off for the larger question.