Extend the Profile List database calls to include selectors

[jhrozek]

Summary

Extend the Profile List database calls to include selectors

Since the profile selectors need to be included when listing profiles, we need to extend the two profile list calls to list selectors as well.

Because the selectors are stored in a different table, but we still need to retrieve them in the same Go struct defined in models, we define a new SQL type in a migration that represents the selectors and tell sqlc to use it in sqlc.yaml.

This new Go struct needs a Scan method as well to convert raw SQL rows into structs as well.

Co-authored-by: Michelangelo Mori mmori@stacklok.com

Fixes: #3721

Change Type

• Bug fix (resolves an issue without affecting existing features)
• Feature (adds new functionality without breaking changes)
• Breaking change (may impact existing functionalities or require documentation updates)
• Documentation (updates or additions to documentation)
• Refactoring or test improvements (no bug fixes or new functionality)

Testing

There are unit tests and I tested the code as part of a larger branch, too.

Review Checklist:

• Reviewed my own code for quality and clarity.
• Added comments to complex or tricky code sections.
• Updated any affected documentation.
• Included tests that validate the fix or feature.
• Checked that related changes are merged.

[jhrozek]

@blkt 👀

[coveralls]

coverage: 52.011% (+0.009%) from 52.002%
when pulling a69dec2 on jhrozek:selector_db
into ff2be84 on stacklok:main.

[coveralls]

coverage: 52.213% (+0.009%) from 52.204%
when pulling 58a29a1 on jhrozek:selector_db
into 6274fe5 on stacklok:main.

[coveralls]

coverage: 52.208% (+0.004%) from 52.204%
when pulling 58a29a1 on jhrozek:selector_db
into 6274fe5 on stacklok:main.

[coveralls]

coverage: 52.208% (+0.004%) from 52.204%
when pulling 58a29a1 on jhrozek:selector_db
into 6274fe5 on stacklok:main.

[jhrozek]

@blkt thank you for the help on tidying up the models!

I did some follow-up cleanup of the code and I think it's ready for review now. I'm not sure if you can give the final ack given that you are the coauthor on the PR now.

[blkt]

LGTM! 🎉

[blkt]

question: why is this override necessary now while it wasn't before?

[jhrozek]

it wasn't, sorry.

[jhrozek]

(longer reply: This was a copy-paste from sqlc docs which I didn't remove by accident)

[blkt]

suggestion (non-blocking): I think you could get rid of the FILTER clause by moving the LEFT JOIN on the outer query. I expect it to work the same given the result is a slice.

Suggested change

	WITH helper AS(
	SELECT pr.id as profid,
	ARRAY_AGG(ROW(ps.id, ps.profile_id, ps.entity, ps.selector, ps.comment)::profile_selector) FILTER (WHERE ps.id IS NOT NULL) AS selectors
	FROM profiles pr
	LEFT JOIN profile_selectors ps
	ON pr.id = ps.profile_id
	WHERE pr.project_id = $1
	GROUP BY pr.id
	)
	SELECT
	sqlc.embed(profiles),
	sqlc.embed(profiles_with_entity_profiles),
	helper.selectors::profile_selector[] AS profiles_with_selectors
	FROM profiles
	JOIN profiles_with_entity_profiles ON profiles.id = profiles_with_entity_profiles.profid
	JOIN helper ON profiles.id = helper.profid
	WHERE profiles.project_id = $1;
	WITH helper AS(
	SELECT pr.id as profid,
	ARRAY_AGG(ROW(ps.id, ps.profile_id, ps.entity, ps.selector, ps.comment)::profile_selector) AS selectors
	FROM profiles pr
	JOIN profile_selectors ps
	ON pr.id = ps.profile_id
	WHERE pr.project_id = $1
	GROUP BY pr.id
	)
	SELECT
	sqlc.embed(profiles),
	sqlc.embed(profiles_with_entity_profiles),
	helper.selectors::profile_selector[] AS profiles_with_selectors
	FROM profiles
	JOIN profiles_with_entity_profiles ON profiles.id = profiles_with_entity_profiles.profid
	LEFT JOIN helper ON profiles.id = helper.profid
	WHERE profiles.project_id = $1;

[jhrozek]

suggestion taken!

[blkt]

suggestion (non-blocking): I think you could get rid of the FILTER clause by moving the LEFT JOIN on the outer query. I expect it to work the same given the result is a slice.

Suggested change

	WITH helper AS(
	SELECT pr.id as profid,
	ARRAY_AGG(ROW(ps.id, ps.profile_id, ps.entity, ps.selector, ps.comment)::profile_selector) FILTER (WHERE ps.id IS NOT NULL) AS selectors
	FROM profiles pr
	LEFT JOIN profile_selectors ps
	ON pr.id = ps.profile_id
	WHERE pr.project_id = $1
	GROUP BY pr.id
	)
	SELECT sqlc.embed(profiles),
	sqlc.embed(profiles_with_entity_profiles),
	helper.selectors::profile_selector[] AS profiles_with_selectors
	FROM profiles
	JOIN profiles_with_entity_profiles ON profiles.id = profiles_with_entity_profiles.profid
	JOIN helper ON profiles.id = helper.profid
	WITH helper AS(
	SELECT pr.id as profid,
	ARRAY_AGG(ROW(ps.id, ps.profile_id, ps.entity, ps.selector, ps.comment)::profile_selector) AS selectors
	FROM profiles pr
	JOIN profile_selectors ps
	ON pr.id = ps.profile_id
	WHERE pr.project_id = $1
	GROUP BY pr.id
	)
	SELECT sqlc.embed(profiles),
	sqlc.embed(profiles_with_entity_profiles),
	helper.selectors::profile_selector[] AS profiles_with_selectors
	FROM profiles
	JOIN profiles_with_entity_profiles ON profiles.id = profiles_with_entity_profiles.profid
	LEFT JOIN helper ON profiles.id = helper.profid

[jhrozek]

suggestion taken!

[coveralls]

coverage: 52.208% (+0.004%) from 52.204%
when pulling 15fc001 on jhrozek:selector_db
into 6274fe5 on stacklok:main.