Improve tracker source and first-party cleanup

Surface every CookieSource value in the console: the trackers page
filter was missing the HTTP option and the source badge helper had no
EXTENSION case, so HTTP-sourced rows could not be filtered and
extension-sourced rows rendered the raw enum string.

On the backend, the mark-first-party verdict now blanks the stale
description on both the catalog row and its uncategorised org tracker
patterns. A terminal non-third-party row keeps no vendor link, so a
description naming the (now-cleared) vendor would be misleading; the
mapping worker only copies descriptions into empty rows and never
clears them, so clearing is done explicitly here.

Signed-off-by: Émile Ré <emile@probo.com>
This commit is contained in:
Émile Ré
2026-06-18 19:37:54 +02:00
parent 3b777e985c
commit 265c56d00b
5 changed files with 97 additions and 7 deletions

View File

@@ -1015,6 +1015,38 @@ WHERE
return result.RowsAffected(), nil
}
// ClearDescriptionByIDs blanks the researched description on the given
// catalog rows without re-arming enrichment or touching the enrichment
// payload. It backs the first-party verdict: a terminal non-third-party
// row keeps no vendor link, so a description that named the (now-cleared)
// vendor would be stale. The verdict is terminal, so rather than re-derive
// a description - which would re-run the mapping agent and re-link a
// vendor - the description simply returns to empty. Returns the number of
// rows updated.
func (ps *CommonTrackerPatterns) ClearDescriptionByIDs(
ctx context.Context,
tx pg.Tx,
ids []gid.GID,
) (int64, error) {
q := `
UPDATE common_tracker_patterns
SET
description = '',
updated_at = NOW()
WHERE
id = ANY(@ids)
`
args := pgx.StrictNamedArgs{"ids": ids}
result, err := tx.Exec(ctx, q, args)
if err != nil {
return 0, fmt.Errorf("cannot clear common tracker pattern description: %w", err)
}
return result.RowsAffected(), nil
}
// RequestEnrichmentByIDs arms enrichment on the given common tracker
// patterns by stamping enrichment_requested_at, which is the only column
// the enrichment worker claims on. It resets enrichment_attempts to 0 so

View File

@@ -1406,6 +1406,49 @@ WHERE
return result.RowsAffected(), nil
}
// ClearDescriptionForUncategorisedByCommonTrackerPatternIDs blanks the
// description on the uncategorised org tracker patterns linked to the
// given common tracker patterns. It pairs with the first-party verdict on
// the catalog row: when the catalog description is cleared because its
// vendor attribution was wrong, the descriptions fanned out to org
// patterns named the same stale vendor and must be cleared too - the
// mapping worker only ever copies a description into an empty org row, it
// never clears one. Like the mapping re-arm it is a global catalog
// operation, so it is intentionally not tenant-scoped, and it leaves
// excluded and user-categorised patterns untouched. The cookie_categories
// subquery is used only for filtering. Returns the number of org patterns
// cleared.
func (tps *TrackerPatterns) ClearDescriptionForUncategorisedByCommonTrackerPatternIDs(
ctx context.Context,
tx pg.Tx,
commonIDs []gid.GID,
) (int64, error) {
q := `
UPDATE tracker_patterns
SET
description = '',
updated_at = NOW()
WHERE
common_tracker_pattern_id = ANY(@common_ids)
AND excluded = false
AND cookie_category_id IN (
SELECT id FROM cookie_categories WHERE kind = @uncategorised_kind
)
`
args := pgx.StrictNamedArgs{
"common_ids": commonIDs,
"uncategorised_kind": CookieCategoryKindUncategorised,
}
result, err := tx.Exec(ctx, q, args)
if err != nil {
return 0, fmt.Errorf("cannot clear uncategorised tracker pattern descriptions: %w", err)
}
return result.RowsAffected(), nil
}
// RequestMappingForUnmappedByInitiatorDomains re-arms mapping on the
// still-unmapped org tracker patterns whose detected trackers share one
// of the given initiator domains, so the mapping worker re-resolves them