[dataquery] Dedup recent queries by content so "No run time" eliminates duplicates

[HenriRabalais]

Description

The "No run times (eliminate duplicates)" filter in the Recent Queries list was deduping on QueryID, but that doesn't actually catch duplicates. Identical queries can get saved under different QueryIDs, since the server dedupes saved queries by an exact string match on the query JSON, so any non-semantic difference (columns added in a different order, an optional visits key) ends up as a distinct QueryID. Keying on QueryID just leaves the duplicates in the list.

Changes

• Dedup the recent queries by a signature of their actual content (columns + filters) instead of QueryID. Columns and filter terms are sorted so their order doesn't matter.
• When two collapse, keep the named one as the representative if there is one, otherwise the most recent run.

Testing Instructions

1. Run a query.
2. Name it in the Recent Queries section.
3. Run the same query again from scratch.
4. Go to Recent Queries and tick "No run times (eliminate duplicates)".
5. The two runs now collapse into a single row (showing the named query).

Related Issues

Resolves #10600

[HenriRabalais]

@CamilleBeau PR has ben made for #10600