perf: group standings entries by round once instead of filtering per round

[Rathoz]

Summary

• StandingsParser.parse previously filtered the full entry list per round three separate times (tiebreakers, placements, final-round definite statuses); replaced with a single entriesByRound grouping pass before those loops.
• Standings.makeRounds previously ran Array.filter over all entries once per round at display time; replaced with one grouping pass keyed by tonumber(entry.roundindex) (the tonumber is preserved since records read from LPDB/wiki vars can have string roundindex).
• Entry order within each round is preserved by sequential insertion, so placement resolution is unchanged.

Based on standings-tests-ai (PR #7647).

How did you test this change?

• busted --run=ci: 605 successes / 0 failures / 0 errors — spec/standings_parser_spec.lua and spec/standings_model_spec.lua lock placements, ordering, and round structure.
• luacheck on both changed files: 0 warnings / 0 errors.

🤖 Generated with Claude Code

[ElectricalBoy]

Lua-Modules/lua/wikis/commons/Standings.lua

Lines 212 to 221 in d9dd250

	Lpdb.executeMassQuery(
	'standingsentry',
	{
	conditions = conditions:toString(),
	order = 'roundindex asc',
	},
	function(record)
	table.insert(standingsEntries, record)
	end
	)

adjusting LPDB query condition to have the returned data to already be sorted would be more optimal, right? like adjusting order to roundindex asc, slotindex asc
then because all data is guaranteed to be sorted we could switch over to Array.groupAdjacentBy

[ElectricalBoy]

Array.groupBy

[Rathoz]

Done in dede56f — switched both manual grouping loops (here and the matching one in Parser.lua) to Array.groupBy.

[Rathoz]

Applied the query-order change (roundindex asc, slotindex asc) and switched to Array.groupBy in dede56f.

groupAdjacentBy unfortunately can't be used here: makeRounds consumes standings.entryRecords, which comes from the LPDB query or from the page-variable path (getStandingsTable reading standings2_N). The var-path entries are in parser output order, which is opponent-major (flatMap(opponents, map(rounds))) — not sorted by round — so adjacency grouping would split each round into one group per opponent run. Same reason the per-round sortBy(position) stays. The grouping has to remain order-independent, which keyed Array.groupBy gives us.