Fix: Add deduplication to event_details CTE to prevent row fan-out

The event_details CTE was reading from market_details_raw without
deduplication. Since market_details_raw contains temporal snapshots
with multiple rows per event_ticker (indicated by last_updated_ts),
the LEFT JOIN on event_ticker was causing a fan-out, producing
duplicate ticker rows that violated the unique_combination_of_columns
test and cascaded incorrect duplicates into kalshi_market_trades.

Added QUALIFY with ROW_NUMBER() to keep only the latest row per
event_ticker based on last_updated_ts DESC, ensuring 1:1 join
cardinality and preventing duplicate ticker rows.

Author: cursoragent

dbt_subprojects/daily_spellbook/models/_projects/kalshi/kalshi_market_details.sql

@@ -54,6 +54,7 @@ with markets as (
 		strike_period,
 		last_updated_ts
 	from {{ source('kalshi', 'market_details_raw') }}
+	qualify row_number() over (partition by event_ticker order by last_updated_ts desc) = 1
 )
 
 select