Fix logic in `find_name_mismatches.py` to correctly identify players missing from projections

- Updated the calculation of `players_missing_from_projections` to accurately reflect the difference between projection players and portfolio players.
- Adjusted the matching process to ensure that the closest matches are derived from the correct player list, enhancing the accuracy of name matching.

global_func/find_name_mismatches.py

@@ -23,7 +23,7 @@ def find_name_mismatches(contest_df, projections_df):
     projection_players_list = list(projection_players)
     
     # Find players in portfolio that are missing from projections
-    players_missing_from_projections = list(portfolio_players - projection_players)
+    players_missing_from_projections = list(projection_players -portfolio_players)
     
     # Automatically handle 100% matches before starting interactive process
     players_to_process = []
@@ -31,7 +31,7 @@ def find_name_mismatches(contest_df, projections_df):
         if not isinstance(player, str):
             st.warning(f"Skipping non-string value: {player}")
             continue
-        closest_matches = process.extract(player, projection_players_list, limit=1)
+        closest_matches = process.extract(player, portfolio_players, limit=1)
         if closest_matches[0][1] == 100:  # If perfect match found
             match_name = closest_matches[0][0]
             # Update all occurrences in contest_df