include ranges rather than sampled values in difficulty metadata dicts (#387)

* update difficulty metadata for logic datasets

* update difficulty metadata for graph datasets

* update difficulty metadata for geometry datasets

* update difficulty metadata for games datasets

* update difficulty metadata for cognition datasets

* update difficulty metadata for arithmetic datasets

* update difficulty metadata for arc datasets

* update difficulty metadata for algorithmic datasets

* update difficulty metadata for algebra datasets

* use tuples

* update tests

* update tests

reasoning_gym/games/boxnet.py

@@ -126,11 +126,14 @@ class BoxnetDataset(ProceduralDataset):
             "question": question,
             "answer": None,
             "metadata": {
-                "difficulty": {
-                    "row_num": row_num,
-                    "column_num": column_num,
-                },
+                "row_num": row_num,
+                "column_num": column_num,
                 "initial_state": pg_dict,
+                "difficulty": {
+                    "row_num": (self.config.min_row_num, self.config.max_row_num),
+                    "column_num": (self.config.min_column_num, self.config.max_column_num),
+                    "box_num": (self.config.min_box_num, self.config.max_box_num),
+                },
             },
         }
 

reasoning_gym/games/emoji_mystery.py

@@ -194,7 +194,10 @@ class EmojiMysteryDataset(ProceduralDataset):
             "answer": secret_sentence,
             "metadata": {
                 "emoji": secret_emoji,
-                "difficulty": {"num_words_in_sentence": len(re.findall(r"\b\w+\b", secret_sentence))},
+                "num_words_in_sentence": len(re.findall(r"\b\w+\b", secret_sentence)),
+                "difficulty": {
+                    "num_words_in_sentence": (self.config.min_words_in_sentence, self.config.max_words_in_sentence),
+                },
             },
         }
 

reasoning_gym/games/futoshiki.py

@@ -84,7 +84,12 @@ class FutoshikiDataset(ProceduralDataset):
                 "puzzle": puzzle,
                 "constraints": constraints,
                 "solution": solution,
-                "difficulty": {"board_size": board_size, "difficulty": difficulty},
+                "board_size": board_size,
+                "difficulty_rating": difficulty,
+                "difficulty": {
+                    "board_size": (self.config.min_board_size, self.config.max_board_size),
+                    "difficulty": (self.config.min_difficulty, self.config.max_difficulty),
+                },
             },
         }
 

reasoning_gym/games/mahjong.py

@@ -122,7 +122,9 @@ class MahjongPuzzleDataset(ProceduralDataset):
             "metadata": {
                 "rounds": rounds,
                 "solution": answer,
-                "difficulty": {"num_rounds": num_rounds},
+                "difficulty": {
+                    "num_rounds": (self.config.min_num_rounds, self.config.max_num_rounds),
+                },
             },
         }
 

reasoning_gym/games/maze.py

@@ -112,8 +112,8 @@ class MazeDataset(ProceduralDataset):
                         "wall": self.wall_char,
                         "path": self.path_char,
                         "difficulty": {
-                            "dist": dist,
-                            "grid_size": size,
+                            "dist": (self.config.min_dist, self.config.max_dist),
+                            "grid_size": (self.config.min_grid_size, self.config.max_grid_size),
                         },
                     },
                 }

reasoning_gym/games/mini_sudoku.py

@@ -197,7 +197,7 @@ class MiniSudokuDataset(ProceduralDataset):
                 "solution": solved_board,
                 "num_empty": num_empty,
                 "difficulty": {
-                    "empty": num_empty,
+                    "empty": (self.config.min_empty, self.config.max_empty),
                 },
             },
         }

reasoning_gym/games/n_queens.py

@@ -137,7 +137,7 @@ class NQueensDataset(ProceduralDataset):
                 "valid_answers": valid_solutions_str,
                 "difficulty": {
                     "n": self.config.n,
-                    "num_removed": num_removed,
+                    "num_removed": (self.config.min_remove, self.config.max_remove),
                 },
             },
         }

reasoning_gym/games/rush_hour.py

@@ -161,7 +161,9 @@ class RushHourDataset(ProceduralDataset):
             "metadata": {
                 "board_config": board_config,
                 "min_moves": min_moves,
-                "difficulty": {"min_moves": min_moves},
+                "difficulty": {
+                    "min_moves": (self.config.min_moves, self.config.max_moves),
+                },
             },
         }
 

reasoning_gym/games/sokoban.py

@@ -65,7 +65,7 @@ class SokobanDataset(ProceduralDataset):
 
         # Make the Sokoban!
         rng = Random(self.seed + idx)
-        gamestr, solution, difficulty = self._generate(
+        gamestr, solution, puzzle_data = self._generate(
             rng=rng,
             min_w=self.config.min_w,
             min_h=self.config.min_h,
@@ -93,7 +93,15 @@ Here is your puzzle:
 """
             + gamestr,
             "answer": solution,
-            "metadata": {"gamestr": gamestr, "difficulty": difficulty},
+            "metadata": {
+                "gamestr": gamestr,
+                "width": puzzle_data["width"],
+                "height": puzzle_data["height"],
+                "difficulty": {
+                    "width": (self.config.min_w, self.config.max_w),
+                    "height": (self.config.min_h, self.config.max_h),
+                },
+            },
         }
 
     def score_answer(self, answer: Optional[str], entry: dict[str, Any]) -> float:

reasoning_gym/games/sudoku.py

@@ -216,7 +216,7 @@ class SudokuDataset(ProceduralDataset):
                 "solution": solved_board,
                 "num_empty": num_empty,
                 "difficulty": {
-                    "num_empty": num_empty,
+                    "empty": (self.config.min_empty, self.config.max_empty),
                 },
             },
         }

reasoning_gym/games/tower_of_hanoi.py

@@ -275,6 +275,9 @@ class HanoiDataset(ProceduralDataset):
                 "target_peg": target_peg,
                 "auxiliary_pegs": auxiliary_pegs,
                 "solution_length": len(solution),
+                "difficulty": {
+                    "num_disks": (self.min_disks, self.max_disks),
+                },
             },
         }
 

reasoning_gym/games/tsumego.py

@@ -270,7 +270,13 @@ class TsumegoDataset(ProceduralDataset):
                 "Specify your move in coordinates (e.g. 'C4' for column C, row 4)"
             ),
             "answer": solution_str,
-            "metadata": {"difficulty": {"board_size": size}, "board": board},
+            "metadata": {
+                "board": board,
+                "board_size": size,
+                "difficulty": {
+                    "board_size": (self.config.min_board_size, self.config.max_board_size),
+                },
+            },
         }
 
     def score_answer(self, answer: Optional[str], entry: dict[str, Any]) -> float: