feat: add palindrome_generation

reasoning_gym/algorithmic/__init__.py

@@ -12,6 +12,7 @@
 from .letter_jumble import LetterJumbleConfig, LetterJumbleDataset
 from .number_filtering import NumberFilteringConfig, NumberFilteringDataset
 from .number_sorting import NumberSortingConfig, NumberSortingDataset
+from .palindrome_generation import PalindromeConfig, PalindromeDataset
 from .sentence_reordering import SentenceReorderingConfig, SentenceReorderingDataset
 from .spell_backward import SpellBackwardConfig, SpellBackwardDataset
 from .word_ladder import WordLadderConfig, WordLadderDataset
@@ -42,4 +43,6 @@
     "TextTransformation",
     "WordLadderConfig",
     "WordLadderDataset",
+    "PalindromeConfig",
+    "PalindromeDataset",
 ]

reasoning_gym/algorithmic/palindrome_generation.py

@@ -0,0 +1,115 @@
+import random
+import string
+from dataclasses import dataclass
+from typing import Any, Dict, Optional
+
+from ..factory import ProceduralDataset, register_dataset
+
+
+@dataclass
+class PalindromeConfig:
+    """
+    Configuration for the palindrome task.
+
+    - min_length: Minimum length of the palindrome.
+    - max_length: Maximum length of the palindrome.
+    - seed: Optional seed for reproducibility.
+    - size: Number of palindrome samples in the virtual dataset.
+    """
+
+    min_length: int = 3
+    max_length: int = 10
+    seed: Optional[int] = None
+    size: int = 50
+
+    def validate(self) -> None:
+        """Validate configuration parameters."""
+        assert self.min_length >= 1, "min_length must be >= 1"
+        assert self.max_length >= self.min_length, "max_length must be >= min_length"
+
+
+class PalindromeDataset(ProceduralDataset):
+    """
+    Generates a set of letters that can be assembled into a palindrome.
+    """
+
+    def __init__(self, config: PalindromeConfig):
+        super().__init__(config=config, seed=config.seed, size=config.size)
+
+    def __getitem__(self, idx: int) -> dict:
+        """
+        Generate a single palindrome task.
+
+        Returns:
+            dict with:
+            - "question": Set of letters to form a palindrome.
+            - "answer": A correct palindrome.
+            - "metadata": Includes letter set and generated palindrome.
+        """
+        rng = random.Random(self.seed + idx)
+        length = rng.randint(self.config.min_length, self.config.max_length)
+        letters = self._generate_palindrome_letters(rng, length)
+        scrambled_letters = rng.sample(letters, len(letters))  # Scramble the order
+        palindrome = self._assemble_palindrome(letters)
+
+        question_str = (
+            "Rearrange these letters to form a palindrome. A palindrome is a word, phrase, or sequence that reads the same forward and backward.\n\n"
+            "For example, if the letters are: a, a, b — a valid palindrome is: aba.\n\n"
+            f"Your letters: {', '.join(scrambled_letters)}\n\n"
+            "What palindrome can you form from these letters?"
+        )
+
+        return {
+            "question": question_str,
+            "answer": palindrome,
+            "metadata": {
+                "letters": scrambled_letters,
+                "generated_palindrome": palindrome,
+            },
+        }
+
+    def _generate_palindrome_letters(self, rng: random.Random, length: int) -> list[str]:
+        """Generate a set of letters that can form a palindrome."""
+        half_length = length // 2
+        letters = rng.choices(string.ascii_lowercase, k=half_length)
+        if length % 2 == 1:
+            middle_letter = rng.choice(string.ascii_lowercase)
+            return letters + [middle_letter] + letters[::-1]
+        return letters + letters[::-1]
+
+    def _assemble_palindrome(self, letters: list[str]) -> str:
+        """Return the palindrome string from the letter set."""
+        return "".join(letters)
+
+    def score_answer(self, answer: Optional[str], metadata: Dict[str, Any]) -> float:
+        """Determine if the solution provided is a valid palindrome.
+        The answer is expected to be a single string
+
+        Expected behavior:
+        - Correct answer (palindrome with only correct letters in the correct quantities) gives 1.0
+        - An answer that is a palindrome, but not with the same letters as provided, gives 0.05
+        - An answer that is a string, but not a palindrome gives 0.02
+        - An empty string gives 0.01.
+        - None gives 0.0.
+        """
+        if answer is None or not isinstance(answer, str):
+            return 0.0  # No answer given
+
+        if answer == "":
+            return 0.01
+
+        answer = answer.strip().lower()
+        expected_letters = metadata["letters"]
+
+        # Check if the answer is a palindrome
+        if answer != answer[::-1]:
+            return 0.02
+
+        # Check if answer contains the same letters as provided (ignoring order)
+        if sorted(answer) != sorted(expected_letters):
+            return 0.05
+
+        return 1.0  # Correct solution
+
+
+register_dataset("palindrome", PalindromeDataset, PalindromeConfig)

tests/test_palindrome.py

@@ -0,0 +1,92 @@
+import pytest
+
+from reasoning_gym.algorithmic.palindrome_generation import PalindromeConfig, PalindromeDataset
+
+
+def test_palindrome_config_validation():
+    """Test that invalid configs raise appropriate errors"""
+    with pytest.raises(AssertionError):
+        config = PalindromeConfig(min_length=0)  # Too short
+        config.validate()
+
+    with pytest.raises(AssertionError):
+        config = PalindromeConfig(min_length=5, max_length=3)  # Invalid range
+        config.validate()
+
+
+def test_palindrome_deterministic():
+    """Test that dataset generates same items with same seed"""
+    config = PalindromeConfig(seed=42, size=10)
+    dataset1 = PalindromeDataset(config)
+    dataset2 = PalindromeDataset(config)
+
+    for i in range(len(dataset1)):
+        assert dataset1[i] == dataset2[i]
+
+
+def test_palindrome_items():
+    """Test basic properties of generated items"""
+    config = PalindromeConfig(min_length=3, max_length=7, size=10, seed=42)
+    dataset = PalindromeDataset(config)
+
+    for item in dataset:
+        assert isinstance(item, dict)
+        assert "question" in item
+        assert "answer" in item
+        assert "metadata" in item
+
+        # Check metadata contains required fields
+        assert "letters" in item["metadata"]
+        assert "generated_palindrome" in item["metadata"]
+
+        # Verify answer is a palindrome
+        palindrome = item["answer"]
+        assert palindrome == palindrome[::-1], f"{palindrome} is not a palindrome"
+
+
+def test_palindrome_randomization():
+    """Test letter randomization in the question"""
+    config = PalindromeConfig(min_length=4, max_length=4, size=10, seed=42)
+    dataset = PalindromeDataset(config)
+
+    for item in dataset:
+        letters = item["metadata"]["letters"]
+        palindrome = item["metadata"]["generated_palindrome"]
+
+        # Ensure the same letters are present but in different order
+        assert sorted(letters) == sorted(palindrome)
+
+
+def test_score_answer():
+    """Test the scoring mechanism for palindrome answers.
+
+    Expected behavior:
+    - Correct answer (palindrome with only correct letters in the correct quantities) gives 1.0
+    - An answer that is a palindrome, but not with the same letters as provided, gives 0.05
+    - An answer that is a string, but not a palindrome gives 0.02
+    - An empty string gives 0.01.
+    - None gives 0.0.
+    """
+    config = PalindromeConfig(min_length=4, max_length=6, size=10, seed=42)
+    dataset = PalindromeDataset(config)
+
+    for item in dataset:
+        correct_answer = item["answer"]
+        metadata = item["metadata"]
+
+        # Correct answer should score 1.0
+        assert dataset.score_answer(correct_answer, metadata) == 1.0
+
+        # Incorrect answer (palindrome, but not correct one) should score 0.05
+        pal_letters = "racecar" if "racecar" != correct_answer else "aba"
+        assert dataset.score_answer(pal_letters, metadata) == 0.05
+
+        # Incorrect answer (not palindrome) should score 0.02
+        wrong_letters = "abcd" if "abcd" != correct_answer else "efgh"
+        assert dataset.score_answer(wrong_letters, metadata) == 0.02
+
+        # Empty String input should score 0.01
+        assert dataset.score_answer("", metadata) == 0.01
+
+        # Empty input should score 0.0
+        assert dataset.score_answer(None, metadata) == 0.0