mirror of
https://github.com/open-thought/reasoning-gym.git
synced 2025-10-09 13:40:09 +03:00
dced3bfc45
* init * fix tests * unify codeio * filtered for libraries not present in reasoning-gym * fix more bounds * puzzle24 * knight swap curriculum * fix number sorting * fix attributes * add validation of config in creation of dataset * dry run for instantiating and validating the datasets * remove unused imports * fix curriculum tests to reference newly updated attribute names
294 lines
11 KiB
Python
294 lines
11 KiB
Python
"""Word ladder task generator"""
|
|
|
|
from collections import deque
|
|
from dataclasses import dataclass
|
|
from random import Random
|
|
from typing import Any, Optional
|
|
|
|
from ..coaching import BaseCurriculum, RangeAttributeDefinition
|
|
from ..data import get_data_file_path
|
|
from ..factory import ProceduralDataset, register_dataset
|
|
|
|
QUESTION_TEMPLATE = """Transform the word ladder '{start}' to '{end}' by changing one letter at a time.
|
|
Provide your answer as a comma-separated sequence of uppercase letters without spaces.
|
|
Each step must be a valid English word."""
|
|
|
|
|
|
DATASET_NAME = "word_ladder"
|
|
|
|
|
|
@dataclass
|
|
class WordLadderConfig:
|
|
"""Configuration for word ladder task generation"""
|
|
|
|
min_word_length: int = 4 # Minimum word length
|
|
max_word_length: int = 4 # Maximum word length
|
|
min_chain_length: int = -1 # Set to -1 for shortest path or a minimum of 3
|
|
max_chain_length: int = -1 # Set to -1 for shortest path or a max
|
|
seed: Optional[int] = None
|
|
size: int = 500
|
|
|
|
def validate(self) -> None:
|
|
"""Validate configuration parameters"""
|
|
assert self.min_word_length >= 3, "min_word_length must be >= 3"
|
|
assert self.max_word_length >= self.min_word_length, "max_word_length must be >= min_word_length"
|
|
assert self.max_word_length <= 5, "max_word_length must be <= 5"
|
|
|
|
# Add size validation
|
|
if self.size > 20000: # Add reasonable upper limit
|
|
raise ValueError("Dataset size too large for this algorithm and constraints")
|
|
|
|
# Modified validation logic
|
|
if self.min_chain_length == -1:
|
|
if self.max_chain_length != -1:
|
|
assert (
|
|
self.max_chain_length >= 3
|
|
), "When min_chain_length=-1 (shortest path), max_chain_length must be -1 or >=3"
|
|
elif self.max_chain_length == -1:
|
|
raise AssertionError("max_chain_length cannot be -1 unless min_chain_length is also -1")
|
|
else:
|
|
assert self.min_chain_length >= 3, "min_chain_length must be 3 or -1"
|
|
assert self.max_chain_length >= self.min_chain_length, "max_chain_length must be >= min_chain_length"
|
|
|
|
def is_valid_path_length(self, length: int) -> bool:
|
|
"""Check if a path length meets the configuration requirements"""
|
|
# When min_chain_length is -1, we accept any path of length >= 3
|
|
if self.min_chain_length == -1:
|
|
if self.max_chain_length == -1:
|
|
return length >= 3
|
|
return 3 <= length <= self.max_chain_length
|
|
|
|
# Otherwise check against both min and max
|
|
return (
|
|
self.min_chain_length <= length <= (self.max_chain_length if self.max_chain_length != -1 else float("inf"))
|
|
)
|
|
|
|
|
|
class WordLadderDataset(ProceduralDataset):
|
|
"""Generates word ladder transformation tasks"""
|
|
|
|
def __init__(self, config: WordLadderConfig):
|
|
self.config = config
|
|
self.word_sets = {}
|
|
self.word_graphs = {}
|
|
self._vocabulary = None # A large list of dictionary words to validate words against
|
|
|
|
# Load words from CSV
|
|
self.word_sets = self._load_words_from_csv(
|
|
min_length=self.config.min_word_length, max_length=self.config.max_word_length
|
|
)
|
|
|
|
# Precompute word graphs for all lengths
|
|
for length in range(self.config.min_word_length, self.config.max_word_length + 1):
|
|
self.word_graphs[length] = self._build_word_graph(length)
|
|
|
|
config.validate()
|
|
super().__init__(config=config, seed=config.seed, size=config.size)
|
|
|
|
@classmethod
|
|
def _load_words_from_csv(cls, min_length: int = 3, max_length: int = 5) -> dict[int, set[str]]:
|
|
"""Load words from CSV file organized by length"""
|
|
# Validate length range before processing
|
|
assert 3 <= min_length <= max_length <= 5, "Word length must be between 3 and 5 inclusive"
|
|
|
|
import csv
|
|
|
|
word_sets = {}
|
|
|
|
try:
|
|
# Get CSV content as string
|
|
with get_data_file_path("words.csv").open("r", encoding="utf-8") as csv_file:
|
|
reader = csv.DictReader(csv_file)
|
|
|
|
for row in reader:
|
|
# Process each word length column using config range
|
|
for length in range(min_length, max_length + 1):
|
|
col_name = f"{length}_letter"
|
|
word = row.get(col_name, "")
|
|
|
|
if not word: # Skip empty entries
|
|
continue
|
|
|
|
word_sets.setdefault(length, set()).add(word.upper())
|
|
|
|
except Exception as e:
|
|
raise RuntimeError(f"Error processing words.csv content: {e}") from e
|
|
|
|
# Validate we have words for each length
|
|
for length in range(min_length, max_length + 1):
|
|
if length not in word_sets or not word_sets[length]:
|
|
raise ValueError(f"No valid words found for length {length}")
|
|
|
|
return word_sets
|
|
|
|
def _get_neighbors(self, word: str, word_set: set[str]) -> set[str]:
|
|
"""Get neighbors from either precomputed graph or by computing on demand"""
|
|
# Try precomputed graph first
|
|
if len(word) in self.word_graphs and word in self.word_graphs[len(word)]:
|
|
return self.word_graphs[len(word)].get(word, set())
|
|
|
|
# Fall back to computing neighbors directly for custom word sets
|
|
neighbors = set()
|
|
for i in range(len(word)):
|
|
for c in "ABCDEFGHIJKLMNOPQRSTUVWXYZ":
|
|
neighbor = word[:i] + c + word[i + 1 :]
|
|
if neighbor != word and neighbor in word_set:
|
|
neighbors.add(neighbor)
|
|
return neighbors
|
|
|
|
def _build_word_graph(self, word_length: int) -> dict[str, set[str]]:
|
|
"""Build graph of word connections for given length, using caching"""
|
|
# Return cached graph if it exists
|
|
if word_length in self.word_graphs:
|
|
return self.word_graphs[word_length]
|
|
|
|
# Build new graph
|
|
word_set = self.word_sets[word_length]
|
|
graph = {}
|
|
|
|
# Build connections
|
|
for word in word_set:
|
|
neighbors = set()
|
|
for i in range(word_length):
|
|
for c in "ABCDEFGHIJKLMNOPQRSTUVWXYZ":
|
|
neighbor = word[:i] + c + word[i + 1 :]
|
|
if neighbor != word and neighbor in word_set:
|
|
neighbors.add(neighbor)
|
|
graph[word] = neighbors
|
|
|
|
# Cache and return
|
|
self.word_graphs[word_length] = graph
|
|
return self.word_graphs[word_length]
|
|
|
|
def _find_path(self, start: str, end: str, word_set: set[str]) -> Optional[list[str]]:
|
|
"""Simplified path finding using BFS for shortest paths"""
|
|
# Early exit if words are direct neighbors
|
|
if end in self._get_neighbors(start, word_set):
|
|
return [start, end]
|
|
|
|
# Use basic BFS for shortest path
|
|
queue = deque([(start, [start])])
|
|
visited = {start}
|
|
|
|
while queue:
|
|
current, path = queue.popleft()
|
|
if current == end:
|
|
if self.config.is_valid_path_length(len(path)):
|
|
return path
|
|
return None
|
|
|
|
for neighbor in self._get_neighbors(current, word_set):
|
|
if neighbor not in visited:
|
|
visited.add(neighbor)
|
|
new_path = path + [neighbor]
|
|
queue.append((neighbor, new_path))
|
|
|
|
return None
|
|
|
|
def _generate_word_pair(self, rng: Random, length: int) -> tuple[str, str, list[str]]:
|
|
"""Simplified word pair generation"""
|
|
word_set = self.word_sets[length]
|
|
words_list = sorted(word_set)
|
|
max_attempts = 100
|
|
|
|
for _ in range(max_attempts):
|
|
start = rng.choice(words_list)
|
|
end = rng.choice(words_list)
|
|
|
|
if start == end:
|
|
continue
|
|
|
|
path = self._find_path(start, end, word_set)
|
|
if path:
|
|
return start, end, path
|
|
|
|
raise RuntimeError(f"Failed to find valid pair for length {length}")
|
|
|
|
def __getitem__(self, idx: int) -> dict:
|
|
"""Generate a single word ladder task"""
|
|
if idx >= self.size:
|
|
raise IndexError(f"Dataset index {idx} out of range for size {self.size}")
|
|
|
|
try:
|
|
rng = Random(self.seed + idx)
|
|
length = rng.randint(self.config.min_word_length, self.config.max_word_length)
|
|
start, end, path = self._generate_word_pair(rng, length)
|
|
except RuntimeError as e:
|
|
# If we run out of valid paths, adjust the virtual size
|
|
self.size = idx
|
|
raise IndexError(f"Dataset exhausted at index {idx}. {str(e)}")
|
|
|
|
return {
|
|
"question": QUESTION_TEMPLATE.format(start=start, end=end),
|
|
"answer": ",".join(path),
|
|
"metadata": {
|
|
"source_dataset": DATASET_NAME,
|
|
"source_index": idx,
|
|
"start_word": start,
|
|
"end_word": end,
|
|
"word_length": length,
|
|
"chain_length": len(path),
|
|
"difficulty": {
|
|
"word_length": (self.config.min_word_length, self.config.max_word_length),
|
|
},
|
|
},
|
|
}
|
|
|
|
def score_answer(self, answer: Optional[str], entry: dict[str, Any]) -> float:
|
|
if not isinstance(answer, str):
|
|
return 0.0
|
|
|
|
answer_words = tuple(s.strip() for s in answer.upper().split(","))
|
|
|
|
metadata = entry["metadata"]
|
|
start_word = metadata["start_word"]
|
|
end_word = metadata["end_word"]
|
|
word_length = len(end_word)
|
|
known_words = self.word_sets[word_length]
|
|
|
|
# Check conditions:
|
|
# 1. start and end word match question
|
|
# 2. all words have the correct length
|
|
# 3. every changed word is a single letter change from the previous word
|
|
# 4. all words are in our vocabulary
|
|
|
|
if len(answer_words) < 2:
|
|
return 0.0
|
|
|
|
if answer_words[0] != start_word or answer_words[-1] != end_word:
|
|
return 0.0
|
|
|
|
if not all(len(w) == word_length for w in answer_words):
|
|
return 0.0
|
|
|
|
for i in range(1, len(answer_words)):
|
|
if sum(1 for a, b in zip(answer_words[i - 1], answer_words[i]) if a != b) != 1:
|
|
return 0.0
|
|
|
|
reward = 1.0
|
|
for word in answer_words:
|
|
if not word in known_words:
|
|
reward *= 0.5
|
|
|
|
return reward
|
|
|
|
|
|
class WordLadderCurriculum(BaseCurriculum):
|
|
def __init__(self):
|
|
super().__init__(WordLadderCurriculum.__name__, WordLadderConfig)
|
|
|
|
# Define attributes
|
|
self._define_attributes(
|
|
RangeAttributeDefinition(
|
|
name="word_length",
|
|
levels=[3, 4, 5],
|
|
description="Length of words in the puzzle",
|
|
lower_field_name="min_word_length",
|
|
upper_field_name="max_word_length",
|
|
ensure_interval=True,
|
|
)
|
|
)
|
|
|
|
|
|
register_dataset(DATASET_NAME, WordLadderDataset, WordLadderConfig, WordLadderCurriculum)
|