DSPy: The framework for programming—not prompting—language models
Python
Add typing to `dspy.datasets.dataset` (#9143)
Max Muoto4193506899db
diff --git a/dspy/datasets/dataset.py b/dspy/datasets/dataset.pyindex f31fda4aa..75407bd5e 100644--- a/dspy/datasets/dataset.py+++ b/dspy/datasets/dataset.py@@ -1,12 +1,24 @@+from __future__ import annotations+import randomimport uuid+from collections.abc import Iterable+from typing import Anyfrom dspy import Examplefrom dspy.dsp.utils import dotdictclass Dataset:- def __init__(self, train_seed=0, train_size=None, eval_seed=0, dev_size=None, test_size=None, input_keys=None):+ def __init__(+ self,+ train_seed: int = 0,+ train_size: int | None = None,+ eval_seed: int = 0,+ dev_size: int | None = None,+ test_size: int | None = None,+ input_keys: list[str] | None = None,+ ) -> None:self.train_size = train_sizeself.train_seed = train_seedself.dev_size = dev_size@@ -19,7 +31,14 @@ class Dataset:self.name = self.__class__.__name__- def reset_seeds(self, train_seed=None, train_size=None, eval_seed=None, dev_size=None, test_size=None):+ def reset_seeds(+ self,+ train_seed: int | None = None,+ train_size: int | None = None,+ eval_seed: int | None = None,+ dev_size: int | None = None,+ test_size: int | None = None,+ ) -> None:self.train_size = train_size or self.train_sizeself.train_seed = train_seed or self.train_seedself.dev_size = dev_size or self.dev_size@@ -37,39 +56,41 @@ class Dataset:del self._test_@property- def train(self):+ def train(self) -> list[Example]:if not hasattr(self, "_train_"):self._train_ = self._shuffle_and_sample("train", self._train, self.train_size, self.train_seed)return self._train_@property- def dev(self):+ def dev(self) -> list[Example]:if not hasattr(self, "_dev_"):self._dev_ = self._shuffle_and_sample("dev", self._dev, self.dev_size, self.dev_seed)return self._dev_@property- def test(self):+ def test(self) -> list[Example]:if not hasattr(self, "_test_"):self._test_ = self._shuffle_and_sample("test", self._test, self.test_size, self.test_seed)return self._test_- def _shuffle_and_sample(self, split, data, size, seed=0):- data = list(data)+ def _shuffle_and_sample(+ self, split: str, data: Iterable[dict[str, Any]], size: int | None, seed: int = 0+ ) -> list[Example]:+ data_list = list(data)# Shuffle the data irrespective of the requested size.base_rng = random.Random(seed)if self.do_shuffle:- base_rng.shuffle(data)+ base_rng.shuffle(data_list)- data = data[:size]- output = []+ data_list = data_list[:size]+ output: list[Example] = []- for example in data:+ for example in data_list:example_obj = Example(**example, dspy_uuid=str(uuid.uuid4()), dspy_split=split)if self.input_keys:example_obj = example_obj.with_inputs(*self.input_keys)@@ -84,19 +105,20 @@ class Dataset:@classmethoddef prepare_by_seed(cls,- train_seeds=None,- train_size=16,- dev_size=1000,- divide_eval_per_seed=True,- eval_seed=2023,- **kwargs,- ):+ train_seeds: list[int] | None = None,+ train_size: int = 16,+ dev_size: int = 1000,+ divide_eval_per_seed: bool = True,+ eval_seed: int = 2023,+ **kwargs: Any,+ ) -> Any:train_seeds = train_seeds or [1, 2, 3, 4, 5]data_args = dotdict(train_size=train_size, eval_seed=eval_seed, dev_size=dev_size, test_size=0, **kwargs)dataset = cls(**data_args)eval_set = dataset.dev- eval_sets, train_sets = [], []+ eval_sets: list[list[Example]] = []+ train_sets: list[list[Example]] = []examples_per_seed = dev_size // len(train_seeds) if divide_eval_per_seed else dev_sizeeval_offset = 0