fix: honor zero values in Dataset.reset_seeds (#9906)

* Fix Dataset.reset_seeds ignoring zero seed/size values

reset_seeds used `x or self.x` to decide whether to apply each argument,
so passing 0 (a valid seed or size) was treated as falsy and silently
dropped, keeping the previous value instead. Use an explicit None check
so 0 is honored.

* Apply suggestion from @greptile-apps[bot]

Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com>
Signed-off-by: Isaac Miller <17116851+isaacbmiller@users.noreply.github.com>

---------

Signed-off-by: Isaac Miller <17116851+isaacbmiller@users.noreply.github.com>
Co-authored-by: alvinttang <alvin@pm.me>
Co-authored-by: Isaac Miller <17116851+isaacbmiller@users.noreply.github.com>
Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com>

Author: 4 people

dspy/datasets/dataset.py

@@ -39,12 +39,12 @@ def reset_seeds(
         dev_size: int | None = None,
         test_size: int | None = None,
     ) -> None:
-        self.train_size = train_size or self.train_size
-        self.train_seed = train_seed or self.train_seed
-        self.dev_size = dev_size or self.dev_size
-        self.dev_seed = eval_seed or self.dev_seed
-        self.test_size = test_size or self.test_size
-        self.test_seed = eval_seed or self.test_seed
+        self.train_size = train_size if train_size is not None else self.train_size
+        self.train_seed = train_seed if train_seed is not None else self.train_seed
+        self.dev_size = dev_size if dev_size is not None else self.dev_size
+        self.dev_seed = eval_seed if eval_seed is not None else self.dev_seed
+        self.test_size = test_size if test_size is not None else self.test_size
+        self.test_seed = eval_seed if eval_seed is not None else self.test_seed
 
         if hasattr(self, "_train_"):
             del self._train_

tests/datasets/test_dataset.py

@@ -36,6 +36,31 @@ def csv_file():
         yield tmp_file.name
 
 
+def test_reset_seeds_accepts_zero():
+    dataset = Dataset(train_seed=5, train_size=10, eval_seed=5, dev_size=10, test_size=10)
+    dataset.reset_seeds(train_seed=0, train_size=0, eval_seed=0, dev_size=0, test_size=0)
+
+    assert dataset.train_seed == 0
+    assert dataset.train_size == 0
+    assert dataset.dev_seed == 0
+    assert dataset.dev_size == 0
+    assert dataset.test_seed == 0
+    assert dataset.test_size == 0
+
+
+def test_reset_seeds_keeps_existing_values_when_omitted():
+    dataset = Dataset(train_seed=5, train_size=10, eval_seed=7, dev_size=20, test_size=30)
+    dataset.reset_seeds(train_seed=1)
+
+    assert dataset.train_seed == 1
+    # Unspecified values should be left untouched.
+    assert dataset.train_size == 10
+    assert dataset.dev_seed == 7
+    assert dataset.dev_size == 20
+    assert dataset.test_seed == 7
+    assert dataset.test_size == 30
+
+
 @pytest.mark.extra
 def test_input_keys(csv_file):
     dataset = CSVDataset(csv_file, input_keys=["content", "question"])