chore(deps): lock file maintenance (#379)

Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com>
Co-authored-by: James Braza <[email protected]>

Author: renovate[bot]

pyproject.toml

@@ -48,7 +48,7 @@ dev = [
     "ipython>=8",  # Pin to keep recent
     "ldp[monitor,nn,rich,server,typing,visualization]",
     "litellm",  # Version ranges are enforced by lmi
-    "mypy>=1.8",  # Pin for mutable-override
+    "mypy>=1.19",  # Pin for zip default detection
     "prek",
     "pydantic~=2.9",  # Pydantic 2.9 changed JSON schema exports 'allOf', so ensure tests match
     "pylint-pydantic",

src/ldp/graph/ops.py

@@ -558,7 +558,7 @@ async def __call__(self, *args, **kwargs) -> OpResult[TOutput_co]:
             # Map positional arguments to keyword arguments to make backward pass easier
             for i_arg, (arg, param) in enumerate(
                 # strict=False b/c not all params in _fwd_args will be in args (i.e. defaults and **kwargs)
-                zip(args, self._fwd_args, strict=False)
+                zip(args, self._fwd_args, strict=False)  # noqa: FURB120
             ):
                 # Don't need to check for too many args or collisions with kwargs, since forward()
                 # will raise an exception anyway

src/ldp/graph/torch_ops.py

@@ -126,7 +126,7 @@ def store_tensor_inputs(
     # See Op.__call__ for some notes on what this is doing.
     for i_arg, (arg, param) in enumerate(
         # strict=False b/c not all params in _fwd_args will be in args (i.e. defaults and **kwargs)
-        zip(tensor_args, fwd_args, strict=False)
+        zip(tensor_args, fwd_args, strict=False)  # noqa: FURB120
     ):
         if param.kind == inspect.Parameter.VAR_POSITIONAL:
             ctx_args = list(tensor_args[i_arg:])

src/ldp/utils.py

@@ -78,7 +78,7 @@ def discounted_returns(
     """
     returns = []
     r = 0.0
-    for reward, term in zip(reversed(rewards), reversed(terminated), strict=False):
+    for reward, term in zip(reversed(rewards), reversed(terminated), strict=True):
         # 1 - term is 0 if the episode has terminated
         r = reward + discount * r * (1 - term)
         returns.append(r)