core: Improve mypy config (#30737)

* Cleanup mypy config
* Add mypy `strict` rules except `disallow_any_generics`,
`warn_return_any` and `strict_equality` (TODO)
* Add mypy `strict_byte` rule
* Add mypy support for PEP702 `@deprecated` decorator
* Bump mypy version to 1.15

---------

Co-authored-by: Eugene Yurtsev <[email protected]>

Author: cbornet

libs/core/langchain_core/_api/beta_decorator.py

@@ -225,10 +225,8 @@ def finalize(wrapper: Callable[..., Any], new_doc: str) -> T:
         new_doc = f".. beta::\n   {details}\n\n{old_doc}\n"
 
         if inspect.iscoroutinefunction(obj):
-            finalized = finalize(awarning_emitting_wrapper, new_doc)
-        else:
-            finalized = finalize(warning_emitting_wrapper, new_doc)
-        return cast("T", finalized)
+            return finalize(awarning_emitting_wrapper, new_doc)
+        return finalize(warning_emitting_wrapper, new_doc)
 
     return beta
 

libs/core/langchain_core/_api/deprecation.py

@@ -152,7 +152,10 @@ def deprecate(
         _package: str = package,
     ) -> T:
         """Implementation of the decorator returned by `deprecated`."""
-        from langchain_core.utils.pydantic import FieldInfoV1, FieldInfoV2
+        from langchain_core.utils.pydantic import (  # type: ignore[attr-defined]
+            FieldInfoV1,
+            FieldInfoV2,
+        )
 
         def emit_warning() -> None:
             """Emit the warning."""
@@ -395,10 +398,8 @@ def finalize(wrapper: Callable[..., Any], new_doc: str) -> T:
 """
 
         if inspect.iscoroutinefunction(obj):
-            finalized = finalize(awarning_emitting_wrapper, new_doc)
-        else:
-            finalized = finalize(warning_emitting_wrapper, new_doc)
-        return cast("T", finalized)
+            return finalize(awarning_emitting_wrapper, new_doc)
+        return finalize(warning_emitting_wrapper, new_doc)
 
     return deprecate
 

libs/core/langchain_core/callbacks/manager.py

@@ -2408,7 +2408,7 @@ def _configure(
                         run_tree.trace_id,
                         run_tree.dotted_order,
                     )
-                    handler.run_map[str(run_tree.id)] = cast("Run", run_tree)
+                    handler.run_map[str(run_tree.id)] = run_tree
     for var, inheritable, handler_class, env_var in _configure_hooks:
         create_one = (
             env_var is not None

libs/core/langchain_core/document_loaders/base.py

@@ -80,7 +80,7 @@ async def alazy_load(self) -> AsyncIterator[Document]:
         iterator = await run_in_executor(None, self.lazy_load)
         done = object()
         while True:
-            doc = await run_in_executor(None, next, iterator, done)  # type: ignore[call-arg, arg-type]
+            doc = await run_in_executor(None, next, iterator, done)
             if doc is done:
                 break
             yield doc  # type: ignore[misc]

libs/core/langchain_core/embeddings/fake.py

@@ -52,7 +52,7 @@ class FakeEmbeddings(Embeddings, BaseModel):
     """The size of the embedding vector."""
 
     def _get_embedding(self) -> list[float]:
-        import numpy as np  # type: ignore[import-not-found, import-untyped]
+        import numpy as np
 
         return list(np.random.default_rng().normal(size=self.size))
 
@@ -109,7 +109,7 @@ class DeterministicFakeEmbedding(Embeddings, BaseModel):
     """The size of the embedding vector."""
 
     def _get_embedding(self, seed: int) -> list[float]:
-        import numpy as np  # type: ignore[import-not-found, import-untyped]
+        import numpy as np
 
         # set the seed for the random generator
         rng = np.random.default_rng(seed)

libs/core/langchain_core/globals.py

@@ -23,7 +23,7 @@ def set_verbose(value: bool) -> None:  # noqa: FBT001
         value: The new value for the `verbose` global setting.
     """
     try:
-        import langchain  # type: ignore[import]
+        import langchain  # type: ignore[import-not-found]
 
         # We're about to run some deprecated code, don't report warnings from it.
         # The user called the correct (non-deprecated) code path and shouldn't get
@@ -57,7 +57,7 @@ def get_verbose() -> bool:
         The value of the `verbose` global setting.
     """
     try:
-        import langchain  # type: ignore[import]
+        import langchain
 
         # We're about to run some deprecated code, don't report warnings from it.
         # The user called the correct (non-deprecated) code path and shouldn't get
@@ -96,7 +96,7 @@ def set_debug(value: bool) -> None:  # noqa: FBT001
         value: The new value for the `debug` global setting.
     """
     try:
-        import langchain  # type: ignore[import]
+        import langchain
 
         # We're about to run some deprecated code, don't report warnings from it.
         # The user called the correct (non-deprecated) code path and shouldn't get
@@ -128,7 +128,7 @@ def get_debug() -> bool:
         The value of the `debug` global setting.
     """
     try:
-        import langchain  # type: ignore[import]
+        import langchain
 
         # We're about to run some deprecated code, don't report warnings from it.
         # The user called the correct (non-deprecated) code path and shouldn't get
@@ -164,7 +164,7 @@ def set_llm_cache(value: Optional["BaseCache"]) -> None:
         value: The new LLM cache to use. If `None`, the LLM cache is disabled.
     """
     try:
-        import langchain  # type: ignore[import]
+        import langchain
 
         # We're about to run some deprecated code, don't report warnings from it.
         # The user called the correct (non-deprecated) code path and shouldn't get
@@ -198,7 +198,7 @@ def get_llm_cache() -> "BaseCache":
         The value of the `llm_cache` global setting.
     """
     try:
-        import langchain  # type: ignore[import]
+        import langchain
 
         # We're about to run some deprecated code, don't report warnings from it.
         # The user called the correct (non-deprecated) code path and shouldn't get

libs/core/langchain_core/indexing/api.py

@@ -394,7 +394,7 @@ def index(
                 if cleanup == "scoped_full":
                     scoped_full_cleanup_source_ids.add(source_id)
             # source ids cannot be None after for loop above.
-            source_ids = cast("Sequence[str]", source_ids)  # type: ignore[assignment]
+            source_ids = cast("Sequence[str]", source_ids)
 
         exists_batch = record_manager.exists([doc.uid for doc in hashed_docs])
 

libs/core/langchain_core/language_models/base.py

@@ -61,7 +61,7 @@ def get_tokenizer() -> Any:
     every time it is called.
     """
     try:
-        from transformers import GPT2TokenizerFast  # type: ignore[import]
+        from transformers import GPT2TokenizerFast  # type: ignore[import-not-found]
     except ImportError as e:
         msg = (
             "Could not import transformers python package. "

libs/core/langchain_core/language_models/chat_models.py

@@ -853,7 +853,7 @@ async def agenerate(
                         run_manager.on_llm_end(
                             LLMResult(
                                 generations=[res.generations],  # type: ignore[list-item, union-attr]
-                                llm_output=res.llm_output,  # type: ignore[list-item, union-attr]
+                                llm_output=res.llm_output,  # type: ignore[union-attr]
                             )
                         )
                         for run_manager, res in zip(run_managers, results)
@@ -1107,7 +1107,7 @@ async def _astream(
                 None,
                 next,
                 iterator,
-                done,  # type: ignore[call-arg, arg-type]
+                done,
             )
             if item is done:
                 break

libs/core/langchain_core/language_models/llms.py

@@ -455,7 +455,7 @@ def batch(
                 inputs[i : i + max_concurrency]
                 for i in range(0, len(inputs), max_concurrency)
             ]
-            config = [{**c, "max_concurrency": None} for c in config]  # type: ignore[misc]
+            config = [{**c, "max_concurrency": None} for c in config]
             return [
                 output
                 for i, batch in enumerate(batches)
@@ -501,7 +501,7 @@ async def abatch(
                 inputs[i : i + max_concurrency]
                 for i in range(0, len(inputs), max_concurrency)
             ]
-            config = [{**c, "max_concurrency": None} for c in config]  # type: ignore[misc]
+            config = [{**c, "max_concurrency": None} for c in config]
             return [
                 output
                 for i, batch in enumerate(batches)
@@ -746,7 +746,7 @@ async def _astream(
                 None,
                 next,
                 iterator,
-                done,  # type: ignore[call-arg, arg-type]
+                done,
             )
             if item is done:
                 break
@@ -1231,7 +1231,7 @@ async def agenerate(
                 stop,
                 run_managers,  # type: ignore[arg-type]
                 new_arg_supported=bool(new_arg_supported),
-                **kwargs,  # type: ignore[arg-type]
+                **kwargs,
             )
         if len(missing_prompts) > 0:
             run_managers = await asyncio.gather(
@@ -1253,7 +1253,7 @@ async def agenerate(
                 stop,
                 run_managers,  # type: ignore[arg-type]
                 new_arg_supported=bool(new_arg_supported),
-                **kwargs,  # type: ignore[arg-type]
+                **kwargs,
             )
             llm_output = await aupdate_cache(
                 self.cache,