Dependency sweep: litellm 1.106.0, lancedb 0.40, pyarrow 26, refreshed lockfile

.semversioner/next-release/patch-20261026003724271651.json

@@ -0,0 +1,4 @@
+{
+  "type": "patch",
+  "description": "Update dependencies (litellm 1.106.0, lancedb 0.40, pyarrow 26, ruff 0.17) and fix resulting lint/type issues."
+}

packages/graphrag-llm/graphrag_llm/completion/completion.py

@@ -10,7 +10,7 @@
 from graphrag_llm.threading.completion_thread_runner import completion_thread_runner
 
 if TYPE_CHECKING:
-    from collections.abc import AsyncIterator, Iterator
+    from collections.abc import AsyncIterator, Generator, Iterator
 
     from graphrag_cache import Cache, CacheKeyCreator
 
@@ -167,7 +167,7 @@ def completion_thread_pool(
         response_handler: "ThreadedLLMCompletionResponseHandler",
         concurrency: int,
         queue_limit: int = 0,
-    ) -> "Iterator[ThreadedLLMCompletionFunction]":
+    ) -> "Generator[ThreadedLLMCompletionFunction, None, None]":
         """Run a completion thread pool.
 
         Args

packages/graphrag-llm/graphrag_llm/completion/lite_llm_completion.py

@@ -4,7 +4,7 @@
 """LLMCompletion based on litellm."""
 
 from collections.abc import AsyncIterator, Iterator
-from typing import TYPE_CHECKING, Any, Unpack
+from typing import TYPE_CHECKING, Any, Unpack, cast
 
 import litellm
 from azure.identity import DefaultAzureCredential, get_bearer_token_provider
@@ -22,6 +22,7 @@
 
 if TYPE_CHECKING:
     from graphrag_cache import Cache, CacheKeyCreator
+    from litellm import CustomStreamWrapper
 
     from graphrag_llm.config import ModelConfig
     from graphrag_llm.metrics import MetricsProcessor, MetricsStore
@@ -262,8 +263,11 @@ def _base_completion(
         if json_object and "response_format" not in new_args:
             new_args["response_format"] = {"type": "json_object"}
 
-        response = litellm.completion(
-            **new_args,
+        response = cast(
+            "ModelResponse | CustomStreamWrapper",
+            litellm.completion(
+                **new_args,
+            ),
         )
         if isinstance(response, ModelResponse):
             return LLMCompletionResponse(**response.model_dump())

packages/graphrag-llm/graphrag_llm/embedding/embedding.py

@@ -10,7 +10,7 @@
 from graphrag_llm.threading.embedding_thread_runner import embedding_thread_runner
 
 if TYPE_CHECKING:
-    from collections.abc import Iterator
+    from collections.abc import Generator
 
     from graphrag_cache import Cache, CacheKeyCreator
 
@@ -93,7 +93,7 @@ def embedding_thread_pool(
         response_handler: "ThreadedLLMEmbeddingResponseHandler",
         concurrency: int,
         queue_limit: int = 0,
-    ) -> "Iterator[ThreadedLLMEmbeddingFunction]":
+    ) -> "Generator[ThreadedLLMEmbeddingFunction, None, None]":
         """Run an embedding thread pool.
 
         Args

packages/graphrag-llm/graphrag_llm/embedding/lite_llm_embedding.py

@@ -3,7 +3,7 @@
 
 """LLMEmbedding based on litellm."""
 
-from typing import TYPE_CHECKING, Any, Unpack
+from typing import TYPE_CHECKING, Any, Unpack, cast
 
 import litellm
 from azure.identity import DefaultAzureCredential, get_bearer_token_provider
@@ -15,6 +15,7 @@
 
 if TYPE_CHECKING:
     from graphrag_cache import Cache, CacheKeyCreator
+    from litellm import EmbeddingResponse
 
     from graphrag_llm.config import ModelConfig
     from graphrag_llm.metrics import MetricsProcessor, MetricsStore
@@ -185,7 +186,7 @@ def _base_embedding(**kwargs: Any) -> LLMEmbeddingResponse:
         kwargs.pop("metrics", None)  # Remove metrics if present
         new_args: dict[str, Any] = {**base_args, **kwargs}
 
-        response = litellm.embedding(**new_args)
+        response = cast("EmbeddingResponse", litellm.embedding(**new_args))
         return LLMEmbeddingResponse(**response.model_dump())
 
     async def _base_embedding_async(**kwargs: Any) -> LLMEmbeddingResponse:

packages/graphrag-llm/graphrag_llm/rate_limit/sliding_window_rate_limiter.py

@@ -6,7 +6,7 @@
 import threading
 import time
 from collections import deque
-from collections.abc import Iterator
+from collections.abc import Generator
 from contextlib import contextmanager
 from typing import Any
 
@@ -62,7 +62,7 @@ def __init__(
             self._stagger = self._period_in_seconds / self._rpp
 
     @contextmanager
-    def acquire(self, token_count: int) -> Iterator[None]:
+    def acquire(self, token_count: int) -> Generator[None, None, None]:
         """
         Acquire Rate Limiter.
 

packages/graphrag-llm/graphrag_llm/threading/completion_thread_runner.py

@@ -7,7 +7,7 @@
 import sys
 import threading
 import time
-from collections.abc import Awaitable, Iterator
+from collections.abc import Awaitable, Generator, Iterator
 from contextlib import contextmanager
 from queue import Empty, Queue
 from typing import TYPE_CHECKING, Protocol, Unpack, runtime_checkable
@@ -147,7 +147,7 @@ def completion_thread_runner(
     concurrency: int,
     queue_limit: int = 0,
     metrics_store: "MetricsStore | None" = None,
-) -> Iterator[ThreadedLLMCompletionFunction]:
+) -> Generator[ThreadedLLMCompletionFunction, None, None]:
     """Run a completion thread pool.
 
     Args

packages/graphrag-llm/graphrag_llm/threading/embedding_thread_runner.py

@@ -7,7 +7,7 @@
 import sys
 import threading
 import time
-from collections.abc import Awaitable, Iterator
+from collections.abc import Awaitable, Generator
 from contextlib import contextmanager
 from queue import Empty, Queue
 from typing import TYPE_CHECKING, Protocol, Unpack, runtime_checkable
@@ -120,7 +120,7 @@ def embedding_thread_runner(
     concurrency: int,
     queue_limit: int = 0,
     metrics_store: "MetricsStore | None" = None,
-) -> Iterator[ThreadedLLMEmbeddingFunction]:
+) -> Generator[ThreadedLLMEmbeddingFunction, None, None]:
     """Run an embedding thread pool.
 
     Args

packages/graphrag-llm/pyproject.toml

@@ -38,7 +38,7 @@ dependencies = [
     "graphrag-cache==3.3.0",
     "graphrag-common==3.3.0",
     "jinja2~=3.1",
-    "litellm==1.100.1",
+    "litellm==1.106.0",
     "nest-asyncio2~=1.7",
     # orjson is required for litellm tool calling. https://github.com/BerriAI/litellm/issues/32993
     "orjson>=3.11.6,<4.0",

packages/graphrag-vectors/pyproject.toml

@@ -35,9 +35,9 @@ dependencies = [
     "azure-identity~=1.25",
     "azure-search-documents~=12.0",
     "graphrag-common==3.3.0",
-    "lancedb~=0.38.0",
+    "lancedb~=0.40.0",
     "numpy~=2.4",
-    "pyarrow~=25.0",
+    "pyarrow~=26.0",
     "pydantic~=2.13",
 ]
 

packages/graphrag/pyproject.toml

@@ -56,7 +56,7 @@ dependencies = [
     "nltk~=3.9.0",
     "numpy~=2.4",
     "pandas~=3.0",
-    "pyarrow~=25.0",
+    "pyarrow~=26.0",
     "pydantic~=2.13",
     "spacy~=3.8",
     "blis~=1.3",