chore!: llama_cpp - drop Python 3.9 and use X|Y typing (#2710)

Author: anakin87

integrations/llama_cpp/pyproject.toml

@@ -7,7 +7,7 @@ name = "llama-cpp-haystack"
 dynamic = ["version"]
 description = 'An integration between the llama.cpp LLM framework and Haystack'
 readme = "README.md"
-requires-python = ">=3.9"
+requires-python = ">=3.10"
 license = "Apache-2.0"
 keywords = []
 authors = [
@@ -18,15 +18,14 @@ classifiers = [
     "License :: OSI Approved :: Apache Software License",
     "Development Status :: 4 - Beta",
     "Programming Language :: Python",
-    "Programming Language :: Python :: 3.9",
     "Programming Language :: Python :: 3.10",
     "Programming Language :: Python :: 3.11",
     "Programming Language :: Python :: 3.12",
     "Programming Language :: Python :: 3.13",
     "Programming Language :: Python :: Implementation :: CPython",
     "Programming Language :: Python :: Implementation :: PyPy",
 ]
-dependencies = ["haystack-ai>=2.19.0", "llama-cpp-python>=0.2.87"]
+dependencies = ["haystack-ai>=2.22.0", "llama-cpp-python>=0.2.87"]
 
 # On macOS GitHub runners, we use a custom index to download pre-built wheels.
 # Installing from source might fail due to missing dependencies (CMake fails with "OpenMP not found")
@@ -92,7 +91,6 @@ known-first-party = ["haystack_integrations"]
 
 
 [tool.ruff]
-target-version = "py39"
 line-length = 120
 
 [tool.ruff.lint]
@@ -136,10 +134,6 @@ ignore = [
     "PLR0913",
     "PLR0915",
 ]
-unfixable = [
-    # Don't touch unused imports
-    "F401",
-]
 
 [tool.ruff.lint.flake8-tidy-imports]
 ban-relative-imports = "parents"

integrations/llama_cpp/src/haystack_integrations/components/generators/llama_cpp/chat/chat_generator.py

@@ -1,7 +1,7 @@
 import json
 from collections.abc import Iterator
 from datetime import datetime, timezone
-from typing import Any, Optional, Union
+from typing import Any
 
 from haystack import component, default_from_dict, default_to_dict, logging
 from haystack.components.generators.utils import _convert_streaming_chunks_to_chat_message
@@ -192,15 +192,15 @@ class LlamaCppChatGenerator:
     def __init__(
         self,
         model: str,
-        n_ctx: Optional[int] = 0,
-        n_batch: Optional[int] = 512,
-        model_kwargs: Optional[dict[str, Any]] = None,
-        generation_kwargs: Optional[dict[str, Any]] = None,
+        n_ctx: int | None = 0,
+        n_batch: int | None = 512,
+        model_kwargs: dict[str, Any] | None = None,
+        generation_kwargs: dict[str, Any] | None = None,
         *,
-        tools: Optional[ToolsType] = None,
-        streaming_callback: Optional[StreamingCallbackT] = None,
-        chat_handler_name: Optional[str] = None,
-        model_clip_path: Optional[str] = None,
+        tools: ToolsType | None = None,
+        streaming_callback: StreamingCallbackT | None = None,
+        chat_handler_name: str | None = None,
+        model_clip_path: str | None = None,
     ):
         """
         :param model: The path of a quantized model for text generation, for example, "zephyr-7b-beta.Q4_0.gguf".
@@ -238,7 +238,7 @@ def __init__(
 
         _check_duplicate_tool_names(flatten_tools_or_toolsets(tools))
 
-        handler: Optional[Llava15ChatHandler] = None
+        handler: Llava15ChatHandler | None = None
         # Validate multimodal requirements
         if chat_handler_name is not None:
             if model_clip_path is None:
@@ -256,7 +256,7 @@ def __init__(
         self.n_batch = n_batch
         self.model_kwargs = model_kwargs
         self.generation_kwargs = generation_kwargs
-        self._model: Optional[Llama] = None
+        self._model: Llama | None = None
         self.tools = tools
         self.streaming_callback = streaming_callback
         self.chat_handler_name = chat_handler_name
@@ -324,10 +324,10 @@ def from_dict(cls, data: dict[str, Any]) -> "LlamaCppChatGenerator":
     def run(
         self,
         messages: list[ChatMessage],
-        generation_kwargs: Optional[dict[str, Any]] = None,
+        generation_kwargs: dict[str, Any] | None = None,
         *,
-        tools: Optional[ToolsType] = None,
-        streaming_callback: Optional[StreamingCallbackT] = None,
+        tools: ToolsType | None = None,
+        streaming_callback: StreamingCallbackT | None = None,
     ) -> dict[str, list[ChatMessage]]:
         """
         Run the text generation model on the given list of ChatMessages.
@@ -435,8 +435,8 @@ def _handle_streaming_response(
 
             if chunk.get("choices") and len(chunk["choices"]) > 0:
                 choice = chunk["choices"][0]
-                delta: Union[ChatCompletionStreamResponseDelta, ChatCompletionStreamResponseDeltaEmpty, dict] = (
-                    choice.get("delta", {})
+                delta: ChatCompletionStreamResponseDelta | ChatCompletionStreamResponseDeltaEmpty | dict = choice.get(
+                    "delta", {}
                 )
 
                 finish_reason = choice.get("finish_reason")

integrations/llama_cpp/src/haystack_integrations/components/generators/llama_cpp/generator.py

@@ -1,4 +1,4 @@
-from typing import Any, Optional, Union
+from typing import Any
 
 from haystack import component, logging
 
@@ -28,10 +28,10 @@ class LlamaCppGenerator:
     def __init__(
         self,
         model: str,
-        n_ctx: Optional[int] = 0,
-        n_batch: Optional[int] = 512,
-        model_kwargs: Optional[dict[str, Any]] = None,
-        generation_kwargs: Optional[dict[str, Any]] = None,
+        n_ctx: int | None = 0,
+        n_batch: int | None = 512,
+        model_kwargs: dict[str, Any] | None = None,
+        generation_kwargs: dict[str, Any] | None = None,
     ):
         """
         :param model: The path of a quantized model for text generation, for example, "zephyr-7b-beta.Q4_0.gguf".
@@ -62,16 +62,16 @@ def __init__(
         self.n_batch = n_batch
         self.model_kwargs = model_kwargs
         self.generation_kwargs = generation_kwargs
-        self.model: Optional[Llama] = None
+        self.model: Llama | None = None
 
     def warm_up(self):
         if self.model is None:
             self.model = Llama(**self.model_kwargs)
 
     @component.output_types(replies=list[str], meta=list[dict[str, Any]])
     def run(
-        self, prompt: str, generation_kwargs: Optional[dict[str, Any]] = None
-    ) -> dict[str, Union[list[str], list[dict[str, Any]]]]:
+        self, prompt: str, generation_kwargs: dict[str, Any] | None = None
+    ) -> dict[str, list[str] | list[dict[str, Any]]]:
         """
         Run the text generation model on the given prompt.