chore: Add type checking to test/human_in_the_loop, test/evaluation, test/document_stores and test/dataclasses

pyproject.toml

@@ -163,7 +163,7 @@ integration-only-slow = 'pytest --maxfail=5 -m "integration and slow" {args:test
 all = 'pytest {args:test}'
 
 # TODO We want to eventually type the whole test folder
-types = "mypy --install-types --non-interactive --cache-dir=.mypy_cache/ {args:haystack test/core/ test/marshal/ test/testing/ test/tracing/}"
+types = "mypy --install-types --non-interactive --cache-dir=.mypy_cache/ {args:haystack test/core/ test/marshal/ test/testing/ test/tracing/ test/human_in_the_loop test/evaluation test/document_stores test/dataclasses}"
 
 [tool.hatch.envs.e2e]
 template = "test"

test/dataclasses/test_chat_message.py

@@ -4,6 +4,7 @@
 
 import json
 import warnings
+from collections.abc import Sequence
 
 import pytest
 
@@ -306,7 +307,7 @@ def test_from_assistant_with_reasoning_string(self):
 
     def test_from_assistant_with_invalid_reasoning(self):
         with pytest.raises(TypeError):
-            ChatMessage.from_assistant(text="text", reasoning=123)
+            ChatMessage.from_assistant(text="text", reasoning=123)  # type: ignore[arg-type]
 
     def test_from_user_with_valid_content(self):
         text = "I have a question."
@@ -354,7 +355,7 @@ def test_from_user_empty_text(self):
         assert message.texts == [""]
 
     def test_from_user_with_content_parts(self, base64_image_string, base64_pdf_string):
-        content_parts = [
+        content_parts: list[TextContent | ImageContent | FileContent | str] = [
             TextContent(text="text"),
             ImageContent(base64_image=base64_image_string),
             FileContent(base64_data=base64_pdf_string),
@@ -393,7 +394,7 @@ def test_from_user_with_content_parts(self, base64_image_string, base64_pdf_stri
     def test_from_user_with_content_parts_fails_unsupported_parts(self):
         with pytest.raises(TypeError):
             ChatMessage.from_user(
-                content_parts=["text part", ToolCall(id="123", tool_name="mytool", arguments={"a": 1})]
+                content_parts=["text part", ToolCall(id="123", tool_name="mytool", arguments={"a": 1})]  # type: ignore[list-item]
             )
 
     def test_from_user_with_content_parts_fails_with_empty_parts(self):
@@ -442,7 +443,10 @@ def test_from_tool_with_valid_content(self):
         assert not message.reasoning
 
     def test_from_tool_with_valid_mixed_content(self, base64_image_string):
-        tool_result = [TextContent(text="Hello"), ImageContent(base64_image=base64_image_string, mime_type="image/png")]
+        tool_result: list[TextContent | ImageContent] = [
+            TextContent(text="Hello"),
+            ImageContent(base64_image=base64_image_string, mime_type="image/png"),
+        ]
         message = ChatMessage.from_tool(
             tool_result=tool_result, origin=ToolCall(tool_name="mytool", arguments={}), error=False
         )
@@ -472,7 +476,10 @@ def test_multiple_text_segments(self):
         assert len(message) == 2
 
     def test_mixed_content(self):
-        content = [TextContent(text="Hello"), ToolCall(id="123", tool_name="mytool", arguments={"a": 1})]
+        content: list[TextContent | ToolCall] = [
+            TextContent(text="Hello"),
+            ToolCall(id="123", tool_name="mytool", arguments={"a": 1}),
+        ]
 
         message = ChatMessage(_role=ChatRole.ASSISTANT, _content=content)
 
@@ -485,16 +492,16 @@ def test_mixed_content(self):
 
     def test_from_function_class_method_removed(self):
         with pytest.raises(AttributeError):
-            ChatMessage.from_function("Result of function invocation", "my_function")
+            ChatMessage.from_function("Result of function invocation", "my_function")  # type: ignore[attr-defined]
 
     def test_chat_message_content_attribute_removed(self):
         message = ChatMessage.from_user(text="This is a message")
         with pytest.raises(AttributeError):
-            message.content
+            message.content  # type: ignore[attr-defined]
 
     def test_chat_message_init_parameters_removed(self):
         with pytest.raises(TypeError):
-            ChatMessage(role="irrelevant", content="This is a message")
+            ChatMessage(role="irrelevant", content="This is a message")  # type: ignore[call-arg]
 
     def test_no_warning_on_init(self):
         with warnings.catch_warnings():
@@ -572,7 +579,7 @@ def test_to_dict_with_invalid_content_type(self):
         text_content = TextContent(text="Hello")
         invalid_content = "invalid"
 
-        message = ChatMessage(_role=ChatRole.ASSISTANT, _content=[text_content, invalid_content])
+        message = ChatMessage(_role=ChatRole.ASSISTANT, _content=[text_content, invalid_content])  # type: ignore[list-item]
 
         with pytest.raises(TypeError):
             message.to_dict()
@@ -888,7 +895,7 @@ def test_to_openai_dict_format_require_tool_call_ids_false(self):
         assert openai_msg == {"role": "tool", "content": "result"}
 
     def test_to_openai_dict_format_tool_message_list_with_unsupported_image(self, base64_image_string):
-        tool_result = [
+        tool_result: Sequence[TextContent | ImageContent] = [
             TextContent(text="first result"),
             ImageContent(base64_image=base64_image_string, mime_type="image/png"),
         ]
@@ -940,6 +947,7 @@ def test_from_openai_dict_format_tool_message(self):
         openai_msg = {"role": "tool", "content": "The weather is sunny", "tool_call_id": "call_123"}
         message = ChatMessage.from_openai_dict_format(openai_msg)
         assert message.role.value == "tool"
+        assert message.tool_call_result is not None
         assert message.tool_call_result.result == "The weather is sunny"
         assert message.tool_call_result.origin.id == "call_123"
 
@@ -951,13 +959,15 @@ def test_from_openai_dict_format_tool_message_list(self):
         }
         message = ChatMessage.from_openai_dict_format(openai_msg)
         assert message.role.value == "tool"
+        assert message.tool_call_result is not None
         assert message.tool_call_result.result == [TextContent(text="first result"), TextContent(text="second result")]
         assert message.tool_call_result.origin.id == "call_123"
 
     def test_from_openai_dict_format_tool_without_id(self):
         openai_msg = {"role": "tool", "content": "The weather is sunny"}
         message = ChatMessage.from_openai_dict_format(openai_msg)
         assert message.role.value == "tool"
+        assert message.tool_call_result is not None
         assert message.tool_call_result.result == "The weather is sunny"
         assert message.tool_call_result.origin.id is None
 

test/dataclasses/test_document.py

@@ -37,7 +37,7 @@ def test_init():
 
 def test_init_with_wrong_parameters():
     with pytest.raises(TypeError):
-        Document(text="")
+        Document(text="")  # type: ignore[call-arg]
 
 
 def test_init_with_parameters():
@@ -53,6 +53,7 @@ def test_init_with_parameters():
     )
     assert doc.id == "1aa43af57c1dbc317241bf55d3067049f334d3b458d95dc72f71a7111f6c1a56"
     assert doc.content == "test text"
+    assert doc.blob is not None
     assert doc.blob.data == blob_data
     assert doc.blob.mime_type == "text/markdown"
     assert doc.meta == {"text": "test text"}
@@ -88,7 +89,7 @@ def test_init_with_legacy_field():
     doc = Document(
         content="test text",
         content_type="text",  # type: ignore
-        id_hash_keys=["content"],  # type: ignore
+        id_hash_keys=["content"],
         score=0.812,
         embedding=[0.1, 0.2, 0.3],
         meta={"date": "10-10-2023", "type": "article"},
@@ -265,7 +266,7 @@ def test_from_dict_with_legacy_field_and_flat_meta():
     ) == Document(
         content="test text",
         content_type="text",  # type: ignore
-        id_hash_keys=["content"],  # type: ignore
+        id_hash_keys=["content"],
         score=0.812,
         embedding=[0.1, 0.2, 0.3],
         meta={"date": "10-10-2023", "type": "article"},

test/dataclasses/test_streaming_chunk.py

@@ -15,15 +15,16 @@
     ToolCallDelta,
     ToolCallResult,
 )
+from haystack.dataclasses.streaming_chunk import FinishReason
 
 
 @component
 class ExampleComponent:
     def __init__(self):
         self.name = "test_component"
 
-    def run(self) -> str:
-        return "Test content"
+    def run(self) -> dict[str, str]:
+        return {"test": "Test content"}
 
 
 def test_create_chunk_with_content_and_metadata():
@@ -143,7 +144,7 @@ def test_create_chunk_with_finish_reason_and_meta():
 
 def test_finish_reason_standard_values():
     """Test all standard finish_reason values including the new Haystack-specific ones."""
-    standard_values = ["stop", "length", "tool_calls", "content_filter", "tool_call_results"]
+    standard_values: list[FinishReason] = ["stop", "length", "tool_calls", "content_filter", "tool_call_results"]
 
     for value in standard_values:
         chunk = StreamingChunk(content="Test content", finish_reason=value)
@@ -271,8 +272,10 @@ def test_from_dict_tool_call_result():
     assert chunk.content == ""
     assert chunk.meta == {"key": "value"}
     assert chunk.index == 0
+    assert chunk.component_info is not None
     assert chunk.component_info.type == "test_streaming_chunk.ExampleComponent"
     assert chunk.component_info.name == "test_component"
+    assert chunk.tool_call_result is not None
     assert chunk.tool_call_result.result == "output"
     assert chunk.tool_call_result.error is False
     assert chunk.tool_call_result.origin.id == "123"
@@ -298,8 +301,10 @@ def test_from_dict_tool_calls():
     assert chunk.content == ""
     assert chunk.meta == {"key": "value"}
     assert chunk.index == 0
+    assert chunk.component_info is not None
     assert chunk.component_info.type == "test_streaming_chunk.ExampleComponent"
     assert chunk.component_info.name == "test_component"
+    assert chunk.tool_calls is not None
     assert chunk.tool_calls[0].tool_name == "test_tool"
     assert chunk.tool_calls[0].index == 0
     assert chunk.finish_reason == "tool_calls"
@@ -325,8 +330,10 @@ def test_from_dict_reasoning():
     assert chunk.content == ""
     assert chunk.meta == {"key": "value"}
     assert chunk.index == 0
+    assert chunk.component_info is not None
     assert chunk.component_info.type == "test_streaming_chunk.ExampleComponent"
     assert chunk.component_info.name == "test_component"
+    assert chunk.reasoning is not None
     assert chunk.reasoning.reasoning_text == "thinking"
     assert chunk.reasoning.extra["step"] == 1
     assert chunk.finish_reason == "stop"

test/document_stores/test_filter_policy.py

@@ -2,6 +2,8 @@
 #
 # SPDX-License-Identifier: Apache-2.0
 
+from typing import Literal
+
 import pytest
 
 from haystack.document_stores.types import FilterPolicy, apply_filter_policy
@@ -168,7 +170,7 @@ def test_merge_comparison_filters_with_same_field():
 
 
 @pytest.mark.parametrize("logical_operator", ["AND", "OR", "NOT"])
-def test_merge_with_custom_logical_operator(logical_operator: str):
+def test_merge_with_custom_logical_operator(logical_operator: Literal["AND", "OR", "NOT"]) -> None:
     """
     Merging with a custom logical operator