Added integration tests for meta (#274)

Added integration tests for `ChatBedrockConverse` for Meta provider. 

2 tests are primarily failing here. 
1. `test_tool_calling`, which is consistently returning a quoted string
input value instead of an expected integer e.g., {input: '3'} instead of
{input: 3}, thus failing the test. This needs further investigation to
determine whether this is a bug with the Bedrock service or an expected
behavior with Llama models when a numeric input is required.
2. `test_tool_message_histories_list_content`, which is failing for
several other providers as well, as Bedrock doesn't seem to allow
conversation block and tool use blocks in the same turn.
        ```python
botocore.errorfactory.ValidationException: An error occurred
(ValidationException) when calling the Converse operation:
messages.1.content: Conversation blocks and tool use blocks cannot be
provided in the same turn.
        ```

libs/aws/tests/integration_tests/chat_models/test_bedrock_converse.py

@@ -70,6 +70,44 @@ def test_tool_calling_with_no_arguments(self, model: BaseChatModel) -> None:
         pass
 
 
+class TestBedrockMetaStandard(ChatModelIntegrationTests):
+    @property
+    def chat_model_class(self) -> Type[BaseChatModel]:
+        return ChatBedrockConverse
+
+    @property
+    def chat_model_params(self) -> dict:
+        return {"model": "us.meta.llama3-2-90b-instruct-v1:0"}
+
+    @property
+    def standard_chat_model_params(self) -> dict:
+        return {"temperature": 0.1, "max_tokens": 100, "stop": []}
+
+    @pytest.mark.xfail(reason="Meta models don't support tool_choice.")
+    def test_structured_few_shot_examples(self, model: BaseChatModel) -> None:
+        pass
+
+    # TODO: This needs investigation, if this is a bug with Bedrock or Llama models,
+    # but this test consistently seem to return single quoted input values {input: '3'}
+    # instead of {input: 3} failing the test. Upon checking with tools with non-numeric
+    # inputs, tool calling seems to work as expected with Bedrock and Llama models.
+    @pytest.mark.xfail(
+        reason="Bedrock Meta models tend to return string values for integer inputs ."
+    )
+    def test_tool_calling(self, model: BaseChatModel) -> None:
+        super().test_tool_calling(model)
+
+    @pytest.mark.xfail(reason="Meta models don't support tool_choice.")
+    def test_tool_calling_with_no_arguments(self, model: BaseChatModel) -> None:
+        pass
+
+    @pytest.mark.xfail(
+        reason="Human messages following AI messages not supported by Bedrock."
+    )
+    def test_tool_message_histories_list_content(self, model: BaseChatModel) -> None:
+        super().test_tool_message_histories_list_content(model)
+
+
 def test_structured_output_snake_case() -> None:
     model = ChatBedrockConverse(
         model="anthropic.claude-3-sonnet-20240229-v1:0", temperature=0