update BaseChatModel to pass standard test_json_mode

[mattf]

Checked

• I searched existing ideas and did not find a similar one
• I added a very descriptive title
• I've clearly described the feature request and motivation for it

Feature request

the default implementation for with_structured_output raises a value error on any kwargs

https://github.com/langchain-ai/langchain/blob/master/libs/core/langchain_core/language_models/chat_models.py#L1174

https://github.com/langchain-ai/langchain/blob/master/libs/core/langchain_core/language_models/chat_models.py#L1283

the tests for json_mode pass method= as a kwargs

https://github.com/langchain-ai/langchain/blob/master/libs/standard-tests/langchain_tests/integration_tests/chat_models.py#L1487

the result is a ValueError, e.g. ValueError: Received unsupported arguments {'method': 'json_mode'}

request: allow ChatModels derived from BaseChatModel to pass test_json_mode without having to override with_structured_output

Motivation

i would like to implement a ChatModel derived from BaseChatModel and minimize maintenance by relying on default methods.

ChatModels that deviate from BaseChatModel's signature to support method param (incomplete list) -

• OpenAI - https://github.com/langchain-ai/langchain/blob/master/libs/partners/openai/langchain_openai/chat_models/base.py#L1391
• Perplexity - https://github.com/langchain-ai/langchain/blob/master/libs/community/langchain_community/chat_models/perplexity.py#L338
• OCI - https://github.com/langchain-ai/langchain/blob/master/libs/community/langchain_community/chat_models/oci_generative_ai.py#L656
• Ollama - https://github.com/langchain-ai/langchain/blob/master/libs/partners/ollama/langchain_ollama/chat_models.py#L830
• Groq - https://github.com/langchain-ai/langchain/blob/master/libs/partners/groq/langchain_groq/chat_models.py#L791
• Mistral - https://github.com/langchain-ai/langchain/blob/master/libs/partners/mistralai/langchain_mistralai/chat_models.py#L710

Proposal (If applicable)

No response

[mattf]

cc @ccurme