community: add AzureAIContentSafetyChain

[Sheepsta300]

Thank you for contributing to LangChain!

• PR title: community: add AzureAIContentSafetyChain

• PR message:

  • Description: The Azure AI Content Safety service checks input against the Azure Content Safety policy. By adding a runnable implementation of this service customers can utilise it in a chain, with other services to monitor input before it's passed to a chat model of some sort. This PR aims to further expose Azure services to LangChain and to allow Azure customers to use The Content Safety service in yet another way.
  • Dependencies: no dependencies added to the class, azure.ai.contentsafety has been added to the extended tests file.

• Add tests and docs:

  1. there are 3 unit tests mocking the underlying API and checking, initialisation, when safe content is passed and when harmful content is passed.
  2. an example notebook showing its use. It lives in docs/docs/integrations directory.

• Lint and test: Linting has passed.

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
langchain	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Dec 16, 2024 6:05am

[efriis]

we're not adding new chains like this to community anymore as we're mostly recommending folks build things like this in langgraph (see the langchain guides on migrating from legacy chains to lcel or langgraph here: https://python.langchain.com/docs/versions/migrating_chains/ )

We're more flexible when it comes to external packages like langchain-azure ones at github.com/langchain-ai/langchain-azure, so if the azure team wants them in, feel free to merge and release from there!