Add argument for custom OpenAI endpoint in EmbedderConfig

[dutchfarao]

We have an OpenAI proxy deployed within our organization. This proxy uses the same logic as the OpenAI API, but has a different endpoint. Currently, the EmbedderConfig has no option to provide a custom endpoint for OpenAI.

Current implementation

embedder_config=EmbedderConfig(embedding_provider='openai', embedding_api_key=xxxxxxxxxxxxx,embedding_model_name='custom_model_name')

which then creates the following client

    @requires_dependencies(["openai"], extras="openai")
    def get_client(self) -> "OpenAI":
        from openai import OpenAI

        return OpenAI(api_key=self.api_key.get_secret_value())

Suggested improvement

The ability to use an OpenAI client with a custom base_url.

embedder_config=EmbedderConfig(embedding_provider='openai', embedding_api_key=xxxxxxxxxxxxx,embedding_model_name='custom_model_name',
base_url='llmproxy.organization.com')

which then creates the following client

    @requires_dependencies(["openai"], extras="openai")
    def get_client(self) -> "OpenAI":
        from openai import OpenAI

        return OpenAI(api_key=self.api_key.get_secret_value(), base_url=self.base_url)

[dutchfarao]

#406