Embedders list token as mandatory init parameter although it is optional

[julian-risch]

Several embedders list token as mandatory init parameter although it is optional

• https://docs.haystack.deepset.ai/docs/optimumtextembedder
• https://docs.haystack.deepset.ai/docs/optimumdocumentembedder
• https://docs.haystack.deepset.ai/docs/sentencetransformerstextembedder
• https://docs.haystack.deepset.ai/docs/sentencetransformersdocumentembedder

We should remove the row containing mandatory init parameters for those four embedders.

The text description could also be more clear.

The component uses a HF_API_TOKEN environment variable by default. Otherwise, you can pass a Hugging Face API token at initialization with token:

Only if the user tries to access a private model or public but gated, the component uses a HF_API_TOKEN environment variable by default or a secret explicitly provided by the user.

[anakin87]

Agree!

Some additions:

• all components based on Hugging Face should support both HF_API_TOKEN and HF_TOKEN enviroment variables
• the token is not only required for private models, but also for public but gated models (e.g models from Meta, Google, Mistral... an example: https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct)