HuggingFace Provider: Text Generation

[Kamilbenkirane]

Sub-issue of #148

Scope

Add text generation via HuggingFace's OpenAI-compatible chat completions endpoint.

• Endpoint: POST https://router.huggingface.co/v1/chat/completions
• Protocol: Reuses existing ChatCompletionsClient (same pattern as Groq/DeepSeek)
• Auth: Authorization: Bearer hf_XXX via HF_TOKEN
• Streaming: SSE

Pre-requisite

Rename HUGGINGFACE_TOKEN to HF_TOKEN in src/celeste/credentials.py and .env.example to match HuggingFace's official env var.

Files

Create: src/celeste/providers/huggingface/{__init__.py, chat/}, src/celeste/modalities/text/providers/huggingface/
Modify: src/celeste/providers/__init__.py, src/celeste/modalities/text/providers/__init__.py, src/celeste/modalities/text/models.py, src/celeste/credentials.py, .env.example