[inference provider] Add regolo.ai as an inference provider

[Andreh7777]

What’s in this PR

This PR integrates RegoloAI as a new Inference Provider for the @huggingface/inference library.

RegoloAI is a green, full-stack AI platform providing optimized APIs for state-of-the-art models. This integration allows users to leverage RegoloAI's infrastructure directly through the Hugging Face JS client.

Changes
Added Regoloai provider implementation in packages/inference.

Supported tasks:

• Conversational (Chat Completion)
• Text Generation
• Text-to-Image (with support for b64_json and url outputs)
• Feature Extraction (Embeddings)
• Updated InferenceClient to support RegoloAI routing and model mapping.
• Added comprehensive test suite in InferenceClient.spec.ts.

Technical Details
Documentation followed: The integration strictly adheres to the Inference Providers' official guide.

Testing: All integration tests for RegoloAI tasks passed successfully using vitest (7/7 tests passed).

~/Desktop/huggingface.js/packages/inference$ npx vitest -t "Regolo AI"

DEV v0.34.6 /home/user/Desktop/huggingface.js/packages/inference

 ↓ HF Inference (50) [skipped]
 ↓ Fal AI (5) [skipped]
 ↓ Featherless (3) [skipped]
 ❯ Regolo AI (7)
   ✓ chatCompletion 425ms
   ✓ chatCompletion stream 316ms
   ✓ chatCompletion multimodal (Vision) 8330ms
   ✓ chatCompletion OCR 741ms
   ✓ textToImage 12536ms
   ✓ featureExtraction 2022ms
   ✓ textGeneration 630ms
 ↓ Replicate (12) [skipped]
 ↓ SambaNova (3) [skipped]
 ↓ Together (4) [skipped]
 ↓ Nebius (5) [skipped]
 ↓ Scaleway (5) [skipped]
 ↓ 3rd party providers (1) [skipped]
 ↓ Fireworks (2) [skipped]
 ↓ Hyperbolic (4) [skipped]
 ↓ Novita (2) [skipped]
 ↓ Black Forest Labs (2) [skipped]
 ↓ Cohere (2) [skipped]
 ↓ Cerebras (2) [skipped]
 ↓ Nscale (3) [skipped]
 ↓ Groq (2) [skipped]
 ↓ ZAI (3) [skipped]
 ↓ OVHcloud (4) [skipped]
 ↓ Wavespeed AI (8) [skipped]
 ↓ PublicAI (2) [skipped]
 ↓ Baseten (2) [skipped]
 ↓ clarifai (2) [skipped]

Test Files 1 passed | 1 skipped (2)
Tests 7 passed | 146 skipped (153)
Start at 14:18:36
Duration 26.60s (transform 1.08s, setup 68ms, collect 1.41s, tests 25.07s, environment 1ms, prepare 201ms)

I look forward to your feedback and am happy to make any necessary adjustments!