[AINode]: Integrate toto as a builtin forecasting model

[graceli02]

Description

Integrate Datadog Toto model into AINode as a built-in forecast model.

What changed

• Added Toto model registration to built-in model map (model_id=toto, repo_id=Datadog/Toto-Open-Base-1.0).
• Added Toto-specific loading path in model_loader.py:
  • Load from local built-in model directory via Toto.from_pretrained(...)
  • Wrap with TotoForecaster for inference pipeline usage.
• Added TotoPipeline for AINode forecast flow:
  • Convert IoTDB targets into Toto MaskedTimeseries.
  • Run Toto forecasting and normalize output to 2D tensors expected by AINode postprocess.
• Vendored Toto runtime source files under iotdb/ainode/core/model/toto/.
• Added required dependencies:
  • gluonts[torch]
  • rotary-embedding-torch
• Added verification assets:
  • verify_toto.py (end-to-end local smoke script)
  • tests/test_toto_pipeline.py (lightweight unit tests with fake forecaster)

Notes

• Toto integration uses a dedicated loader path instead of Transformers auto_map.
• Output shape handling in Toto pipeline was adjusted to satisfy AINode ForecastPipeline.postprocess contract (2D per batch item).

---

This PR has:

• been self-reviewed.
  • concurrent read
  • concurrent write
  • concurrent read and write
• added documentation for new or modified features or behaviors.
• added Javadocs for most classes and all non-trivial methods.
• added or updated version, license, or notice information
• added comments explaining the "why" and the intent of the code wherever would not be obvious
  for an unfamiliar reader.
• added unit tests or modified existing tests to cover new code paths, ensuring the threshold
  for code coverage.
• added integration tests.
• been tested in a test IoTDB cluster.

---

Key changed/added classes (or packages if there are too many classes) in this PR