[DOC] Add model usage examples (v2)

[ramanbansal1]

Reference Issues/PRs

Related to: #2377

What does this implement/fix? Explain your changes.

This PR adds doctest-style usage examples to the v2 package classes for stable v2 models. The examples are minimal, CI-compliant docstrings that demonstrate the recommended high-
level package interface for the v2 API, helping users quickly adopt the new architecture.

• Added docstring usage examples for v2 package classes:

  • TFT_pkg_v2 (Temporal Fusion Transformer)
  • PatchTST_pkg_v2 (PatchTST)
  • DLinear_pkg_v2 (DLinear)
  • SCINet_pkg_v2 (SCINet)
  • SOFTS_pkg_v2 (SOFTS)
  • FreTS_pkg_v2 (FreTS)

• Each example contains:

  • Minimal synthetic dataset creation using the v2 TimeSeries API
  • Proper data module initialization (EncoderDecoderTimeSeriesDataModule or TslibDataModule)
  • Package-level model instantiation with model_cfg, trainer_cfg, and datamodule_cfg
  • Demonstration of the high-level fit/predict workflow
  • Doctest-compliant formatting with SKIP directives for Lightning-dependent operations

• Examples are CI-ready:

  • Use proper doctest prompts (>>> and ...) for doctest runner compatibility
  • Mark heavy operations (fit/predict) with  # doctest: +SKIP  to avoid Lightning dependency issues in CI
  • All lines ≤ 88 characters to comply with repository linting rules
  • Use type-safe assertions instead of numeric comparisons for robustness
  • Follow the exact v2 API structure (TimeSeries + DataModule patterns)

What should a reviewer concentrate their feedback on?

• Examples correctness:

  • Are the imports and model calls accurate for the current v2 codebase?
  • Do the examples correctly use the TimeSeries and DataModule v2 API?
  • Are the model_cfg parameters correct for each model?

• CI compliance:

  • Do the doctest examples parse correctly with  python -m doctest ?
  • Are the SKIP directives appropriately placed for Lightning-dependent operations?
  • Do line lengths comply with the 88-character limit?

• API consistency:

  • Do the examples demonstrate the recommended high-level package interface?
  • Are the data module parameters (max_encoder_length, context_length, etc.) correctly configured?
  • Are the loss functions and model parameters appropriate for each model type?

• Documentation quality:

  • Are the examples minimal yet complete enough to be copy-paste runnable?
  • Do they provide sufficient context for users to understand the v2 workflow?
  • Is the distinction between v1 and v2 API clear in the examples?

Did you add any tests for the change?

• No. These are documentation docstring examples only. The existing test suite for v2 models (test_all_estimators_v2.py ) includes doctest execution tests that will validate  these examples.

Any other comments?

• The v2 API is still experimental (users see warnings when importing TimeSeries and DataModule classes), but these examples follow the current stable v2 patterns.
• Examples use minimal synthetic data to avoid external dependencies and ensure they work in CI environments without Lightning.
• The package-level interface (model_cfg, trainer_cfg, datamodule_cfg) is the recommended way to use v2 models, as demonstrated in these examples.
• Future work could add similar examples to the corresponding model classes (_v2.py files) in addition to the package classes.

PR checklist

• [x] The PR title starts with either [ENH], [MNT], [DOC], or [BUG]. ([DOC])
• [ ] Added/modified tests
• [ ] Used pre-commit hooks when committing to ensure that code is compliant with hooks. Install hooks with pre-commit install . To run hooks independent of commit, execute p
re-commit run --all-files

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
⚠️ Please upload report for BASE (main@67c354f). Learn more about missing BASE report.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #2408   +/-   ##
=======================================
  Coverage        ?   88.20%           
=======================================
  Files           ?      200           
  Lines           ?    11170           
  Branches        ?        0           
=======================================
  Hits            ?     9852           
  Misses          ?     1318           
  Partials        ?        0           

Flag	Coverage Δ
cpu	88.20% <ø> (?)
pytest	88.20% <ø> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[phoeenniixx]

Thanks @ramanbansal1!
But if you look at the issues #2407 and #2377, we are planning to make some changes to the API, so these examples will be overridden very soon.
That is why I would suggest working on v1 examples for now