Normalize decoder targets when using EncoderNormalizer

[HarshRajSinghania]

Summary
Normalize decoder targets when EncoderDecoderTimeSeriesDataModule uses a per-sequence target normalizer (EncoderNormalizer).

Motivation
Fixes #2360. The dataset fitted EncoderNormalizer on the encoder window and applied it only to target_past. y stayed on the raw scale, so the training target did not match the normalized encoder history.

Implementation

• Add ScalerAdapter.transform_sequence, which reuses the state just fitted by fit_transform_sequence and does not re-fit.
• In the dataset __getitem__, fit on the encoder window and transform both target_past and the decoder target with that state. The forecast horizon is not used to fit the scale.
• Non-per-sequence columns are left unchanged, matching the existing sequence helper.

Testing

• python -m pytest tests/test_data/test_data_module.py -q -o addopts= — 43 passed, including new test_encoder_normalizer_scales_decoder_target.
• ruff check on the changed adapter, data module, and test file passed.

The regression test checks that y and target_past match standard scaling fitted on the encoder window only (mean and unbiased std, plus float epsilon).

[codecov]

Codecov Report

❌ Patch coverage is 88.88889% with 2 lines in your changes missing coverage. Please review.
⚠️ Please upload report for BASE (main@e437602). Learn more about missing BASE report.

Files with missing lines	Patch %	Lines
pytorch_forecasting/adapters/scaler_adapters.py	85.71%	2 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #2457   +/-   ##
=======================================
  Coverage        ?   88.82%           
=======================================
  Files           ?      218           
  Lines           ?    11687           
  Branches        ?        0           
=======================================
  Hits            ?    10381           
  Misses          ?     1306           
  Partials        ?        0           

Flag	Coverage Δ
cpu	88.82% <88.88%> (?)
pytest	88.82% <88.88%> (?)

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.