Add SleepWakeClassification task for DREAMT

[diegofariasc]

Contributor: Diego Farias Castro (diegof4@illinois.edu)
Type of contribution: task
Link to original paper: https://proceedings.mlr.press/v248/wang24a.html

High-level description:
This PR adds a new standalone task, SleepWakeClassification, on top of the existing DREAMTDataset. The task supports epoch-level sleep-vs-wake prediction from multimodal wrist-worn wearable signals. It turns each DREAMT record into fixed-length epochs, extracts features from accelerometer, temperature, blood volume pulse, and electrodermal activity signals, adds temporal context features, and assigns a binary sleep/wake label to each epoch.

Implementation summary:

• Adds SleepWakeClassification in pyhealth/tasks/sleep_wake_classification.py
• Exports the task in pyhealth/tasks/__init__.py
• Adds task API documentation in docs/api/tasks/pyhealth.tasks.sleep_wake_classification.rst
• Registers the task page in docs/api/tasks.rst
• Includes an example ablation script in examples/dreamt_sleep_wake_classification_lightgbm.py
• Includes task tests in tests/core/test_sleep_wake_classification.py

Reproducibility scope:
This PR focuses on the task side of the paper. It makes the sleep-wake prediction setting available inside PyHealth so the generated samples can be used in new experiments and ablation studies.

Task behavior:

• Input: wearable records from the existing DREAMTDataset
• Output: epoch-level samples with patient_id, record_id, epoch_index,
  features, and binary label
• Labels: wake maps to 1; sleep stages (REM, N1, N2, N3) map to 0
• Features: accelerometer summaries, temperature summaries, BVP-based HRV
  features, and EDA-based SCR features
• Temporal context: Gaussian smoothing, temporal derivative, and rolling
  variance for each base feature

File guide:

• pyhealth/tasks/sleep_wake_classification.py: task implementation
• pyhealth/tasks/__init__.py: public task export
• docs/api/tasks/pyhealth.tasks.sleep_wake_classification.rst: task docs
• docs/api/tasks.rst: task index update
• examples/dreamt_sleep_wake_classification_lightgbm.py: example and ablation workflow
• tests/core/test_sleep_wake_classification.py: task unit tests