Add optional pipeline noise reduction module with Gaussian and CLAHE methods

[Copilot]

This introduces an optional noise-reduction step for noisy raw/processed images so users can improve segmentation quality by inserting denoising at any pipeline point. The new step is pipeline-compatible and configurable for both simple smoothing and contrast-adaptive enhancement.

• New pipeline module: pyopia.noise

  • Added ReduceNoise as a pipeline-compatible class.
  • Supports:
    • method='gaussian' for blur-based denoising
    • method='clahe' for adaptive local contrast equalization
  • Configurable input/output keys:
    • image_source selects which image in Pipeline.data to process.
    • output_key allows writing to a chosen key.
    • If output_key is not provided, output now defaults to im_denoised.

• Method-specific configuration

  • Gaussian: gaussian_sigma
  • CLAHE: clahe_clip_limit, clahe_nbins
  • Invalid method names raise a clear ValueError.

• Targeted tests added

  • New test module pyopia/tests/test_noise.py covering:
    • Gaussian smoothing behavior
    • CLAHE output constraints
    • Invalid method handling
    • End-to-end use through Pipeline.run_step(...)
    • Default output behavior (im_denoised)

• Documentation

  • Added a new example under Advanced analysis: docs/notebooks/noise_reduction.md
  • Linked in docs table of contents (docs/_toc.yml)

[steps.noisereduction]
pipeline_class = "pyopia.noise.ReduceNoise"
method = "gaussian"
image_source = "imraw"
output_key = "im_denoised"
gaussian_sigma = 1.0

[nepstad]

@copilot Update default value for output_key to "im_denoised". Add new documentation example under "Advanced analysis".

[Copilot]

@copilot Update default value for output_key to "im_denoised". Add new documentation example under "Advanced analysis".

Implemented in bf94e9b: ReduceNoise now defaults to writing output to im_denoised when output_key is not provided, tests were updated for the new default behavior, and I added a new Advanced analysis documentation page at docs/notebooks/noise_reduction.md (linked from docs/_toc.yml).

[ralph-sj]

Looks good to me.