docs: add missing docstrings, Raises sections, and fix Args in networ…

[chhayankjain]

Description

Fix several docstring issues across network blocks in monai/networks/blocks/:

• localnet_block.py: add class and __init__ docstrings with Args and Raises sections to ResidualBlock and
  LocalNetResidualBlock; fix broken indentation and incorrect parameter name (kernel_initializer → initializer) in
  LocalNetFeatureExtractorBlock
• selfattention.py: add Raises section to SABlock.__init__
• crossattention.py: add Raises section to CrossAttentionBlock.__init__
• patchembedding.py: add Raises section to PatchEmbeddingBlock.__init__
• upsample.py: add Raises section to UpSample.__init__
• downsample.py: add Raises sections to DownSample.__init__ and SubpixelDownsample.__init__

No functional changes.

Types of changes

• Non-breaking change (fix or new feature that would not break existing functionality).

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 1f410674-eb85-423a-8f89-cbbf606df12a

📥 Commits

Reviewing files that changed from the base of the PR and between a1bf6f1 and a66d670.

📒 Files selected for processing (6)

• monai/networks/blocks/crossattention.py
• monai/networks/blocks/downsample.py
• monai/networks/blocks/localnet_block.py
• monai/networks/blocks/patchembedding.py
• monai/networks/blocks/selfattention.py
• monai/networks/blocks/upsample.py

✅ Files skipped from review due to trivial changes (5)

• monai/networks/blocks/patchembedding.py
• monai/networks/blocks/selfattention.py
• monai/networks/blocks/crossattention.py
• monai/networks/blocks/downsample.py
• monai/networks/blocks/upsample.py

---

📝 Walkthrough

Walkthrough

This PR updates docstrings across multiple MONAI network block classes, adding explicit Raises: sections that document ValueError conditions for constructor parameters (crossattention, selfattention, patchembedding, downsample, upsample, localnet_block). Additionally, LocalNetFeatureExtractorBlock.__init__ renames a parameter from kernel_initializer to initializer (default "kaiming_uniform"). No runtime logic changes beyond the parameter rename.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	Title is truncated ('networ…') and partially relates to changes—mentions docstrings, Raises sections, and Args fixes across network blocks.
Description check	✅ Passed	Description covers all modified files with specific changes and notes no functional impact. Matches template with Types of changes section.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 2

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@monai/networks/blocks/downsample.py`:
- Around line 133-134: Update the docstring ValueError clause for the downsample
block to precisely match the implemented check: state that the error is raised
only when mode is "maxpool" or "avgpool", pre_conv == "default", out_channels !=
in_channels, and in_channels is not provided; locate the exception description
in the downsample module's docstring (the section documenting raised exceptions
for parameters mode/pre_conv/in_channels/out_channels) and change the sentence
to include the additional condition out_channels != in_channels so the docs
mirror the actual code path.

In `@monai/networks/blocks/localnet_block.py`:
- Around line 302-303: The docstring for the initializer parameter in
LocalNetBlock (parameter name: initializer) claims the default is
"kaiming_uniform" but the implementation calls nn.init.kaiming_normal_; make
them consistent by either changing the docstring default to "kaiming_normal"
(and document accepted values accordingly) or changing the implementation to use
nn.init.kaiming_uniform_ when initializer=="kaiming_uniform"; update the
initializer handling logic in the LocalNetBlock __init__ (or the function that
performs weight initialization) and the docstring so the parameter name, allowed
string values, and applied nn.init.* call match exactly.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 07796a91-596c-4fbc-97ad-52cf3abecb3d

📥 Commits

Reviewing files that changed from the base of the PR and between 586dea1 and a1bf6f1.

📒 Files selected for processing (6)

• monai/networks/blocks/crossattention.py
• monai/networks/blocks/downsample.py
• monai/networks/blocks/localnet_block.py
• monai/networks/blocks/patchembedding.py
• monai/networks/blocks/selfattention.py
• monai/networks/blocks/upsample.py