Actionable comments posted: 4

docs/sf/providers/aws/guide/python.md (1)

342-358: 📐 Maintainability & Code Quality | 🔵 Trivial | 💤 Low value

Minor: combined example block breaks from surrounding doc convention.

The preceding "Lambda Layer" section splits custom and functions config into separate YAML blocks; this new example merges both into one, which is a small stylistic inconsistency.

🤖 Prompt for 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.

In `@docs/sf/providers/aws/guide/python.md` around lines 342 - 358, The new Python
layer example in the AWS guide breaks the existing documentation style by
merging the `custom` and `functions` sections into a single YAML block. Update
the example near the `pythonRequirements` and `api` configuration so it matches
the surrounding “Lambda Layer” convention: keep the `custom` configuration and
the `functions` configuration as separate YAML snippets while preserving the
same `PydanticLambdaLayer` and `WebLambdaLayer` references.

packages/serverless/lib/plugins/python/index.js (1)

127-132: 🎯 Functional Correctness | 🔵 Trivial | ⚡ Quick win

Invalid options.layers is silently dropped without any warning.

When custom.pythonRequirements.layers is misconfigured (not an object, or an array), it's deleted with no log message. Elsewhere in this same getter, misconfiguration (dockerImage+dockerFile) throws, and docker-related misconfig triggers this.warningLogged/this.log.warning. A user who accidentally sets layers to the wrong shape will see no layers built and no explanation why.

💡 Suggested fix: warn on invalid shape

     if (
       options.layers != null &&
       (typeof options.layers !== 'object' || Array.isArray(options.layers))
     ) {
+      if (this.log) {
+        this.log.warning(
+          'custom.pythonRequirements.layers must be an object map of layer definitions; ignoring invalid value.',
+        )
+      } else {
+        this.serverless.cli.log(
+          'WARNING: custom.pythonRequirements.layers must be an object map of layer definitions; ignoring invalid value.',
+        )
+      }
       delete options.layers
     }

🤖 Prompt for 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.

In `@packages/serverless/lib/plugins/python/index.js` around lines 127 - 132, The
invalid options.layers handling in pythonRequirements should not silently delete
misconfigured values; update the getter in the Python plugin to emit a warning
before removing layers when it is not an object or is an array. Follow the
existing validation patterns used nearby in this same getter, such as the
dockerImage/dockerFile conflict and docker-related warning logic, and use
this.warningLogged or this.log.warning so users know their
custom.pythonRequirements.layers setting was ignored.

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 `@packages/serverless/lib/plugins/python/lib/layer.js`:
- Around line 126-131: The empty-requirements fast path in the layer zip flow
only checks requirementsTxtPath with fse.existsSync, so a zero-byte
requirements.txt still falls through and gets packaged. Update the branch in the
layer.js logic that builds the rootZip/writeZip path to treat an empty
requirements.txt the same as a missing one by checking file size as well as
existence, so the JSZip clean empty zip path is used whenever the generated
requirements file has no content.
- Around line 112-165: The symlink creation in zipNamedLayerRequirements is not
awaited, so failures can be swallowed and the layer zip may never be created;
update the async flow to await the file operation and let errors fail the build.
Apply the same fix in zipRequirements as well, since it has the same un-awaited
fse.symlink pattern. Use the existing zipCachePath/targetZipPath branch logic
and keep the Windows copySync path unchanged.

In `@packages/serverless/lib/plugins/python/lib/pip.js`:
- Around line 1067-1079: The empty requirements flow is inconsistent between
installRequirementsForFile and zipNamedLayerRequirements, causing an empty
requirements.txt to be packaged instead of treated as “no reqs.” Update
installRequirementsForFile to signal the empty-file case more explicitly (for
example by returning null/undefined) or update zipNamedLayerRequirements to
check the requirements file size in addition to existsSync before zipping. Make
the behavior consistent using the existing symbols generateRequirementsFile,
installRequirementsForFile, and zipNamedLayerRequirements so an empty layer
produces a clean empty python/ artifact without requirements.txt.
- Around line 1052-1065: Add an explicit guard in the Python requirements layer
flow before `path.isAbsolute(requirementsFile)` is called, since
`requirementsFile` may be missing or non-string when schema validation is
skipped. Update the logic in `pip.js` around the `absRequirementsFile`
resolution to first validate `requirementsFile` and throw a `ServerlessError`
with a clear message and the existing
`PYTHON_REQUIREMENTS_LAYER_REQUIREMENTS_FILE_INVALID` code, rather than allowing
a native `TypeError` to escape. Use the existing `requirementsFile` and
`layerName` handling in this block to keep the error consistent with the current
file-existence check.

---

Nitpick comments:
In `@docs/sf/providers/aws/guide/python.md`:
- Around line 342-358: The new Python layer example in the AWS guide breaks the
existing documentation style by merging the `custom` and `functions` sections
into a single YAML block. Update the example near the `pythonRequirements` and
`api` configuration so it matches the surrounding “Lambda Layer” convention:
keep the `custom` configuration and the `functions` configuration as separate
YAML snippets while preserving the same `PydanticLambdaLayer` and
`WebLambdaLayer` references.

In `@packages/serverless/lib/plugins/python/index.js`:
- Around line 127-132: The invalid options.layers handling in pythonRequirements
should not silently delete misconfigured values; update the getter in the Python
plugin to emit a warning before removing layers when it is not an object or is
an array. Follow the existing validation patterns used nearby in this same
getter, such as the dockerImage/dockerFile conflict and docker-related warning
logic, and use this.warningLogged or this.log.warning so users know their
custom.pythonRequirements.layers setting was ignored.

Fix all unresolved CodeRabbit comments on this PR:

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

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: 5ef6c132-10c6-430e-92b1-c80848a65d3e

Reviewing files that changed from the base of the PR and between b5ff282 and 8df2d23cbf5fce325c4fe4e11607dd005ef35993.

• docs/sf/providers/aws/guide/python.md
• packages/serverless/lib/plugins/python/index.js
• packages/serverless/lib/plugins/python/lib/layer.js
• packages/serverless/lib/plugins/python/lib/pip.js
• packages/serverless/test/unit/lib/plugins/python/layer.test.js
• packages/serverless/test/unit/lib/plugins/python/pip.test.js
• packages/sf-core/tests/python/test.js
• packages/sf-core/tests/python/tests/layers_multi/requirements/pydantic.txt
• packages/sf-core/tests/python/tests/layers_multi/requirements/ulid.txt
• packages/sf-core/tests/python/tests/layers_multi/serverless.yml