Feat: `--known-enum-bases` option to support custom enum-like base cl… by imtoopunkforyou · Pull Request #3513 · wemake-services/wemake-python-styleguide

imtoopunkforyou · 2025-08-23T12:49:05Z

Checklist

I have double checked that there are no unrelated changes in this pull request (old patches, accidental config files, etc)
I have created at least one test case for the changes I have made
I have updated the documentation for the changes I have made
I have added my changes to the CHANGELOG.md

Related issues

Closes When inheriting an enum, WPS115 occurs. #3497

codecov · 2025-08-23T12:54:01Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 100.00%. Comparing base (f4c70bc) to head (f13641b).
⚠️ Report is 42 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff            @@
##            master     #3513   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files          363       363           
  Lines        12033     12071   +38     
  Branches       820       823    +3     
=========================================
+ Hits         12033     12071   +38

☔ View full report in Codecov by Sentry.
📢 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.

imtoopunkforyou · 2025-08-23T12:56:12Z

@CodiumAI-Agent /improve

QodoAI-Agent · 2025-08-23T12:57:06Z

PR Code Suggestions ✨

Latest suggestions up to f13641b

Category Suggestion Impact

Possible issue

Make normalization order-stable

Preserve the intended insertion order for normalized to avoid non-deterministic
behavior caused by converting a set to a tuple. Replace the set-comprehension with
an ordered, duplicate-filtering approach. This ensures consistent matching and
stable test results across runs and Python versions.

wemake_python_styleguide/logic/naming/enums.py [87-92]

 enum_bases = config.known_enum_bases
 if enum_bases:
-    normalized: tuple[str, ...] = tuple({
-        name for base in enum_bases for name in (base, base.split('.')[-1])
-    })
+    seen: set[str] = set()
+    normalized_list: list[str] = []
+    for base in enum_bases:
+        for name in (base, base.split('.')[-1]):
+            if name not in seen:
+                seen.add(name)
+                normalized_list.append(name)
+    normalized: tuple[str, ...] = tuple(normalized_list)
     return _has_one_of_base_classes(defn, normalized)

Suggestion importance[1-10]: 7

__

Why: Replacing a set-comprehension with an order-preserving de-duplication avoids non-determinism from set ordering and stabilizes behavior; the fix correctly reflects the intended change and fits the surrounding code.

Medium

General

Cache normalized enum bases

Avoid repeating normalization work for each class by caching the expanded
known_enum_bases on the options object. This reduces per-class overhead in large
codebases where the visitor processes many classes. Use a simple attribute cache
with a fallback computation.

wemake_python_styleguide/logic/naming/enums.py [74-94]

 def has_enum_like_base_with_config(
     defn: ast.ClassDef,
     config: ValidatedOptions,
 ) -> bool:
     """
     Tells if some class has `Enum` or semantically similar class as its base.
 
     This function also checks user-defined enum-like base classes from
     configuration.
     """
     if has_enum_like_base(defn):
         return True
 
     enum_bases = config.known_enum_bases
     if enum_bases:
-        normalized: tuple[str, ...] = tuple({
-            name for base in enum_bases for name in (base, base.split('.')[-1])
-        })
-        return _has_one_of_base_classes(defn, normalized)
+        # Cache expanded enum bases on the options object to avoid recomputation:
+        expanded = getattr(config, '_expanded_known_enum_bases', None)
+        if expanded is None:
+            seen: set[str] = set()
+            normalized_list: list[str] = []
+            for base in enum_bases:
+                for name in (base, base.split('.')[-1]):
+                    if name not in seen:
+                        seen.add(name)
+                        normalized_list.append(name)
+            expanded = tuple(normalized_list)
+            setattr(config, '_expanded_known_enum_bases', expanded)
+        return _has_one_of_base_classes(defn, expanded)
 
     return False

Suggestion importance[1-10]: 5

__

Why: Caching can reduce repeated normalization overhead and the code is consistent with the PR, but it mutates the options object with ad-hoc attributes and offers minor performance gains for typical cases.

Low

Previous suggestions

Suggestions up to commit f13641b

Category	Suggestion	Impact
Possible issue	Preserve order; avoid set dedup Preserve the configured order and avoid set-based deduplication to prevent accidental loss of intended precedence and to keep duplicate dotted/short names harmless. Use an ordered sequence and only append the short-name variant when it differs. wemake_python_styleguide/logic/naming/enums.py [87-92] enum_bases = config.known_enum_bases if enum_bases: - normalized: tuple[str, ...] = tuple({ - name for base in enum_bases for name in (base, base.split('.')[-1]) - }) + normalized_list: list[str] = [] + for base in enum_bases: + normalized_list.append(base) + short = base.split('.')[-1] + if short != base: + normalized_list.append(short) + normalized: tuple[str, ...] = tuple(normalized_list) return _has_one_of_base_classes(defn, normalized) Suggestion importance[1-10]: 6 __ Why: The existing code uses a set, which needlessly reorders items; preserving order while adding short-name variants is a sensible, low-risk improvement, though order is unlikely to affect correctness here.	Low

Suggestions up to commit f13641b

Category	Suggestion	Impact
General	Add missing enum import in test The test code references `enum.Enum` without importing `enum`, causing parsing or runtime errors in test setup. Inline an `import enum` in the snippet to ensure the AST parses consistently across environments. tests/test_visitors/test_ast/test_naming/test_class_attributes.py [48-56] custom_enum_base_class = """ +import enum + class BaseEnum(enum.Enum): def __init__(self, short: str, long: str) -> None: self.short = short self.long = long class ConcreteEnum(BaseEnum): HELLO = 'h', 'hello' WORLD = 'w', 'world' """ Suggestion importance[1-10]: 8 __ Why: The snippet uses `enum.Enum` without importing `enum`; adding the import makes the test snippet self-contained and prevents parse/runtime errors, directly improving test correctness.	Medium
General	Import enum in multi-base test This snippet also uses `enum.Enum` without importing `enum`, which will break AST parsing in tests. Include `import enum` at the top of the snippet to make tests reliable. tests/test_visitors/test_ast/test_naming/test_class_attributes.py [60-75] multiple_custom_enum_bases = """ +import enum + class FirstEnum(enum.Enum): pass class SecondEnum(enum.Enum): pass class FirstEnumSuccessor(FirstEnum): HELLO = 'hello' WORLD = 'world' class SecondEnumSuccessor(SecondEnum): PING = 'pong' class MultiInheritanceEnum(FirstEnum, SecondEnum): MULTI = 'multi' """ Suggestion importance[1-10]: 8 __ Why: This snippet also references `enum.Enum` without an import; adding `import enum` is correct and important to ensure the AST parses and tests run reliably.	Medium
Possible issue	Ensure deterministic, trimmed base matching Avoid using a set to build `normalized` because it makes matching order non-deterministic and may drop duplicates that could be significant. Also strip whitespace from configured names to prevent subtle mismatches. Build a tuple deterministically while de-duplicating conservatively. wemake_python_styleguide/logic/naming/enums.py [87-92] enum_bases = config.known_enum_bases if enum_bases: - normalized: tuple[str, ...] = tuple({ - name for base in enum_bases for name in (base, base.split('.')[-1]) - }) + names: list[str] = [] + for base in enum_bases: + b = base.strip() + if not b: + continue + for name in (b, b.split('.')[-1]): + if name not in names: + names.append(name) + normalized: tuple[str, ...] = tuple(names) return _has_one_of_base_classes(defn, normalized) Suggestion importance[1-10]: 6 __ Why: The change avoids set-induced nondeterminism and trims whitespace, which is a reasonable robustness improvement; it's accurate and aligns with the code context, but impact is moderate since order likely doesn't affect matching.	Low

Suggestions up to commit 2fdedc7

Category	Suggestion	Impact
Possible issue	Normalize configured enum base names Normalize configured base names to match how `_has_one_of_base_classes` expects fully qualified and short names. Without normalization, users passing plain names like `MyEnum` or qualified ones inconsistently may not match bases like `module.MyEnum`. Expand the configured names to include both the raw and module-qualified variants as supported. wemake_python_styleguide/logic/naming/enums.py [74-90] def has_enum_like_base_with_config( defn: ast.ClassDef, config: ValidatedOptions, ) -> bool: """ Tells if some class has `Enum` or semantically similar class as its base. This function also checks user-defined enum-like base classes from configuration. """ if has_enum_like_base(defn): return True - if config.known_enum_bases: - return _has_one_of_base_classes(defn, config.known_enum_bases) + bases = tuple(config.known_enum_bases or ()) + if bases: + # Add both short and dotted variants for robustness: + normalized: tuple[str, ...] = tuple({ + name for base in bases for name in {base, base.split('.')[-1]} + }) + return _has_one_of_base_classes(defn, normalized) return False Suggestion importance[1-10]: 6 __ Why: The suggestion is plausible and could improve matching by handling both dotted and short base names, but the current code may already work with short names and there's no evidence `_has_one_of_base_classes` requires this normalization. Impact is moderate, not critical.	Low

sobolevn · 2025-08-23T19:49:21Z

+    options,
+):
+    """Testing that custom enum base classes work with configuration."""
+    code = """


we define examples on a module level

fixed in f13641b

…asses

sobolevn

Please, also don't forget to add new configuration option to the related violations docs.

sobolevn · 2025-08-24T11:56:28Z

+        normalized: tuple[str, ...] = tuple({
+            name for base in enum_bases for name in (base, base.split('.')[-1])
+        })
+        return _has_one_of_base_classes(defn, normalized)


Suggested change

normalized: tuple[str, ...] = tuple({

name for base in enum_bases for name in (base, base.split('.')[-1])

})

return _has_one_of_base_classes(defn, normalized)

return _has_one_of_base_classes(

defn,

tuple({

name

for base in enum_bases

for name in (base, base.split('.')[-1])

}),

)

: tuple[str, ...] annotation looks suspicious :)

I'll fix it next week.

imtoopunkforyou · 2025-08-24T12:04:18Z

Please, also don't forget to add new configuration option to the related violations docs.

oh, yes, okay. I'll add it in a week

sobolevn

I want to make a release, with this feature in it :)
You can address issues in the follow-up PR, since they are minor.

imtoopunkforyou · 2025-08-25T10:18:09Z

I want to make a release, with this feature in it :) You can address issues in the follow-up PR, since they are minor.

I'll create a new PR with a link to this one as soon as I make the changes.

sobolevn reviewed Aug 23, 2025

View reviewed changes

imtoopunkforyou added 3 commits August 24, 2025 12:51

Feat: --known-enum-bases option to support custom enum-like base cl…

5b1f908

…asses

fix: Normalize configured enum base names

97f6f3e

fix: move test code examples to module level and improve test structure

f13641b

sobolevn reviewed Aug 24, 2025

View reviewed changes

sobolevn approved these changes Aug 25, 2025

View reviewed changes

sobolevn merged commit c5c229c into wemake-services:master Aug 25, 2025
9 checks passed

imtoopunkforyou mentioned this pull request Aug 27, 2025

docs: add configuration examples for known-enum-bases option #3518

Closed

4 tasks

Uh oh!

Conversation

imtoopunkforyou commented Aug 23, 2025

Checklist

Related issues

Uh oh!

codecov Bot commented Aug 23, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

imtoopunkforyou commented Aug 23, 2025

Uh oh!

QodoAI-Agent commented Aug 23, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

PR Code Suggestions ✨

Previous suggestions

Uh oh!

sobolevn Aug 23, 2025

Choose a reason for hiding this comment

Uh oh!

imtoopunkforyou Aug 24, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

sobolevn left a comment

Choose a reason for hiding this comment

Uh oh!

sobolevn Aug 24, 2025

Choose a reason for hiding this comment

Uh oh!

sobolevn Aug 24, 2025

Choose a reason for hiding this comment

Uh oh!

imtoopunkforyou Aug 24, 2025

Choose a reason for hiding this comment

Uh oh!

imtoopunkforyou commented Aug 24, 2025

Uh oh!

sobolevn left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

imtoopunkforyou commented Aug 25, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

codecov Bot commented Aug 23, 2025 •

edited

Loading

QodoAI-Agent commented Aug 23, 2025 •

edited

Loading

imtoopunkforyou Aug 24, 2025 •

edited

Loading