Import analyser fix

[aseembits93]

User description

just match class name when method is not imported
for example
from code_to_optimize.topological_sort import Graph # topologicalSort not imported

---

PR Type

Bug fix, Tests

---

Description

• Match class-only imports to methods

• Avoid missing class.method targets

• Update E2E to assert test count

---

Diagram Walkthrough

flowchart LR
  A["ImportFrom analyzer"] -- "fallback: match class to class.method" --> B["Target function detection"]
  B -- "found_qualified_name set" --> C["Unit test discovery"]
  C -- "validated by" --> D["End-to-end test updated (expected_unit_tests=1)"]

Loading

File Walkthrough

	Relevant files
Bug fix	discover_unit_tests.py Fallback: match class import to class.method                          codeflash/discovery/discover_unit_tests.py Add class-to-method fallback matching. Set found flags when class name matches. Handle dotted target functions via split. Add TODO about cross-file class name collisions. +11/-0
Tests	end_to_end_test_topological_sort_worktree.py E2E: assert single discovered unit test                                    tests/scripts/end_to_end_test_topological_sort_worktree.py Assert expected_unit_tests equals 1. Align E2E with new discovery behavior. +1/-0

---

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.

---

Codeflash Bot seems not to be a GitHub user. You need a GitHub account to be able to sign the CLA. If you have already a GitHub account, please add the email address used for this commit to your account.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[github-actions]

PR Reviewer Guide 🔍

Here are some key observations to aid the review process:

⏱️ Estimated effort to review: 2 🔵🔵⚪⚪⚪

🧪 PR contains tests

🔒 No security concerns identified

⚡ Recommended focus areas for review Overmatching Risk Class-name matching may incorrectly associate methods from different modules/classes sharing the same class name, potentially causing false positives in test discovery. # Check if any target function is a method of the imported class/module # e.g., importing Graph and looking for Graph.topologicalSort # TODO will pick up all tests which have the same class name (could be coming from a different file) for target_func in fnames: if "." in target_func: class_name, method_name = target_func.split(".", 1) if aname == class_name: self.found_any_target_function = True self.found_qualified_name = target_func return Ambiguous Split Splitting on the first dot assumes a single-level class.method; deeper qualifications (e.g., pkg.Class.method or Outer.Inner.method) may not be handled correctly. if "." in target_func: class_name, method_name = target_func.split(".", 1) if aname == class_name: Performance Consideration Iterating over all function names for each ImportFrom without early filtering could be inefficient on large fnames; consider pre-indexing by class name. for target_func in fnames: if "." in target_func: class_name, method_name = target_func.split(".", 1) if aname == class_name: self.found_any_target_function = True self.found_qualified_name = target_func return

[github-actions]

PR Code Suggestions ✨

Explore these optional code suggestions:

Category	Suggestion	Impact
Possible issue	Validate dotted name structure Validate the split to avoid unexpected behavior for malformed names (e.g., leading/trailing dots) and protect against empty class or method segments. This prevents incorrect matches and potential crashes if later logic assumes non-empty parts. codeflash/discovery/discover_unit_tests.py [303-305] if "." in target_func: - class_name, method_name = target_func.split(".", 1) + parts = target_func.split(".", 1) + if len(parts) == 2 and parts[0] and parts[1]: + class_name, method_name = parts + else: + continue Suggestion importance[1-10]: 6 __ Why: Adding defensive checks against malformed dotted names is reasonable and low-risk, improving robustness without changing behavior for valid inputs. Impact is moderate as current code likely handles normal cases but could be safer with validation.	Low
General	Reduce class-name collision risk Guard against false positives by ensuring aname refers to a class import (not a module alias) and matching only when the import kind aligns. Also restrict matches to ast.ImportFrom with node.module resolving to the expected module if available to avoid cross-file class name collisions. codeflash/discovery/discover_unit_tests.py [302-308] for target_func in fnames: if "." in target_func: class_name, method_name = target_func.split(".", 1) - if aname == class_name: + # Only match when the imported alias equals the class name and this is a 'from X import Y' form + if aname == class_name and isinstance(node, ast.ImportFrom) and node.module: self.found_any_target_function = True self.found_qualified_name = target_func return Suggestion importance[1-10]: 4 __ Why: The concern about class-name collisions is valid, but the added check isinstance(node, ast.ImportFrom) is redundant inside visit_ImportFrom, and requiring node.module may incorrectly skip valid cases (e.g., relative imports). The improvement is minor and partially misapplies context.	Low

[codeflash-ai]

⚡️ Codeflash found optimizations for this PR

📄 38% (0.38x) speedup for ImportAnalyzer.visit_Attribute in codeflash/discovery/discover_unit_tests.py

⏱️ Runtime : 114 microseconds → 82.6 microseconds (best of 250 runs)

A dependent PR with the suggested changes has been created. Please review:

• ⚡️ Speed up method ImportAnalyzer.visit_Attribute by 38% in PR #868 (import-analyser-fix) #869

If you approve, it will be merged into this PR (branch import-analyser-fix).

[codeflash-ai]

This PR is now faster! 🚀 @misrasaurabh1 accepted my optimizations from:

• ⚡️ Speed up method ImportAnalyzer.visit_Attribute by 38% in PR #868 (import-analyser-fix) #869