[FIX] Revert the lsp optimization logic

[mohammedahmed18]

PR Type

Bug fix, Enhancement

---

Description

• Revert LSP optimization initialization flow

• Simplify patch creation API usage

• Improve error handling on patch failure

• Return task identifier in response

---

Diagram Walkthrough

flowchart LR
  InitOld["Server-held init result"] -- reuse --> OptimRun["Run optimization"]
  OptimRun -- best found --> PatchCreate["Create patch from worktree"]
  PatchCreate -- success --> RespondOK["LSP success response"]
  PatchCreate -- failure --> RespondErr["LSP error response"]

Loading

File Walkthrough

	Relevant files
Enhancement	perform_optimization.py Revert to server-provided init and simplify patch flow      codeflash/lsp/features/perform_optimization.py Reuse server.current_optimization_init_result and existing optimizer. Remove module prep and optimizer creation logic. Replace metadata-based patch creation with simplified API. Add failure check, return task_id, adjust response fields. +17/-56

---

[github-actions]

PR Reviewer Guide 🔍

Here are some key observations to aid the review process:

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

🧪 No relevant tests

🔒 No security concerns identified

⚡ Recommended focus areas for review Possible None access The code assumes server.current_optimization_init_result and server.optimizer.current_function_optimizer are present and valid without checks; if either is None or not initialized, this can raise exceptions during optimization start. def sync_perform_optimization(server: CodeflashLanguageServer, params) -> dict[str, str]: # noqa: ANN001 server.show_message_log(f"Starting optimization for function: {params.functionName}", "Info") should_run_experiment, code_context, original_helper_code = server.current_optimization_init_result function_optimizer = server.optimizer.current_function_optimizer current_function = function_optimizer.function_to_optimize Patch creation check create_diff_patch_from_worktree return value is only checked for truthiness; clarify and handle different failure modes (e.g., exceptions vs. empty path) to avoid silent failures and ensure consistent error messaging. patch_path = create_diff_patch_from_worktree( server.optimizer.current_worktree, relative_file_paths, function_to_optimize_qualified_name ) if not patch_path: return { "functionName": params.functionName, "status": "error", "message": "Failed to create a patch for optimization", } Missing patch metadata Response no longer includes patch id or original file path metadata, which may be relied upon by clients; confirm API contract and update consumers or include minimal identifiers. server.show_message_log(f"Optimization completed for {params.functionName} with {speedup:.2f}x speedup", "Info") return { "functionName": params.functionName, "status": "success", "message": "Optimization completed successfully", "extra": f"Speedup: {speedup:.2f}x faster", "patch_file": str(patch_path), "task_id": params.task_id, "explanation": best_optimization.explanation_v2, }

[github-actions]

PR Code Suggestions ✨

Explore these optional code suggestions:

Category	Suggestion	Impact
Possible issue	Add initialization guards Guard against missing initialization state to prevent attribute errors. If the current optimizer or init result is absent, return a clear error instead of dereferencing None. codeflash/lsp/features/perform_optimization.py [12-14] +if not getattr(server, "current_optimization_init_result", None): + return {"functionName": params.functionName, "status": "error", "message": "Optimization not initialized"} should_run_experiment, code_context, original_helper_code = server.current_optimization_init_result -function_optimizer = server.optimizer.current_function_optimizer + +function_optimizer = getattr(server.optimizer, "current_function_optimizer", None) +if not function_optimizer: + return {"functionName": params.functionName, "status": "error", "message": "No function optimizer found"} current_function = function_optimizer.function_to_optimize Suggestion importance[1-10]: 7 __ Why: Guarding server.current_optimization_init_result and server.optimizer.current_function_optimizer prevents attribute errors; it's contextually accurate given the new code directly dereferences them. Moderate impact as it's defensive error handling.	Medium
General	Pre-validate patch prerequisites Validate relative_file_paths before creating the patch to avoid generating empty or invalid diffs. Also ensure server.optimizer.current_worktree is present to prevent failures. codeflash/lsp/features/perform_optimization.py [92-101] +if not relative_file_paths: + return { + "functionName": params.functionName, + "status": "error", + "message": "No files changed during optimization", + } + +if not getattr(server.optimizer, "current_worktree", None): + return { + "functionName": params.functionName, + "status": "error", + "message": "Worktree not available for patch creation", + } + patch_path = create_diff_patch_from_worktree( server.optimizer.current_worktree, relative_file_paths, function_to_optimize_qualified_name ) if not patch_path: return { "functionName": params.functionName, "status": "error", "message": "Failed to create a patch for optimization", } Suggestion importance[1-10]: 6 __ Why: Validating relative_file_paths and presence of current_worktree reduces failure modes before calling create_diff_patch_from_worktree. It's reasonable and consistent with nearby code, offering robustness but not fixing a critical bug.	Low
	Handle missing optional fields Protect against division by zero or invalid baseline runtimes when computing speedup. If the baseline is zero or non-positive, return an explicit error to avoid ZeroDivisionError and misleading results. codeflash/lsp/features/perform_optimization.py [103-113] server.show_message_log(f"Optimization completed for {params.functionName} with {speedup:.2f}x speedup", "Info") return { "functionName": params.functionName, "status": "success", "message": "Optimization completed successfully", "extra": f"Speedup: {speedup:.2f}x faster", "patch_file": str(patch_path), - "task_id": params.task_id, + "task_id": getattr(params, "task_id", None), "explanation": best_optimization.explanation_v2, } Suggestion importance[1-10]: 3 __ Why: The change only makes task_id access defensive with getattr, which is a minor improvement; it does not address the earlier-mentioned speedup validation and offers limited impact on correctness.	Low