refactor: simplify autograd padding indices

[yaugenst-flex]

Greptile Overview

Updated On: 2025-11-06 10:08:29 UTC

Greptile Summary

Refactored _get_pad_indices function to delegate padding logic to NumPy's built-in pad function instead of manually implementing each mode (edge, reflect, symmetric, wrap). This simplification reduces code complexity from ~15 lines of mode-specific logic to a single delegation call with proper error handling.

Key changes:

• Removed manual index computation for edge, reflect, symmetric, and wrap modes
• Added delegation to onp.pad(onp.arange(n), pad_width, mode=mode) to generate indices
• Preserved constant mode behavior (unchanged)
• Improved error handling with try/catch block for unsupported modes
• Maintained compatibility by converting result back to appropriate numpy module (autograd.numpy when needed)

The refactoring maintains identical behavior while improving maintainability and leveraging well-tested NumPy functionality.

Confidence Score: 5/5

• This PR is safe to merge - it's a clean refactoring that simplifies code without changing behavior
• The refactoring delegates to NumPy's well-tested pad function instead of reimplementing padding logic, reducing code complexity while maintaining identical behavior. Comprehensive test coverage exists for all padding modes, and the change properly handles error cases.
• No files require special attention

Important Files Changed

File Analysis

Filename	Score	Overview
tidy3d/plugins/autograd/functions.py	5/5	Simplified _get_pad_indices by delegating to NumPy's pad function instead of manual mode implementations - improves maintainability without changing behavior

Sequence Diagram

sequenceDiagram
    participant Caller
    participant pad
    participant _get_pad_indices
    participant numpy_pad as onp.pad
    
    Caller->>pad: array, pad_width, mode
    alt mode == "constant"
        pad->>pad: Use np.pad directly
    else other modes
        pad->>_get_pad_indices: n, pad_width, mode, numpy_module
        alt mode == "constant"
            _get_pad_indices->>_get_pad_indices: Return arange indices
        else other modes (edge, reflect, symmetric, wrap)
            _get_pad_indices->>numpy_pad: arange(n), pad_width, mode
            numpy_pad-->>_get_pad_indices: padded indices
            _get_pad_indices->>_get_pad_indices: Convert to numpy_module
        end
        _get_pad_indices-->>pad: indices
        pad->>pad: Index array using indices
    end
    pad-->>Caller: padded array

Loading

[greptile-apps]

_{1 file reviewed, no comments}

_{Edit Code Review Agent Settings | Greptile}

[github-actions]

Diff Coverage

Diff: origin/develop...HEAD, staged and unstaged changes

• tidy3d/plugins/autograd/functions.py (85.7%): Missing lines 71

Summary

• Total: 7 lines
• Missing: 1 line
• Coverage: 85%

tidy3d/plugins/autograd/functions.py

Lines 67-75

  67         return numpy_module.zeros(total_pad, dtype=int)
  68 
  69     pad_left, pad_right = pad_width
  70     if mode == "constant":
! 71         return numpy_module.arange(-pad_left, n + pad_right)
  72 
  73     try:
  74         indices = onp.pad(onp.arange(n), (pad_left, pad_right), mode=mode)
  75     except ValueError as error:

---