[ENH] Add comprehensive tests for mindist and pointwise distance modules + 5 critical bug removal

[Adityakushwaha2006]

Description

The mindist and pointwise distance modules currently lack dedicated test coverage.

mindist module - 4 files need tests:

• _sax.py
• _sfa.py
• _dft_sfa.py
• _paa_sax.py

pointwise module - 4 files need tests:

• _euclidean.py
• _manhattan.py
• _squared.py
• _minkowski.py

Why This Matters

Distance calculations are fundamental to classification, clustering, and regression in aeon. Without tests, refactoring could introduce bugs and edge cases may not be handled correctly.

Currently only test_symbolic_mindist.py exists for mindist, and pointwise has no dedicated tests.

Proposed Solution

Add comprehensive test suites covering:

• Correctness with known values
• Mathematical properties (symmetry, non-negativity, etc.)
• Edge cases (empty series, extreme values, NaN/Inf)
• Lower bounding properties for mindist functions
• Pairwise distance computations

[Adityakushwaha2006]

@aeon-actions-bot assign @Adityakushwaha2006

[Adityakushwaha2006]

I've completed implementing comprehensive test coverage for the mindist and pointwise distance modules.

What Was Added

8 new test files with 127 test functions total:
Pointwise module (78 tests):

• Euclidean distance (20 tests)
• Manhattan distance (18 tests)
• Squared distance (17 tests)
• Minkowski distance (23 tests)
  Mindist module (49 tests):
• SAX distance (14 tests)
• SFA distance (12 tests)
• DFT-SFA distance (11 tests)
• PAA-SAX distance (12 tests)

[Adityakushwaha2006]

BUGS FOUND:
During test development, the comprehensive test suite exposed five pre-existing bugs in production code:

1. List operations - Incorrect use of .shape[0] on lists instead of len()
2. Built-in functions - Python max()/min() incompatible with compiled code
3. Type inference - Implicit casting and operations causing compilation errors
4. Array dimensions - 2D arrays incorrectly passed to functions expecting 1D arrays
5. Weight handling - Pairwise function tried to index weights incorrectly, expecting per-instance weights when weights should be shared across all comparisons
   All bugs fixed across distance calculation functions. These issues were hidden because the code had no test coverage and was never executed before.

[Adityakushwaha2006]

The PR has been completed, and is ready for review : )
The abovementioned goals were achieved, all Cl tests pass and validate fully.

Unassigning myself for now, let me know if theres any changes required in the PR.

@aeon-actions-bot unassign @Adityakushwaha2006

[GiGiKoneti]

Hi! I’d like to work on adding tests and fixing bugs for the mindist / pointwise distance modules.
I plan to start by writing comprehensive tests and then fixing any uncovered bugs.
Could anyone point me to the best starting files or guidelines?

[Adityakushwaha2006]

Hey @GiGiKoneti , the same has already been completed in the PR #3186 , related to this issue.
This issue has been marked complete and is pending review.
Feel free to suggest improvements in the related PR.

[Adityakushwaha2006]

@GiGiKoneti if you're a new contributor , it'll be best to go through the available good first issues and the contributors guide to get started : )

There is need for expansive test coverage, do let me know if you'd need help in finding cases to write them for.