[FIX] OWDBSCAN: handle empty X

[ajdapretnar]

Issue

Fixes #7105.

Description of changes

Handle empty X (often from plain Corpus).

Includes

• Code changes
• Tests
• Documentation

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 88.73%. Comparing base (5872eb6) to head (7d6d8d2).

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #7107      +/-   ##
==========================================
- Coverage   88.73%   88.73%   -0.01%     
==========================================
  Files         332      332              
  Lines       73450    73455       +5     
==========================================
+ Hits        65176    65177       +1     
- Misses       8274     8278       +4     

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[VesnaT]

I think self.data.X is None could never be True.
Table.X is always a np.ndarray, but sometimes empty.

[ajdapretnar]

I thought as much. Thanks for checking!

[VesnaT]

To avoid the duplicated code, I'd rather the code was places right after the

        if not self.check_data_size(data):
            data = None

block, and implemented similarly (not necessarily with an additional function, but by setting the data to None before assigning it to self.data). In this case self.data_normalized would also be properly set.

[ajdapretnar]

@VesnaT Fixed, you can check again.