fix: handle test series names with multiple dashes in fix_series

[xickomesquita]

Summary

• fix_series() in run_test.py assumed series names had exactly one dash (e.g. TestData-7), extracting the number with split('-')[1]
• When a name contained multiple dashes the split produced unexpected segments, raising ValueError and breaking the test runner series counter

Fix

Extract the numeric part from the last dash-separated segment instead of the second one, with a try/except fallback to 0 on any parse error:

parts = max_test_data_series[0].split('-')
if len(parts) > 1 and parts[-1].isdigit():
    max_test_data_series = int(parts[-1])
else:
    max_test_data_series = 0

Test plan

• Run barista test suite with series names containing a single dash — behavior unchanged
• Run with series names containing multiple dashes — no longer raises ValueError
• Run with malformed series name — safely falls back to 0