[PWGDQ] fix match type assignment in ML training data producer

[aferrero2707]

The match type assigned to candidate MFT-MCH matches in the ML training data producer task was wrongly set to "fake" in cases where the MCH track is pairable but the corresponding MFT track does not have an associated MC particle. In such cases the match type is now correrctly assigned as " wrong".

[github-actions]

O2 linter results: ❌ 1 errors, ⚠️ 0 warnings, 🔕 5 disabled

[aferrero2707]

@iarsene @mcoquet642 @XiaozhiBai @mguilbau this is a bug fix for the task that produces the ML training input data for the MFT-MCH matching. I have split the PR in two commits, the first with the actual bugfix and the second with the unrelated linter fixes.

Could you please have a look? Thanks!

[alibuild]

Error while checking build/O2Physics/code-check for d412958 at 2026-08-22 04:21:

## sw/BUILD/O2Physics-code-check-latest/log
--
========== List of issues found ==========
++ echo 'Found 2 errors and 1 warnings.'
Found 2 errors and 1 warnings.
++ [[ 2 -gt 0 ]]
++ cat /sw/BUILD/d204f4ddda8ae0cc4defe2ec7574433c1c86924e/O2Physics-code-check/errors.txt
PWGDQ/Tasks/mftMchMatcher.cxx:49:1: error: included header TMath.h is not used directly [misc-include-cleaner,-warnings-as-errors]
PWGDQ/Tasks/mftMchMatcher.cxx:318:61: error: no header providing "o2::constants::math::PI" is directly included [misc-include-cleaner,-warnings-as-errors]
++ [[ 1 -gt 0 ]]
++ cat /sw/BUILD/d204f4ddda8ae0cc4defe2ec7574433c1c86924e/O2Physics-code-check/warnings.txt
PWGDQ/Tasks/mftMchMatcher.cxx:406:14: warning: redundant boolean literal in conditional return statement [readability-simplify-boolean-expr]
++ [[ 2 -gt 0 ]]
++ exit 1
--

Full log here.

[alibuild]

Error while checking build/O2Physics/o2 for d412958 at 2026-08-22 16:46:

## sw/BUILD/O2Physics-latest/log
c++: fatal error: Killed signal terminated program cc1plus
ninja: build stopped: subcommand failed.

Full log here.