Fix Kendrick base consistency between MSPeak and MolecularFormula

[robertyoung3]

Summary

Two related fixes for Kendrick base handling:

1. Inconsistent default in MolecularFormulaBase: When a molecular formula has no parent peak, __init__ used a hardcoded {"C": 1, "H": 2} fallback instead of reading from MSParameters.ms_peak.kendrick_base. This meant user-configured Kendrick bases were ignored for orphan formulas.

2. Stale formula KMD after base change: MSPeak.change_kendrick_base() recalculated KMD for the peak but not for its associated molecular formulas. When change_kendrick_base_all_mspeaks() is called after formula assignment (as in priorityAssignment.py and findOxygenPeaks.py), formulas retained stale KMD values.

Changes

MolecularFormulaFactory.py

# Before: hardcoded fallback
kendrick_dict_base = {"C": 1, "H": 2}

# After: reads from MSParameters
from corems.encapsulation.factory.parameters import MSParameters
kendrick_dict_base = MSParameters.ms_peak.kendrick_base

MSPeakClasses.py — added to change_kendrick_base():

# Also update all associated molecular formulas
for mf in self.molecular_formulas:
    mf._kmd, mf._kendrick_mass, mf._nominal_km = mf._calc_kmd(kendrick_dict_base)

Test plan

• Verified change_kendrick_base_all_mspeaks() is the canonical entry point (called in priorityAssignment.py:273 and findOxygenPeaks.py:92)
• That method iterates all peaks calling mspeak.change_kendrick_base(), which now also updates formulas
• Existing test in tests/test_mspeak.py:32 exercises change_kendrick_base