fix(es,ca): guard to_currency() cents-part for integer inputs

[fbindakheel]

Bug

to_currency() crashes with IndexError for Spanish (es) and Catalan (ca) when the input is an integer that resolves to a whole currency unit (zero cents). For example:

num2words(100, lang='es', to='currency')  # → IndexError: list index out of range
num2words(0,   lang='es', to='currency')  # → IndexError: list index out of range

This affects all whole-unit integer inputs — 0, 100 (1 euro), 200, 2100, etc.

Root cause

Both lang_ES.py and lang_CA.py override to_currency() by calling the base implementation, then splitting the result on the separator (con / amb) to get [dollars_part, cents_part]. They then unconditionally apply grammar fixes to list_result[1].

However, the base to_currency() omits the cents segment entirely when the input has no cents (i.e. integer input where right == 0 and has_decimal is False). This leaves list_result with only one element, so list_result[1] raises IndexError.

The existing tests only used float inputs like 1.00, 21.00 — floats always produce a cents segment (con cero céntimos), so the integer path was never exercised.

Fix

Guard the cents-part post-processing with if len(list_result) > 1: in both files. The dollars-part handling is unaffected.

Tests

Added regression tests in tests/test_es.py and tests/test_ca.py covering integer whole-unit amounts and feminine currencies. Verified the new tests fail on the original code and pass with the fix. All 1495 tests pass.

num2words(0,    lang='es', to='currency')  # → 'cero euros'
num2words(100,  lang='es', to='currency')  # → 'un euro'
num2words(2100, lang='es', to='currency')  # → 'veintiún euros'
num2words(100,  lang='es', to='currency', currency='GBP')  # → 'una libra'