Fix kwarg_help() crash on current pandas (trailing-comma .loc indexing)

[heykav]

Bug

mpf.kwarg_help() raises ValueError: zip() argument 2 is longer than argument 1 on current pandas (tested with pandas 3.0.6, but the same failure occurs on any pandas 2.x that enforces strict=True zip in _LocIndexer._multi_take). This is a 100% reproduction on the existing tests/test_kwarg_help.py::test_kwarg_help test.

Root cause

df_wrapcols() in src/mplfinance/_kwarg_help.py selects each row with:

row = df.loc[ix,]

The trailing comma turns ix, into the 1-tuple (ix,). Older pandas treated a 1-tuple key to .loc the same as the bare scalar. Current pandas routes any tuple key through _getitem_tuple -> _multi_take, which does:

d = {axis: self._get_listlike_indexer(key, axis)
     for (key, axis) in zip(tup, self.obj._AXIS_ORDERS, strict=True)}

self.obj._AXIS_ORDERS for a DataFrame has length 2 (index, columns), but tup here has length 1, so zip(..., strict=True) raises ValueError. Minimal repro outside mplfinance:

import pandas as pd
df = pd.DataFrame({'a': [1, 2], 'b': [3, 4]})
df.loc[0,]   # ValueError: zip() argument 2 is longer than argument 1

Fix

Drop the stray trailing comma so the key is a plain scalar, which is what was actually intended (df.loc[ix]), and is a strict subset of the existing behavior — no other code path is touched.

Verification

• tests/test_kwarg_help.py::test_kwarg_help: fails on master (ValueError: zip() argument 2 is longer than argument 1), passes after this change.
• Red/green confirmed by reverting the one-line change and re-running: fails without the fix, passes with it.
• Full tests/ suite: 43 failed / 8 passed before -> 42 failed / 9 passed after, with no new failures introduced. The remaining failures are pre-existing and unrelated (mostly image-comparison baseline drift from newer matplotlib rendering, and an ATR/.loc indexing bug in the P&F/renko code path that is already tracked in open PR Add traditional box scaling for P&F charts and fix ATR indexing bug #699).

Environment used for verification: fresh clone, Python 3.12 venv, pip install -e ., pandas 3.0.6 / numpy 2.5.3 / matplotlib 3.11.2 (current latest as of testing).

🤖 Generated with Claude Code