BUG: Resample multi-series indicator inputs consistently

[Jaco-Ren]

Summary

• resample additional time-indexed indicator inputs alongside the primary series
• apply the existing OHLCV aggregation defaults to each input and align their complete periods
• add an ATR-style regression test that exercises the real Strategy.data array path

Motivation and Context

Fixes #1243.

resample_apply() currently resamples only its primary series. Additional inputs passed through *args remain at the source frequency, so a call such as:

resample_apply('D', atr, self.data.High, self.data.Low, self.data.Close, 14)

passes daily High values together with minute-level Low and Close values. Multi-series indicators therefore receive incompatible lengths and indexes and may fail before producing a result.

The change recognizes pandas Series/DataFrames and Strategy data arrays only when they share the primary series' original index. Those inputs are resampled with their field-specific defaults (High=max, Low=min, Close=last, etc.); scalar parameters and unrelated arguments remain unchanged. The existing label='right' behavior is preserved so a higher-timeframe value becomes available only after that period is complete, avoiding look-ahead bias.

Testing

• confirmed the new regression test fails before the implementation because Low and Close remain raw _Array values
• python -m unittest backtesting.test._test.TestLib.test_resample_apply backtesting.test._test.TestLib.test_resample_apply_multiple_series
• python -m backtesting.test - 82 passed, 1 skipped
• git diff --check

Types of changes

• Bug fix
• Tests
• Documentation