FIX: support RangeIndex inputs in auto-formatting

[biru-codeastromer]

Summary

This fixes an internal auto-format failure for valid integer-indexed / RangeIndex time series.

Closes #390.

Problem

load_data_source accepts integer-indexed pandas data, but the auto-format path assumes every index exposes datetime-like frequency metadata. With RangeIndex, that raised an internal AttributeError on index.freq, logged a warning, and fell back to the unformatted path.

Changes

• guard frequency inspection so it only runs for datetime / period indexes
• preserve the existing Integer frequency metadata for non-datetime indexes
• add a regression test covering the integer-indexed pandas load path with auto-format enabled

Verification

• .venv/bin/ruff check src/sktime_mcp/runtime/executor.py tests/test_data_sources.py
• .venv/bin/ruff format --check src/sktime_mcp/runtime/executor.py tests/test_data_sources.py
• .venv/bin/pytest tests/test_data_sources.py -q
• .venv/bin/pytest -q