Add optional wrap_fractional to interpolate_path for unfolded k-paths

[eyjafjallac]

Summary

• Adds a new optional argument wrap_fractional: bool = True to qten.bands.interpolate_path.
• Preserves current behavior by default (wrap_fractional=True), where rebased waypoints are wrapped into the canonical fractional cell.
• Enables unfolded path construction with wrap_fractional=False, so rebased waypoint coordinates are kept as-is (no modulo-cell wrapping), which helps when plotting paths that should follow the original label sequence without collapsing equivalent points.

Why

When interpolating a path in a target reciprocal lattice different from the source KPointSet lattice, canonical fractional wrapping can cause distinct labeled waypoints to map onto the same folded point. This makes path ticks/labels appear merged in bandstructure plots.
This PR makes that wrapping behavior configurable without breaking existing callers.

API Change

• interpolate_path(recip, waypoints, points, n_points=100, wrap_fractional=True) -> BzPath

Behavior:

• wrap_fractional=True (default): uses rebased_wp.fractional() (existing behavior)
• wrap_fractional=False: uses raw rebased_wp coordinates

Tests

• Existing interpolation/path tests remain passing.
• Added test:
  • test_interpolate_path_can_preserve_unwrapped_rebased_waypoints
  • Verifies that with wrap_fractional=False, an endpoint like (3/2, -1/4) is preserved instead of wrapped to (1/2, 3/4).

Validation

• uv run pytest "tests/test_bands.py" -k "interpolate_path" passed
• uv run pytest "tests/test_plots.py" -k "bandstructure and bz_path" passed

Example Usage

path = Q.interpolate_path(
    blocked_diamond.dual,
    ["Gamma", "L", "U", "X", "W", "L", "K", "Gamma", "U", "W", "Gamma", "X"],
    points=kpts,
    n_points=500,
    wrap_fractional=False,
)

[Alphaharrius]

LGTM!