feat: allow negative FOV overlap to introduce spacing between tiles

[mpiersonsmela]

Summary

Allows the FOV Overlap spinbox in both Flexible and Wellplate multipoint acquisition modes to accept negative values, which introduce spacing between adjacent FOVs (useful for sparse sampling).

Motivation

This is useful for sparse imaging over a large well, in cases where cells may not be evenly distributed (so looking at the center of the well may not be representative).

Today the spinbox is bounded setRange(0, 99). The downstream math already does the right thing for negatives:

step_size_mm = fov_size_mm * (1 - overlap_percent / 100)

A negative overlap (e.g. -50%) produces a step ~1.5× the FOV — tiles end up spaced apart rather than overlapping. The only thing blocking the feature was the input bound.

Changes

• software/control/widgets.py — FlexibleMultiPointWidget: setRange(0, 99) → setRange(-1000, 99)
• software/control/widgets.py — WellplateMultiPointWidget: setRange(0, 99) → setRange(-1000, 99)
• software/control/microscope_control_server.py — MCP server's overlap_percent Pydantic field: ge=0, le=50 → ge=-1000, le=99 so YAML / API callers can also pass negative overlap. (Note: the previous le=50 was already inconsistent with the UI bound of 99; this aligns them.)

A lower bound of -1000% gives a step size of up to ~11× the FOV — effectively unlimited for any realistic plate, while keeping the spinbox UI well-behaved.

Safety

• Out-of-bounds FOVs generated by extreme spacing are already filtered downstream by ScanCoordinates.validate_coordinates() against SOFTWARE_POS_LIMIT, so no additional bounds checking is needed.
• control/core/geometry_utils.py already guards step_size <= 0 for the pathological overlap >= 100% case.
• No changes to scan-coordinate or geometry logic.

Test plan

• Spinbox accepts -50 in both wellplate and flexible mode; live FOV preview shows tiles spaced ~1.5× FOV apart
• Very negative values (e.g. -500) don't crash near stage limits — out-of-range FOVs are silently dropped
• YAML load with overlap_percent: -25 passes Pydantic validation
• Positive overlap (e.g. 10%) behaves unchanged — regression check

🤖 Generated with Claude Code