feat: add adapter trimming option to remove soft-clipped regions from signal

[jayhesselberth]

Summary

Add a new "Trim Adapters" checkbox in the Plot Options panel that removes soft-clipped (adapter) regions from signal data before plotting. This is useful for focusing on the aligned portion of reads without adapter noise.

Changes

Python Backend:

• Add trim_adapters parameter to plot_read(), plot_reads(), plot_aggregate()
• Add _apply_adapter_trimming_to_reads() helper function in squiggy/plotting.py that:
  • Trims signal based on soft-clip base positions from move table
  • Trims sequence and quality scores
  • Adjusts reference_start/reference_end coordinates
  • Adjusts query_to_ref mapping for new indices
  • Ensures move table base count matches sequence length
• Add soft-clip fields (query_start_offset, query_end_offset) to AlignedRead dataclass

TypeScript Frontend:

• Add trimAdapters to PlotOptionsState interface
• Add checkbox UI in Plot Options panel (Aggregate section)
• Wire parameter through SquiggyRuntimeAPI.generateAggregatePlot()

Tests:

• Add tests/test_adapter_trimming.py with 13 comprehensive tests covering:
  • No soft-clipping (unchanged)
  • Soft-clip at start/end/both ends
  • Quality scores trimming
  • query_to_ref adjustment
  • Move table base count consistency
  • Edge cases

Important Note

Adapter trimming removes the actual adapter sequences that are soft-clipped in the BAM alignment. The poly-A tail that may be visible in tRNA plots is part of the reference sequence, not adapter - it will remain visible after trimming because it's part of the aligned region.

Test plan

• All 735 Python tests pass
• All 564 TypeScript tests pass
• Build compiles successfully
• Manual testing in Positron with POD5/BAM data

🤖 Generated with Claude Code