Pomoxis

Added
- Read mean Qscore and alignment flag added to `stats_from_bam` output.
- Add option to `filter_bam` to keep unmapped reads that pass (non-alignment) filters.
- Enable multithreaded bam file reading and writing in `filter_bam`.
Changed
- Pin numpy<2.0.

Added
- `mini_align` can take a bam as input and optionally retain all or subset of bam tags.
Changed
- Drop porechop dependency from pypi package due to blocking package uploads. Add warning to the user if porechop is called by `mini_assemble` but does not exist.

Changed
- `subsample_bam` and `coverage_from_bam` now have unified read filtering options and logic.
Added
- `filter_bam` to filter a bam with the same logic used in `subsample_bam`.
Fixed
- `subsample_bam` was previously subsampling proportionally before filtering resulting in lower than expected depth.

Changed
- `subsample_bam`: `--force_low_coverage` saves contigs with coverage below the target
- `subsample_bam`: `--force_non_primary` saves multimapping for the subsampled reads
- `coverage_from_bam`: `--primary_only` considers only primary reads when computing the depth
- `bedtools`: upgraded to v2.31
- `porechop`: switched to using Artic version
Added
- Option `-C` for `mini_align` to copy fastx comments into bam tags
Fixed
- Minor compatibility fixes to support `pandas>=2.0`

Changed
- `subsample_bam`: `--quality` filtering now uses mean error probability, not mean of quality scores as previously.
- `subsample_bam`: enable filtering for proportional subsampling.

Fixed
- Fix crashes in `subsample_bam` with alignment filtering and `common_errors_from_bam`
- `assess_assembly -H` uses correct output directory.
- Handling of comments in bed files.
Changed
- Added `Q(sub)` to summary output.
- Ported bed file handling from `intervaltrees` to [`ncls`](https://github.com/biocore-ntnu/ncls), speeding up assessment and multithreading efficiency.
Added