Clint Valentine

One mockup for a new class: ```python @dataclass(frozen=True) class Template(Iterable): r1: AlignedSegment | None = None r2: AlignedSegment | None = None r1_auxiliaries: list[AlignedSegment] = field(default_factory=list) r2_auxiliaries: list[AlignedSegment] = field(default_factory=list)...

@msto here's a second draft with some your feedback incorporated. Some of the method names are a lot longer but are also now much less implicit: ```python @dataclass(frozen=True) class Template(Iterable):...

@msto I've been confused by this too. Although the SAM spec has a "null value" for these attributes (`*`), pysam additionally allows setting them as `None`. I haven't yet found...

^ Writing a `array('B', [9])` to a pysam alignment that is then written to the SAM/BAM file will roundtrip back as a `None`. I've accidentally blended the concepts in my...

Any obvious way to unit test this?

Hey @coderabbitai how would you write a unit test for this PR that would fail with the previous implementation but succeed with the new implementation?

What if folks start using [`os.devnull`](https://docs.python.org/3/library/os.html#os.devnull) as the path for when no path is requested?

> I did run into one new question - the help mentions storing the sample and molecular barcodes in SAM tags. Is that just a helpful guide for downstream usage,...

Most of this was solved in this recently merged PR (and release [v2.5.0](https://github.com/fulcrumgenomics/fgbio/releases/tag/2.5.0) now exists): - https://github.com/fulcrumgenomics/fgbio/pull/1026 I still have a lingering issue with how hard-clipping is handled: - https://github.com/fulcrumgenomics/fgbio/issues/1030...

Same issue here :(