Nezar Abdennur

> As of v0.19, [convolve](https://docs.scipy.org/doc/scipy/reference/generated/scipy.signal.convolve.html#scipy.signal.convolve) automatically chooses this method or the direct method based on an estimation of which is faster. https://docs.scipy.org/doc/scipy/reference/generated/scipy.signal.fftconvolve.html Maybe replace anyway?

What if the API were based on dataframes, instead of abstract alignment records?

Question: which name do we like better, `compress_runs` or `merge_runs`? It basically very similar to `merge`, but instead of combining all intervals which transitively overlap (thus, sharing a unique cluster...

Now I'm wondering if these couldn't just be special cases of `cluster` and `merge`.

I migrated the strand-awareness logic to a new function `bioframe.shift`, basically following the proposal [here](https://github.com/open2c/bioframe/issues/144#issuecomment-2037230552). Unlike `expand`, the semantics of `shift` always treats negative values as upstream and positive as...

Why `col_dtypes` as opposed to `dtypes`?

Instead of manually adding all UCSC table schemas, we should consider an approach to build the schema "database" that automatically sources table schemas from UCSC and translates them into something...

Btw, there's this which got sandboxed. It's slow, as you might expect: https://github.com/open2c/bioframe/blob/fbd129c1444cef7c34edce067027ab5f65172fe8/bioframe/sandbox/gtf_io.py

Agreed that there's plenty of room for improvement. I would advocate for now keeping it as a function downstream from `read_table`, to be applied on an unparsed dataframe column or...

That should be it. I think `chromsizes` is always a Series[str, int], being indexed by position. The docstring seems to be wrong.