[Biopython-dev] Ideas for Biopython 2.0

[Peter Cock]

On Wed, Jun 21, 2017 at 4:09 PM, Tiago Antão <tiagoantao at gmail.com> wrote:
> If people are stuck on Python 2, they do have a simple solution: use
> Biopython 1. That would probably be maintained at least until 2020.

I would say would likely be maintained until 2020 only - in line with the
proposal to sign up to http://www.python3statement.org/

http://mailman.open-bio.org/pipermail/biopython-dev/2017-June/021739.html

> If Biopython 2 goes ahead (a big if) then I doubt it would be launched very
> soon: in the big scheme of things we are fast approaching 2020, and I doubt
> a first release of Biopython 2 would occur before 2018.

Yes.

> With regards to the Python 3 version I would suggest supporting only the
> most recent when we start Biopython 3. In any case nothing before 3.5.

Yes. I would be very tempted to recommend Python 3.6 onwards, but
certainly at least Python 3.5 seems sensible.

> In my mind the most important discussion is the module system. All the rest
> seems easy (lots of work still) in comparison.

By this do you mean agreeing a new scheme for where to put things,
or the related but even harder challenge of how to allow sub-module
installation under the Biopython namespace using actual sub-packages
on PyPI? [*]. If anyone knows of a Python based effort like this, that
would be really helpful for the practical details.

As far as I could see, the scikit community just use a common prefix
giving scikit-learn, scikit-neuralnetwork, etc.

Peter

[*] I think this is possible, see the backports namespace as an example,
but that is fiddly and I don't fully understand it myself despite having used
it in https://github.com/peterjc/backports.lzma