petl 2.0 roadmap and maintenance

103 views
Skip to first unread message

Alistair Miles

unread,
Aug 7, 2019, 10:24:49 AM8/7/19
to pytho...@googlegroups.com
Hi all,

Following on from the 1.3.0 release, here are a couple of thoughts about what to do next. 

My own priority is to maintain compatibility with upstream packages like intervaltree and openpyxl, and to reduce maintenance effort. In this vein there are three work items I think would be worth putting in scope for a 2.0 release:

(1) Drop support for Python 2.7. It is near end of life, and many packages in the pydata ecosystem have dropped support already. This would imply removing all Python 2 specific code, which would simplify the code base and reduce maintenance effort. https://github.com/petl-developers/petl/issues/478 

(2) Migrate from nosetests to pytest. Nosetests is no longer maintained, which is not necessarily a major problem, however using pytest would allow some simplification of testing code, and so probably worth doing. https://github.com/petl-developers/petl/issues/453

(3) Change the default value of the new `petl.config.failonerror` option from False to True. This would mean that transformation functions would by default raise any exceptions that occur during value transormations, rather than silently swallowing errors and replacing values as is the current behaviour. Several people have argued this should change, but it is a substantial change so it's waited for a major version bump. https://github.com/petl-developers/petl/issues/365

If anyone has any suggestions or objections, please comment on the issues linked above.

If there are any other changes or new features you would like to include in the roadmap, please do say. However please bear in mind that the current maintainers don't have bandwidth to add new features, and so you would have to lead the development work yourself (which you would be welcome to do).

Best wishes,
Alistair 
--

Alistair Miles
Head of Epidemiological Informatics
Centre for Genomics and Global Health
Big Data Institute
Li Ka Shing Centre for Health Information and Discovery
University of Oxford
Old Road Campus
Headington
Oxford
OX3 7LF
United Kingdom

Please feel free to resend your email and/or contact me by other means if you need an urgent reply.

Reply all
Reply to author
Forward
0 new messages