I've got some work progress, although right now I'm stuck on handling the for comprehensions in Python 3 type inference (see https://github.com/holdenk/beam/tree/try-py3-by-hand )
Personally I'll probably take a look into Python on other runners after this rather than Dataflow. On Sat, Nov 4, 2017 at 11:17 PM, Ahmet Altay <[email protected]> wrote: > For reference https://issues.apache.org/jira/browse/BEAM-1251 is the > umbrella issue tracking python3 support in the core SDK. There needs to be > additional runner specific work (e.g. DataflowRunner needs to use python3 > binary on its workers) once the core work is completed. > > Ahmet > > On Thu, Nov 2, 2017 at 12:58 PM, Jesse Anderson <[email protected] > > > wrote: > > > Holden is being modest in her contributions to Python frameworks, > > especially Apache Spark. > > > > On Thu, Nov 2, 2017 at 12:55 PM Holden Karau <[email protected]> > wrote: > > > > > Hi! So this is something I'm currently working on (e.g. in between > > checking > > > my e-mails :p). If you want to help join in we can split up the work > into > > > smaller components and parallelize the process a bit :) Always happy to > > see > > > more folks who care about Python 3 support. > > > > > > On Thu, Nov 2, 2017 at 12:44 PM, Lukasz Cwik <[email protected] > > > > > wrote: > > > > > > > Contributions are always welcome to improve progress. > > > > > > > > You can always vote/watch the Python 3 JIRA issue as this helps > people > > > know > > > > what others are looking for. > > > > > > > > On Thu, Nov 2, 2017 at 10:33 AM, Yue Yang <[email protected]> > wrote: > > > > > > > > > Hello, > > > > > I wonder what is the schedule to support python 3. It seems that > > the > > > > > progess is very slow. > > > > > Thanks. > > > > > > > > > > > > > > > > > > > > > -- > > > Twitter: https://twitter.com/holdenkarau > > > > > -- > > Thanks, > > > > Jesse > > > -- Twitter: https://twitter.com/holdenkarau
