This sounds reasonable to me. Thank you. Nam, does it make sense to you?
On Fri, May 8, 2020 at 11:53 AM Robert Bradshaw wrote:
> I'd really like to not see this work go to waste, both the original
> revision, the further efforts Nam has done in making it more manageable to
> review, and the
That's good news! Thanks for sharing.
Another datapoint, here are a few of Beam's dependencies that no longer
release new py2 artifacts (I looked at REQUIRED_PACKAGES + aws, gcp, and
interactive extras):
hdfs
numpy
pyarrow
ipython
There are more if we include transitive dependencies and
Thanks for the heads up Thomas. Please let us know as soon as possible what
you find.
On Fri, May 8, 2020 at 2:43 PM Thomas Weise wrote:
> I could not find a way for the hidden pip install to succeed within the
> virtual environment.
>
> Also, heads up that we found another issue related to
Hi Everyone,
We use FileIO's writeDynamic to write dynamically to separate groups based
on an attribute's value in the input PCollection.
I wanted to check if there is a way to make sharding as runner dependent?
Many thanks,
Amit
I'd really like to not see this work go to waste, both the original
revision, the further efforts Nam has done in making it more manageable to
review, and the work put into reviewing this so far, so we can get the
benefits of being on Hugo. How about this for a concrete proposal:
(1) We get
I could not find a way for the hidden pip install to succeed within the
virtual environment.
Also, heads up that we found another issue related to timer encoding that
looks like a release blocker. Details to follow.
On Fri, May 8, 2020 at 9:53 AM Udi Meiri wrote:
> +Chad Dombrova , who added
Here's a script that we could run on the old and new sites that should
quickly catch any major issues but not get caught up in formatting minutia.
On Fri, May 8, 2020 at 10:23 AM Robert Bradshaw wrote:
> On Fri, May 8, 2020 at 9:58 AM Aizhamal Nurmamat kyzy
> wrote:
>
>> I understand the
On Fri, May 8, 2020 at 9:58 AM Aizhamal Nurmamat kyzy
wrote:
> I understand the difficulty, and this certainly comes with lessons learned
> for future similar projects.
>
> To your questions Robert:
> (1 and 2) I will commit to review the text in the resulting pages. I will
> try and use some
Also an update from +Nam Bui :
"This commit [1] is up-to-date. So I walked through all of markdown files.
Apart from Syntax changes between Jekyll & Hugo, if there were any
differences in contents regarding to removed/added/modified, I would have
double check the text with the current website. I
The window information should still be there. Beam propagates windows
through PCollection, and I don't think WriteFiles does anything explicit to
stop that.
Can you try this with the direct runner to see what happens there?
What is your windowing on this PCollection?
Reuven
On Fri, May 8, 2020
I understand the difficulty, and this certainly comes with lessons learned
for future similar projects.
To your questions Robert:
(1 and 2) I will commit to review the text in the resulting pages. I will
try and use some automation to extract visible text from each page and diff
it with the
+Chad Dombrova , who added _find_protoc_gen_mypy.
I'm guessing that the code
in _install_grpcio_tools_and_generate_proto_files creates a kind of
virtualenv, but it only works well for staging Python modules and not
binaries like protoc-gen-mypy.
(I assume there's a reason why it doesn't invoke
I'm -0 on merging as-is. I have the same concerns as Robert and he's voiced
them very well so I won't waste time re-airing them.
(2) I spot checked the content, pulled out some common patterns, and
> it mostly looks good, but there were also some issues (e.g. several
> pages were replaced with
It hasn't been 3 months yet, but I wanted to call out a milestone that
Python 3 downloads crossed the 50% threshold on pypi, if just briefly.
On Thu, Feb 13, 2020 at 12:40 AM Ismaël Mejía wrote:
>
> > I would suggest re-evaluating this within the next 3 months again. We need
> > to balance
Start with the contribution guide[1] and pick up some of the starter
tasks[2] to learn parts of the codebase.
1: https://beam.apache.org/contribute/
2: https://s.apache.org/beam-starter-tasks
On Thu, May 7, 2020 at 11:58 PM deepak kumar wrote:
> Hi There,
> Can anyone suggest the best way to
I got the same behavior using Spark Runner (with Spark 2.4.3), window
information was missing.
Just to clarify, the combiner after TextIO had different results. In Flink
runner the files names were dropped, and in Spark the combination process
happened twice, duplicating data. I think it is
Hi There,
Can anyone suggest the best way to get started on Beam APIs , so i can
start contributing to the codebase?
Thanks
Deepak
This is a tough situation.
It would have been much better if this transition was structured in
such a way that the review was more manageable (e.g. the suggestion of
scripts, not mixing in voluminous unnecessary changes like whitespace,
and not updating content), and possibly even incrementally
18 matches
Mail list logo