Recaping here:

We all agree that SDF is the way to go for future implementations of sources. It enables us to get rid of the source interfaces. However, SDF does not solve the lack of streaming sources in Python.

The expansion PR (thanks btw!) solves the problem of expanding/translating URNs known to an ExpansionService. That is a more programmatic way of replacing language-specific performs, instead of relying on translators directly in the Runner.

What is unsolved is the configuration of sources from a foreign environment. In my opinion this is the most pressing issue for Python sources, because what is PubSubIO worth in Python if you cannot configure it?

What about this:

I think it is worth adding a JSON configuration option for all existing Java sources. That way, we could easily configure them as part of the expansion request (which would contain a JSON configuration). I'll probably fork a thread to discuss this in more detail, but would like to hear your thoughts.

-Max

On 01.02.19 13:08, Robert Bradshaw wrote:
On Thu, Jan 31, 2019 at 6:25 PM Maximilian Michels <m...@apache.org <mailto:m...@apache.org>> wrote:

    Ah, I thought you meant native Flink transforms.

    Exactly! The translation code is already there. The main challenge
    is how to
    programmatically configure the BeamIO from Python. I suppose that is
    also an
    unsolved problem for cross-language transforms in general.


This is what https://github.com/apache/beam/pull/7316 does.

For a particular source, one would want to define a URN and corresponding payload, then (probably) a CompositeTransform in Python that takes the users arguments, packages them into the payload, applies the ExternalTransform, and returns the results. How to handle arbitrary UDFs embedded in sources is still TBD.

    For Matthias' pipeline with PubSubIO we can build something
    specific, but for
    the general case there should be way to initialize a Beam IO via a
    configuration
    map provided by an external environment.


I thought quite a bit about how we could represent expansions statically (e.g. have some kind of expansion template that could be used, at least in many cases, as data without firing up a separate process. May be worth doing eventually, but we run into the same issues that were discussed at https://github.com/apache/beam/pull/7316#discussion_r249996455 ).

If one is already using a portable runner like Flink, having the job service process automatically also serve up an expansion service for various URNs it knows and cares about is probably a pretty low bar. Flink could serve up things it would rather get back untouched in a transform with a special flink runner urn.

As Ahmet mentions, SDF is better solution. I hope it's not that far away, but even once it comes we'll likely want the above framework to invoke the full suite of Java IOs even after they're running on SDF themselves.

- Robert

    On 31.01.19 17:36, Thomas Weise wrote:
     > Exactly, that's what I had in mind.
     >
     > A Flink runner native transform would make the existing unbounded
    sources
     > available, similar to:
     >
     >
    
https://github.com/apache/beam/blob/2e89c1e4d35e7b5f95a622259d23d921c3d6ad1f/runners/flink/src/main/java/org/apache/beam/runners/flink/FlinkStreamingTransformTranslators.java#L167
     >
     >
     >
     >
     > On Thu, Jan 31, 2019 at 8:18 AM Maximilian Michels
    <m...@apache.org <mailto:m...@apache.org>
     > <mailto:m...@apache.org <mailto:m...@apache.org>>> wrote:
     >
     >     Wouldn't it be even more useful for the transition period if
    we enabled Beam IO
     >     to be used via Flink (like in the legacy Flink Runner)? In
    this particular
     >     example, Matthias wants to use PubSubIO, which is not even
    available as a
     >     native
     >     Flink transform.
     >
     >     On 31.01.19 16:21, Thomas Weise wrote:
     >      > Until SDF is supported, we could also add Flink runner
    native transforms for
     >      > selected unbounded sources [1].
     >      >
     >      > That might be a reasonable option to unblock users that
    want to try Python
     >      > streaming on Flink.
     >      >
     >      > Thomas
     >      >
     >      > [1]
     >      >
     >
    
https://github.com/lyft/beam/blob/release-2.10.0-lyft/runners/flink/src/main/java/org/apache/beam/runners/flink/LyftFlinkStreamingPortableTranslations.java
     >      >
     >      >
     >      > On Thu, Jan 31, 2019 at 6:51 AM Maximilian Michels
    <m...@apache.org <mailto:m...@apache.org>
     >     <mailto:m...@apache.org <mailto:m...@apache.org>>
     >      > <mailto:m...@apache.org <mailto:m...@apache.org>
    <mailto:m...@apache.org <mailto:m...@apache.org>>>> wrote:
     >      >
     >      >      > I have a hard time to imagine how can we map in a
    generic way
     >      >     RestrictionTrackers into the existing
    Bounded/UnboundedSource, so I would
     >      >     love to hear more about the details.
     >      >
     >      >     Isn't it the other way around? The SDF is a
    generalization of
     >     UnboundedSource.
     >      >     So we would wrap UnboundedSource using SDF. I'm not
    saying it is
     >     trivial, but
     >      >     SDF offers all the functionality that UnboundedSource
    needs.
     >      >
     >      >     For example, the @GetInitialRestriction method would
    call split on the
     >      >     UnboundedSource and the restriction trackers would
    then be used to
     >     process the
     >      >     splits.
     >      >
     >      >     On 31.01.19 15:16, Ismaël Mejía wrote:
     >      >      >> Not necessarily. This would be one way. Another
    way is build an SDF
     >      >     wrapper for UnboundedSource. Probably the easier path
    for migration.
     >      >      >
     >      >      > That would be fantastic, I have heard about such
    wrapper multiple
     >      >      > times but so far there is not any realistic
    proposal. I have a hard
     >      >      > time to imagine how can we map in a generic way
    RestrictionTrackers
     >      >      > into the existing Bounded/UnboundedSource, so I
    would love to hear
     >      >      > more about the details.
     >      >      >
     >      >      > On Thu, Jan 31, 2019 at 3:07 PM Maximilian Michels
    <m...@apache.org <mailto:m...@apache.org>
     >     <mailto:m...@apache.org <mailto:m...@apache.org>>
     >      >     <mailto:m...@apache.org <mailto:m...@apache.org>
    <mailto:m...@apache.org <mailto:m...@apache.org>>>> wrote:
     >      >      >>
     >      >      >>   > In addition to have support in the runners,
    this will require a
     >      >      >>   > rewrite of PubsubIO to use the new SDF API.
     >      >      >>
     >      >      >> Not necessarily. This would be one way. Another
    way is build an SDF
     >      >     wrapper for
     >      >      >> UnboundedSource. Probably the easier path for
    migration.
     >      >      >>
     >      >      >> On 31.01.19 14:03, Ismaël Mejía wrote:
     >      >      >>>> Fortunately, there is already a pending PR for
    cross-language
     >      >     pipelines which
     >      >      >>>> will allow us to use Java IO like PubSub in
    Python jobs.
     >      >      >>>
     >      >      >>> In addition to have support in the runners, this
    will require a
     >      >      >>> rewrite of PubsubIO to use the new SDF API.
     >      >      >>>
     >      >      >>> On Thu, Jan 31, 2019 at 12:23 PM Maximilian Michels
     >     <m...@apache.org <mailto:m...@apache.org>
    <mailto:m...@apache.org <mailto:m...@apache.org>>
     >      >     <mailto:m...@apache.org <mailto:m...@apache.org>
    <mailto:m...@apache.org <mailto:m...@apache.org>>>> wrote:
     >      >      >>>>
     >      >      >>>> Hi Matthias,
     >      >      >>>>
     >      >      >>>> This is already reflected in the compatibility
    matrix, if you look
     >      >     under SDF.
     >      >      >>>> There is no UnboundedSource interface for
    portable pipelines.
     >     That's a
     >      >     legacy
     >      >      >>>> abstraction that will be replaced with SDF.
     >      >      >>>>
     >      >      >>>> Fortunately, there is already a pending PR for
    cross-language
     >      >     pipelines which
     >      >      >>>> will allow us to use Java IO like PubSub in
    Python jobs.
     >      >      >>>>
     >      >      >>>> Thanks,
     >      >      >>>> Max
     >      >      >>>>
     >      >      >>>> On 31.01.19 12:06, Matthias Baetens wrote:
     >      >      >>>>> Hey Ankur,
     >      >      >>>>>
     >      >      >>>>> Thanks for the swift reply. Should I change
    this in the
     >     capability matrix
     >      >      >>>>>
    <https://s.apache.org/apache-beam-portability-support-table> then?
     >      >      >>>>>
     >      >      >>>>> Many thanks.
     >      >      >>>>> Best,
     >      >      >>>>> Matthias
     >      >      >>>>>
     >      >      >>>>> On Thu, 31 Jan 2019 at 09:31, Ankur Goenka
    <goe...@google.com <mailto:goe...@google.com>
     >     <mailto:goe...@google.com <mailto:goe...@google.com>>
     >      >     <mailto:goe...@google.com <mailto:goe...@google.com>
    <mailto:goe...@google.com <mailto:goe...@google.com>>>
     >      >      >>>>> <mailto:goe...@google.com
    <mailto:goe...@google.com> <mailto:goe...@google.com
    <mailto:goe...@google.com>>
     >     <mailto:goe...@google.com <mailto:goe...@google.com>
    <mailto:goe...@google.com <mailto:goe...@google.com>>>>> wrote:
     >      >      >>>>>
     >      >      >>>>>       Hi Matthias,
     >      >      >>>>>
     >      >      >>>>>       Unfortunately, unbounded reads including
    pubsub are not yet
     >      >     supported for
     >      >      >>>>>       portable runners.
     >      >      >>>>>
     >      >      >>>>>       Thanks,
     >      >      >>>>>       Ankur
     >      >      >>>>>
     >      >      >>>>>       On Thu, Jan 31, 2019 at 2:44 PM Matthias
    Baetens
     >      >     <baetensmatth...@gmail.com
    <mailto:baetensmatth...@gmail.com> <mailto:baetensmatth...@gmail.com
    <mailto:baetensmatth...@gmail.com>>
     >     <mailto:baetensmatth...@gmail.com
    <mailto:baetensmatth...@gmail.com> <mailto:baetensmatth...@gmail.com
    <mailto:baetensmatth...@gmail.com>>>
     >      >      >>>>>       <mailto:baetensmatth...@gmail.com
    <mailto:baetensmatth...@gmail.com>
     >     <mailto:baetensmatth...@gmail.com
    <mailto:baetensmatth...@gmail.com>>
     >      >     <mailto:baetensmatth...@gmail.com
    <mailto:baetensmatth...@gmail.com>
     >     <mailto:baetensmatth...@gmail.com
    <mailto:baetensmatth...@gmail.com>>>>> wrote:
     >      >      >>>>>
     >      >      >>>>>           Hi everyone,
     >      >      >>>>>
     >      >      >>>>>           Last few days I have been trying to
    run a streaming
     >      >     pipeline (code on
     >      >      >>>>>           Github
    <https://github.com/matthiasa4/beam-demo>) on a
     >      >     Flink Runner.
     >      >      >>>>>
     >      >      >>>>>           I am running a Flink cluster locally
    (v1.5.6
>      >      >>>>>  <https://flink.apache.org/downloads.html>)
     >      >      >>>>>           I have built the SDK Harness
    Container: /./gradlew
     >      >      >>>>>           :beam-sdks-python-container:docker/
     >      >      >>>>>           and started the JobServer: /./gradlew
>      >      >>>>>  :beam-runners-flink_2.11-job-server:runShadow
     >      >      >>>>>           -PflinkMasterUrl=localhost:8081./
     >      >      >>>>>
     >      >      >>>>>           I run my pipeline with: /env/bin/python
     >     streaming_pipeline.py
     >      >      >>>>>           --runner=PortableRunner
    --job_endpoint=localhost:8099
     >      >     --output xxx
     >      >      >>>>>           --input_subscription xxx
    --output_subscription xxx/
     >      >      >>>>>           /
     >      >      >>>>>           /
     >      >      >>>>>           All this is running inside a Ubuntu
    (Bionic) in a
     >     Virtualbox.
     >      >      >>>>>
     >      >      >>>>>           The job submits fine, but
    unfortunately fails after
     >     a few
     >      >     seconds with
     >      >      >>>>>           the error attached.
     >      >      >>>>>
     >      >      >>>>>           Anything I am missing or doing wrong?
     >      >      >>>>>
     >      >      >>>>>           Many thanks.
     >      >      >>>>>           Best,
     >      >      >>>>>           Matthias
     >      >      >>>>>
     >      >      >>>>>
     >      >
     >

Reply via email to