[GitHub] incubator-beam pull request #940: [BEAM-] Making Dataflow Py...

[katsiapis <gi...@git.apache.org>]

GitHub user katsiapis opened a pull request:

    https://github.com/apache/incubator-beam/pull/940

    [BEAM-<Jira issue 625>] Making Dataflow Python Materialized PCollection representation more efficient (3 of several).

    Be sure to do all of the following to help us incorporate your contribution
    quickly and easily:
    
    - Refactoring up common functionality from TextFileSink to FileSink.
    
    - Removing no-longer needed overrides from TextFileSink
      (since their implementation is identical to FileSink).
    
    - Introducing NativeFileSink and NativeFileSinkWriter that capture the common
      patterns of file-related native sinks. The _NativeFileSink will now automatically
      append a compression-appropriate suffix (if any) based on the compression type (and
      similarly for the non-native FileSink).
    
    - NativeTextFileSink now inherits from NativeFileSink.
    
    - TextFileWriter now inherits from NativeFileWriter.
    
    - Introducing NativeFileSource and NativeFileSourceWriter that capture the common
      patterns of file-related native sources.
    
    - TextFileSource now inherits from NativeFileSource.
    
    - TextFileReader now inherits from NativeFileSourceReader.
    
    - Introducing CompressionTypes.AUTO, in analogy to Dataflow Java's TextIO.CompressionType.
      This is now the new default for NativeFileSource, NativeFileSink, FileSink (and all
      their subclasses).
    
    - Moving compression_type and mime_type handling to ChannelFactory.open().
    
    - Changing compression level from 9 to Z_DEFAULT_COMPRESSION as that better matches
      disk/cpu tradeoffs.
      This was showcased in the comments of this CL.
    
    - Removing CompressionTypes.DEFLATE since it's not a standard on-disk format (and not
      recognized by the SDK).
    
    - Augmenting the functionality of _CompressedFile and GcsBufferedWriter (supporting more
      standard file operations on them, needed by various Sources/Sinks).
    
    ---
    
    efficient (3 of several).

You can merge this pull request into a Git repository by running:

    $ git pull https://github.com/katsiapis/incubator-beam python

Alternatively you can review and apply these changes as the patch at:

    https://github.com/apache/incubator-beam/pull/940.patch

To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:

    This closes #940
    
----
commit 0ed53b32e3bd56a6cd9fccd17bbe941fe70da177
Author: Gus Katsiapis <ka...@katsiapis-linux.mtv.corp.google.com>
Date:   2016-09-09T22:34:39Z

    Making Dataflow Python Materialized PCollection representation more
    efficient (3 of several).

----


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---

[GitHub] incubator-beam pull request #940: [BEAM-625] Making Dataflow Python Material...

[katsiapis <gi...@git.apache.org>]

Github user katsiapis closed the pull request at:

    https://github.com/apache/incubator-beam/pull/940


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---