SPARK-1637: Clean up examples for 1.0 #571

techaddict · 2014-04-27T08:54:19Z

Move all of them into subpackages of org.apache.spark.examples (right now some are in org.apache.spark.streaming.examples, for instance, and others are in org.apache.spark.examples.mllib)
Move Python examples into examples/src/main/python
Update docs to reflect these changes

….apache.spark.streaming.examples, for instance, and others are in org.apache.spark.examples.mllib)

techaddict · 2014-04-27T08:55:30Z

streaming/src/main/scala/org/apache/spark/streaming/util/RawTextHelper.scala

@@ -22,10 +22,10 @@ import org.apache.spark.SparkContext._
 import org.apache.spark.util.collection.OpenHashMap
 import scala.collection.JavaConversions.mapAsScalaMap

-private[streaming]
+private[spark]


I'm not sure about this, needed by RawNetworkGrep and KafkaWordCount. There has to be better way.

We shouldn't use any private methods in examples. The usage of RawTextHelper in KafkaWordCount is very simple. warmUp is used in RawNetworkGrep, is it necessary? @tdas

Well, the RawTextHelper was essentially the way to provide data for running benchmarks. However, I havent quite maintained it, so that example code has been languishing in unknown state. I am inclined to keep it there as it is, as long as it hidden behind private[*]. Though I dont see the logic behind changing it from private[streaming] to private[spark]

AmplabJenkins · 2014-04-27T08:57:58Z

Can one of the admins verify this patch?

techaddict · 2014-04-27T19:35:27Z

@mateiz @rxin We already have KMeans, LogisticRegressionWithSGD and other implementation examples as main functions in their respective objects.

mateiz · 2014-04-28T20:28:09Z

Jenkins, add to whitelist and test this please

AmplabJenkins · 2014-04-28T20:32:59Z

Merged build triggered.

AmplabJenkins · 2014-04-28T20:33:09Z

Merged build started.

AmplabJenkins · 2014-04-28T21:07:05Z

Merged build finished. All automated tests passed.

AmplabJenkins · 2014-04-28T21:07:06Z

All automated tests passed.
Refer to this link for build results: https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/14539/

mengxr · 2014-04-28T23:30:07Z

@techaddict FYI, I moved main methods to examples/mllib in #584

techaddict · 2014-04-29T01:33:44Z

@mateiz @rxin so this is complete then ?

mengxr · 2014-04-29T05:11:23Z

One question: is org.apache.spark.examples a good place for examples? Examples here have access to private[spark] classes. If we are not careful, users may have trouble compiling their own apps if they use the examples as templates to build their own.

techaddict · 2014-04-29T05:15:47Z

@mengxr ya exactly, thats why i commented on the private[spark] change. I think better way would be to make the examples use there own methods, in case of private methods.

mengxr · 2014-04-29T05:51:41Z

If Apache allows to use a name space not starting with org.apache, I would recommend using spark.examples instead of org.apache.spark.examples. It also saves some typing with spark-submit.

pwendell · 2014-04-29T06:06:26Z

@mengxr I'm pretty sure we are only allowed to release classes in the org.apache namespace (unfortunately).

mengxr · 2014-04-29T16:40:29Z

I know this is a little weird, but either org.apache.examples.spark or org.apache.spark_examples can help prevent using private[spark] classes.

techaddict · 2014-04-29T16:48:01Z

@mengxr IMO we should re-implement the necessary private functions in the examples, they are not too many just 2-3 methods. It will make using/modifying the examples easier for users.

tdas · 2014-04-29T18:40:26Z

I have a general question about what should file and package organization of the examples be. For example, for Spark Streaming, is org.apache.spark.examples.streaming.* better than org.apache.spark.streaming.examples.* ? I kind of feel that the former makes mores sense, as it fits better when we refer to them as "spark streaming examples", or "graphx examples".

mengxr · 2014-04-29T20:24:52Z

@tdas I'm a little confused. Which one do you prefer, examples.streaming or streaming.examples?

tdas · 2014-04-29T20:40:07Z

Oops, sorry for the confusion. I meant "latter" instead of "former". That is, I prefer org.apache.spark.streaming.examples over org.apache.spark.examples.streaming., because

Its natural to refer to them as "spark streaming examples"
Consistent with rest of the API, where all the streaming related stuff, library and examples alike, is in org.apache.spark.streaming

tdas · 2014-04-29T20:41:54Z

Bringing @marmbrus into this discussion as he had added the Spark SQL examples in org.apache.spark.sql.examples.

marmbrus · 2014-04-29T21:46:41Z

No strong opinions here other than we should be consistent.

tdas · 2014-04-29T22:10:47Z

Okay. As I said earlier, I find it more consistent to keep it as "org.apache.spark.streaming.examples.*"

techaddict · 2014-04-30T03:53:22Z

@mengxr so what changes should i make other than streaming one suggested by tdas.

mengxr · 2014-04-30T05:02:18Z

@techaddict I don't see anything necessary except the one using private[spark] or private[streaming] methods. For consistency, org.apache.examples may be better because many examples are already under this package name.

techaddict · 2014-04-30T05:12:10Z

@mengxr what do you suggest how should we resolve the private[spark] problem ? I'm short on idea.

…m examples

AmplabJenkins · 2014-04-30T05:32:57Z

Build triggered.

AmplabJenkins · 2014-04-30T05:33:05Z

Build started.

mengxr · 2014-04-30T06:39:37Z

I would say using org.apache.spark.examples for now, which requires little code change. Then we try to avoid using private[spark] in examples. Maybe there is an automatic way to detect it.

techaddict · 2014-04-30T06:43:21Z

@mengxr @47ef86c392badc58052a0414115e49c2970b31eb looks good ?

AmplabJenkins · 2014-04-30T06:47:54Z

Build finished. All automated tests passed.

AmplabJenkins · 2014-04-30T06:47:54Z

All automated tests passed.
Refer to this link for build results: https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/14582/

mengxr · 2014-04-30T06:56:39Z

@techaddict It looks good to me.

techaddict · 2014-04-30T18:55:37Z

@mateiz ready for merging

tdas · 2014-05-01T00:04:52Z

No, I have strong feelings against "org.apache.spark.examples.streaming.*" This is entirely inconsistent with rest of the API.

techaddict · 2014-05-06T09:43:02Z

@tdas ok I'll move streaming examples to "org.apache.spark.streaming.examples.*" as you suggest. Any other changes @mateiz

mateiz · 2014-05-06T19:42:21Z

@tdas @techaddict let's not move them IMO. There are two reasons: First, when you browse the folder tree, you want to get into spark/examples and immediately see subfolders, instead of wondering "should I go into spark/examples or spark/streaming or spark/mllib or whatever". Second, it makes it less likely for us to use private[streaming] APIs in there.

mateiz · 2014-05-06T19:43:01Z

BTW the other point of this was to make them consistent across packages, and MLlib and all the Python examples are under examples/ already.

techaddict · 2014-05-06T19:45:04Z

@mateiz So this is good to merge. ?
On May 7, 2014 1:13 AM, "Matei Zaharia" [email protected] wrote:

BTW the other point of this was to make them consistent across packages,
and MLlib and all the Python examples are under examples/ already.

—
Reply to this email directly or view it on GitHub.

tdas · 2014-05-06T19:52:39Z

@mateiz I dont think that decision is that big a deal. Its fairly obvious if one sees a streaming folder inside spark/ that those are streaming examples. It feels more inconsistent to have spark.streaming.XXX suddenly become spark.XXX.streaming when everything else is spark.streaming.XXX

mateiz · 2014-05-07T00:29:19Z

Talked with TD offline and he was okay with changing the package name, so I've merged this in. Thanks Sandeep!

- [x] Move all of them into subpackages of org.apache.spark.examples (right now some are in org.apache.spark.streaming.examples, for instance, and others are in org.apache.spark.examples.mllib) - [x] Move Python examples into examples/src/main/python - [x] Update docs to reflect these changes Author: Sandeep <[email protected]> This patch had conflicts when merged, resolved by Committer: Matei Zaharia <[email protected]> Closes #571 from techaddict/SPARK-1637 and squashes the following commits: 47ef86c [Sandeep] Changes based on Discussions on PR, removing use of RawTextHelper from examples 8ed2d3f [Sandeep] Docs Updated for changes, Change for java examples 5f96121 [Sandeep] Move Python examples into examples/src/main/python 0a8dd77 [Sandeep] Move all Scala Examples to org.apache.spark.examples (some are in org.apache.spark.streaming.examples, for instance, and others are in org.apache.spark.examples.mllib)

- [x] Move all of them into subpackages of org.apache.spark.examples (right now some are in org.apache.spark.streaming.examples, for instance, and others are in org.apache.spark.examples.mllib) - [x] Move Python examples into examples/src/main/python - [x] Update docs to reflect these changes Author: Sandeep <[email protected]> This patch had conflicts when merged, resolved by Committer: Matei Zaharia <[email protected]> Closes #571 from techaddict/SPARK-1637 and squashes the following commits: 47ef86c [Sandeep] Changes based on Discussions on PR, removing use of RawTextHelper from examples 8ed2d3f [Sandeep] Docs Updated for changes, Change for java examples 5f96121 [Sandeep] Move Python examples into examples/src/main/python 0a8dd77 [Sandeep] Move all Scala Examples to org.apache.spark.examples (some are in org.apache.spark.streaming.examples, for instance, and others are in org.apache.spark.examples.mllib) (cherry picked from commit a000b5c) Signed-off-by: Matei Zaharia <[email protected]>

…e PR apache#571 was last updated. This resulted in Compilation Errors.

@mateiz

…e PR #571 was last updated. This resulted in Compilation Errors. cc @mateiz project not compiling currently. Author: Sandeep <[email protected]> Closes #673 from techaddict/SPARK-1637-HOTFIX and squashes the following commits: b512f4f [Sandeep] [SPARK-1637][HOTFIX] There are some Streaming examples added after the PR #571 was last updated. This resulted in Compilation Errors.

@mateiz

…e PR #571 was last updated. This resulted in Compilation Errors. cc @mateiz project not compiling currently. Author: Sandeep <[email protected]> Closes #673 from techaddict/SPARK-1637-HOTFIX and squashes the following commits: b512f4f [Sandeep] [SPARK-1637][HOTFIX] There are some Streaming examples added after the PR #571 was last updated. This resulted in Compilation Errors. (cherry picked from commit fdae095) Signed-off-by: Patrick Wendell <[email protected]>

SPARK-1072 Use binary search when needed in RangePartioner Author: Holden Karau <[email protected]> Closes apache#571 and squashes the following commits: f31a2e1 [Holden Karau] Swith to using CollectionsUtils in Partitioner 4c7a0c3 [Holden Karau] Add CollectionsUtil as suggested by aarondav 7099962 [Holden Karau] Add the binary search to only init once 1bef01d [Holden Karau] CR feedback a21e097 [Holden Karau] Use binary search if we have more than 1000 elements inside of RangePartitioner

- [x] Move all of them into subpackages of org.apache.spark.examples (right now some are in org.apache.spark.streaming.examples, for instance, and others are in org.apache.spark.examples.mllib) - [x] Move Python examples into examples/src/main/python - [x] Update docs to reflect these changes Author: Sandeep <[email protected]> This patch had conflicts when merged, resolved by Committer: Matei Zaharia <[email protected]> Closes apache#571 from techaddict/SPARK-1637 and squashes the following commits: 47ef86c [Sandeep] Changes based on Discussions on PR, removing use of RawTextHelper from examples 8ed2d3f [Sandeep] Docs Updated for changes, Change for java examples 5f96121 [Sandeep] Move Python examples into examples/src/main/python 0a8dd77 [Sandeep] Move all Scala Examples to org.apache.spark.examples (some are in org.apache.spark.streaming.examples, for instance, and others are in org.apache.spark.examples.mllib)

@mateiz

…e PR apache#571 was last updated. This resulted in Compilation Errors. cc @mateiz project not compiling currently. Author: Sandeep <[email protected]> Closes apache#673 from techaddict/SPARK-1637-HOTFIX and squashes the following commits: b512f4f [Sandeep] [SPARK-1637][HOTFIX] There are some Streaming examples added after the PR apache#571 was last updated. This resulted in Compilation Errors.

Surfacing decoders on KafkaInputDStream

The new role "install-docker" install docker-ce by using get.docker.com script, it support multiple OS and arch, and keep consistent docker-ce version, mark role "install-docker-ce" as deprecated, and apply new role in all of docker-machine jobs. Update tag of docker-machine job from latest-release to 18.06 in order to avoid confusions. Related-Bug: theopenlab/openlab#308

Move all Scala Examples to org.apache.spark.examples (some are in org…

0a8dd77

….apache.spark.streaming.examples, for instance, and others are in org.apache.spark.examples.mllib)

techaddict reviewed Apr 27, 2014
View reviewed changes

techaddict added 2 commits April 27, 2014 14:30

Move Python examples into examples/src/main/python

5f96121

Docs Updated for changes, Change for java examples

8ed2d3f

Changes based on Discussions on PR, removing use of RawTextHelper fro…

47ef86c

…m examples

techaddict closed this May 7, 2014

techaddict added a commit to techaddict/spark that referenced this pull request May 7, 2014

[SPARK-1637][HOTFIX] There are some Streaming examples added after th…

b512f4f

…e PR apache#571 was last updated. This resulted in Compilation Errors.

meisam pushed a commit to meisam/spark that referenced this pull request Sep 24, 2014

Merge pull request apache#571 from Reinvigorate/sm-kafka-serializers

cfcda95

Surfacing decoders on KafkaInputDStream

SPARK-1637: Clean up examples for 1.0 #571

SPARK-1637: Clean up examples for 1.0 #571

Conversation

techaddict commented Apr 27, 2014

techaddict Apr 27, 2014

Choose a reason for hiding this comment

mengxr Apr 29, 2014

Choose a reason for hiding this comment

tdas Apr 29, 2014

Choose a reason for hiding this comment

AmplabJenkins commented Apr 27, 2014

techaddict commented Apr 27, 2014

mateiz commented Apr 28, 2014

AmplabJenkins commented Apr 28, 2014

AmplabJenkins commented Apr 28, 2014

AmplabJenkins commented Apr 28, 2014

AmplabJenkins commented Apr 28, 2014

mengxr commented Apr 28, 2014

techaddict commented Apr 29, 2014

mengxr commented Apr 29, 2014

techaddict commented Apr 29, 2014

mengxr commented Apr 29, 2014

pwendell commented Apr 29, 2014

mengxr commented Apr 29, 2014

techaddict commented Apr 29, 2014

tdas commented Apr 29, 2014

mengxr commented Apr 29, 2014

tdas commented Apr 29, 2014

tdas commented Apr 29, 2014

marmbrus commented Apr 29, 2014

tdas commented Apr 29, 2014

techaddict commented Apr 30, 2014

mengxr commented Apr 30, 2014

techaddict commented Apr 30, 2014

AmplabJenkins commented Apr 30, 2014

AmplabJenkins commented Apr 30, 2014

mengxr commented Apr 30, 2014

techaddict commented Apr 30, 2014

AmplabJenkins commented Apr 30, 2014

AmplabJenkins commented Apr 30, 2014

mengxr commented Apr 30, 2014

techaddict commented Apr 30, 2014

tdas commented May 1, 2014

techaddict commented May 6, 2014

mateiz commented May 6, 2014

mateiz commented May 6, 2014

techaddict commented May 6, 2014

tdas commented May 6, 2014

mateiz commented May 7, 2014