Make shell component serialisation plugable #697
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
The new implementation separates the various concerns formerly mixed up in the example code. Also, we are now using tick tuples (introduced in Storm 0.8) instead of spawning manual threads for carrying out periodical tasks. Lastly, we add 192 unit tests for the new implementation that brings the test coverage for the Rolling Count from 0% to almost 100%. Note: Adding those unit tests required changes to the build (m2-pom.xml), notably new test dependencies and moving the existing Java code from src/jvm/* to src/jvm/main/*. The latter was required so that the test runner triggered by Maven can tell code (src/jvm/main) and tests (src/jvm/test) apart.
Complete refactoring of the Rolling Count example
SlidingWindowCounterTest: add missing @test annotation to unit test
Yes, we iterate through the list twice, but we no longer have a function that does two things.
This commit is best explained by describing the behavior of RollingTopWords in its absence. Let's say the word "Bieber" appears in the tweet stream at a very high rate for half an hour, and then is *never heard again*. This will insert the word "Bieber" into our rankings object downstream. Note that there is nothing to *remove* the word "Bieber" from the rankings object downstream. Let us assume that, during its half-hour, the word "Bieber" appears *more than five times as often* as other highly-ranked words. Then the very *last* "Bieber" report (in which "Bieber" has only appeared for the first minute of the five-minute window) will still be able to hold a place in the rankings. Objects are only lowered in the rankings by being outranked, or by *appearing again* with a lower score. *With* this change, we wipe zeros, and *then* wipe a slot. This means that when we wipe the slot containing the last Bieber mentions, we don't at that moment remove "Bieber" from our map. The *next* report will include "Bieber" with a score of zero, removing "Bieber" from the ranking object downstream.
added kafkaspout lag to transactionaltridentkafkaspout
Ensure zeros get reported.
update to storm 0.8.2
…artitions is ever called
…or-storm STORM-196: When JVM_OPTS are set, storm jar fails to detect storm.jar from environment
…o detect storm.jar from environment
…args Conflicts: pom.xml storm-core/src/clj/backtype/storm/ui/core.clj
Utils#readCommandLineOpts (STORM-173)
…bator-storm STORM-194: Support list of strings in *.worker.childopts, handle spaces
…com/dschiavu/incubator-storm STORM-173: Treat command line "-c" option number config values as such
Conflicts: conf/defaults.yaml
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This pull request makes multilang components plugable (issue #373) and relates to issue #654, in that it is now possible to implement protocol buffer serialisation.
Changes in this pull request:
Moved multilang classes to multilang directory.
Added BoltMsg, ShellMSg and SpoutMsg objects. These objects are used when communicating with a serializer, instead of raw JSON.
All JSON serialisation has been moved to a new JsonSerializer.
A serialiser is selected when creating a new ShellBolt or ShellSpout, by passing it in to the constructor. A default option of no serialiser automatically selects the built-in JsonSerializer.