You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@edgent.apache.org by qu...@apache.org on 2016/07/26 22:31:33 UTC
[49/51] [partial] incubator-quarks-website git commit: from
05ad5e885c21d745f87c23c280f222cbca457b89
http://git-wip-us.apache.org/repos/asf/incubator-quarks-website/blob/a13739b3/content/algolia_search.json
----------------------------------------------------------------------
diff --git a/content/algolia_search.json b/content/algolia_search.json
index 1bb9ec3..9214db2 100644
--- a/content/algolia_search.json
+++ b/content/algolia_search.json
@@ -13,7 +13,7 @@
"keywords": "",
"url": "../docs/committers",
"summary": "",
-"body": "## Commit activityTo see commit activity for Apache Quarks, click [here](https://github.com/apache/incubator-quarks/pulse).## How to become a committerYou can become a committer by contributing to Apache Quarks. Qualifications for a new committer include:* **Sustained Contributions**: Potential committers should have a history of contributions to Quarks. They will create pull requests over a period of time.* **Quality of Contributions**: Potential committers should submit code that adds value to Quarks, including tests and documentation as needed. They should comment in a positive way on issues and pull requests, providing guidance and input to improve Quarks.* **Community Involvement**: Committers should participate in discussions in a positive way, triage and fix bugs, and interact with users who submit questions. They will remain courteous, and helpful, and encourage new members to join the Quarks community."
+"body": "## Commit activityTo see commit activity for Apache Edgent, click [here]({{ site.data.project.source_repository_mirror }}/pulse).## How to become a committerYou can become a committer by contributing to Apache Edgent. Qualifications for a new committer include:* **Sustained Contributions**: Potential committers should have a history of contributions to Edgent. They will create pull requests over a period of time.* **Quality of Contributions**: Potential committers should submit code that adds value to Edgent, including tests and documentation as needed. They should comment in a positive way on issues and pull requests, providing guidance and input to improve Edgent.* **Community Involvement**: Committers should participate in discussions in a positive way, triage and fix bugs, and interact with users who submit questions. They will remain courteous, and helpful, and encourage new members to join the Edgent community."
},
@@ -21,12 +21,12 @@
{
-"title": "Common Quarks operations",
+"title": "Common Edgent operations",
"tags": "",
"keywords": "",
-"url": "../docs/common-quarks-operations",
+"url": "../docs/common-edgent-operations",
"summary": "",
-"body": "In the [Getting started guide](quarks-getting-started), we covered a Quarks application where we read from a device's simulated temperature sensor. Yet Quarks supports more operations than simple filtering. Data analysis and streaming require a suite of functionality, the most important components of which will be outlined below.## TStream.map()`TStream.map()` is arguably the most used method in the Quarks API. Its two main purposes are to perform stateful or stateless operations on a stream's tuples, and to produce a `TStream` with tuples of a different type from that of the calling stream.### Changing a TStream's tuple typeIn addition to filtering tuples, `TStream`s support operations that *transform* tuples from one Java type to another by invoking the `TStream.map()` method.This is useful in cases such as calculating the floating point average of a list of `Integer`s, or tokenizing a Java String into a list of `String`s. To demonstrate this, let's say we have a `TStream
` which contains a few lines, each of which contains multiple words:```javaTStream lines = topology.strings( \"this is a line\", \"this is another line\", \"there are three lines now\", \"and now four\");```We then want to print the third word in each line. The best way to do this is to convert each line to a list of `String`s by tokenizing them. We can do this in one line of code with the `TStream.map()` method:```javaTStream > wordsInLine = lines.map(tuple -> Arrays.asList(tuple.split(\" \")));```Since each tuple is now a list of strings, the `wordsInLine` stream is of type `List`. As you can see, the `map()` method has the ability to change the type of the `TStream`. Finally, we can use the `wordsInLine` stream to print the third word in each line.```javawordsInLine.sink(list -> System.out.println(list.get(2)));```As mentioned in the [Getting started guide](quarks-getting-started), a `TStream` can be parameterized to any serializable Java type, including ones created
by the user.### Performing stateful operationsIn all previous examples, the operations performed on a `TStream` have been stateless; keeping track of information over multiple invocations of the same operation has not been necessary. What if we want to keep track of the number of Strings sent over a stream? To do this, we need our `TStream.map()` method to contain a counter as state.This can be achieved by creating an anonymous `Function` class, and giving it the required fields.```javaTStream streamOfStrings = ...;TStream counts = streamOfStrings.map(new Function() { int count = 0; @Override public Integer apply(String arg0) { count = count + 1; return count; }});```The `count` field will now contain the number of `String`s which were sent over `streamOfStrings`. Although this is a simple example, the anonymous `Function` passed to `TStream.map()` can contain any kind of state! This could be a `HashMap`, a running list of tuples, or any serializable Java t
ype. The state will be maintained throughout the entire runtime of your application."
+"body": "In the [Getting started guide](edgent-getting-started), we covered an Edgent application where we read from a device's simulated temperature sensor. Yet Edgent supports more operations than simple filtering. Data analysis and streaming require a suite of functionality, the most important components of which will be outlined below.## TStream.map()`TStream.map()` is arguably the most used method in the Edgent API. Its two main purposes are to perform stateful or stateless operations on a stream's tuples, and to produce a `TStream` with tuples of a different type from that of the calling stream.### Changing a TStream's tuple typeIn addition to filtering tuples, `TStream`s support operations that *transform* tuples from one Java type to another by invoking the `TStream.map()` method.This is useful in cases such as calculating the floating point average of a list of `Integer`s, or tokenizing a Java String into a list of `String`s. To demonstrate this, let's say we have a `TStrea
m` which contains a few lines, each of which contains multiple words:```javaTStream lines = topology.strings( \"this is a line\", \"this is another line\", \"there are three lines now\", \"and now four\");```We then want to print the third word in each line. The best way to do this is to convert each line to a list of `String`s by tokenizing them. We can do this in one line of code with the `TStream.map()` method:```javaTStream > wordsInLine = lines.map(tuple -> Arrays.asList(tuple.split(\" \")));```Since each tuple is now a list of strings, the `wordsInLine` stream is of type `List`. As you can see, the `map()` method has the ability to change the type of the `TStream`. Finally, we can use the `wordsInLine` stream to print the third word in each line.```javawordsInLine.sink(list -> System.out.println(list.get(2)));```As mentioned in the [Getting started guide](edgent-getting-started), a `TStream` can be parameterized to any serializable Java type, including ones created
by the user.### Performing stateful operationsIn all previous examples, the operations performed on a `TStream` have been stateless; keeping track of information over multiple invocations of the same operation has not been necessary. What if we want to keep track of the number of Strings sent over a stream? To do this, we need our `TStream.map()` method to contain a counter as state.This can be achieved by creating an anonymous `Function` class, and giving it the required fields.```javaTStream streamOfStrings = ...;TStream counts = streamOfStrings.map(new Function() { int count = 0; @Override public Integer apply(String arg0) { count = count + 1; return count; }});```The `count` field will now contain the number of `String`s which were sent over `streamOfStrings`. Although this is a simple example, the anonymous `Function` passed to `TStream.map()` can contain any kind of state! This could be a `HashMap`, a running list of tuples, or any serializable Java
type. The state will be maintained throughout the entire runtime of your application."
},
@@ -34,12 +34,12 @@
{
-"title": "Apache Quarks community",
+"title": "Apache Edgent community",
"tags": "",
"keywords": "",
"url": "../docs/community",
"summary": "",
-"body": "Every volunteer project obtains its strength from the people involved in it. We invite you to participate as much or as little as you choose.You can:* Use our project and provide a feedback.* Provide us with the use-cases.* Report bugs and submit patches.* Contribute code, javadocs, documentation.Visit the [Contributing](http://www.apache.org/foundation/getinvolved.html) page for general Apache contribution information. If you plan to make any significant contribution, you will need to have an Individual Contributor License Agreement [\\(ICLA\\)](https://www.apache.org/licenses/icla.txt) on file with Apache.## Mailing listGet help using {{ site.data.project.short_name }} or contribute to the project on our mailing lists:{% if site.data.project.user_list %}* [site.data.project.user_list](mailto:{{ site.data.project.user_list }}) is for usage questions, help, and announcements. [subscribe](mailto:{{ site.data.project.user_list_subscribe }}?subject=send this email to subscribe
), [unsubscribe](mailto:{{ site.data.project.dev_list_unsubscribe }}?subject=send this email to unsubscribe), [archives]({{ site.data.project.user_list_archive_mailarchive }}){% endif %}* [{{ site.data.project.dev_list }}](mailto:{{ site.data.project.dev_list }}) is for people who want to contribute code to {{ site.data.project.short_name }}. [subscribe](mailto:{{ site.data.project.dev_list_subscribe }}?subject=send this email to subscribe), [unsubscribe](mailto:{{ site.data.project.dev_list_unsubscribe }}?subject=send this email to unsubscribe), [Apache archives]({{ site.data.project.dev_list_archive }}), [mail-archive.com archives]({{ site.data.project.dev_list_archive_mailarchive }})* [{{ site.data.project.commits_list }}](mailto:{{ site.data.project.commits_list }}) is for commit messages and patches to {{ site.data.project.short_name }}. [subscribe](mailto:{{ site.data.project.commits_list_subscribe }}?subject=send this email to subscribe), [unsubscribe](mailto:{{ site.data.pro
ject.commits_list_unsubscribe }}?subject=send this email to unsubscribe), [Apache archives]({{ site.data.project.commits_list_archive }}), [mail-archive.com archives]({{ site.data.project.commits_list_archive_mailarchive }})## Issue trackerWe use Jira here: [https://issues.apache.org/jira/browse/{{ site.data.project.jira }}](https://issues.apache.org/jira/browse/{{ site.data.project.jira }})### Bug reportsFound bug? Create an issue in [Jira](https://issues.apache.org/jira/browse/{{ site.data.project.jira }}).Before submitting an issue, please:* Verify that the bug does in fact exist* Search the issue tracker to verify there is no existing issue reporting the bug you've found* Consider tracking down the bug yourself in the {{ site.data.project.short_name }} source and submitting a pull request along with your bug report. This is a great time saver for the {{ site.data.project.short_name }} developers and helps ensure the bug will be fixed quickly.### Feature requestsEnhancement reque
sts for new features are also welcome. The more concrete the request is and the better rationale you provide, the greater the chance it will incorporated into future releases. To make a request, create an issue in [Jira](https://issues.apache.org/jira/browse/{{ site.data.project.jira }}).## Source codeThe project sources are accessible via the [source code repository]({{ site.data.project.source_repository }}) which is also mirrored in [GitHub]({{ site.data.project.source_repository_mirror }}).When you are considering a code contribution, make sure there is an [Jira issue](https://issues.apache.org/jira/browse/{{ site.data.project.jira }}) that describes your work or the bug you are fixing. For significant contributions, please discuss your proposed changes in the issue so that others can comment on your plans. Someone else may be working on the same functionality, so it's good to communicate early and often. A committer is more likely to accept your change if there is clear informa
tion in the issue.To contribute, [fork](https://help.github.com/articles/fork-a-repo/) the [mirror]({{ site.data.project.source_repository_mirror }}) and issue a [pull request](https://help.github.com/articles/using-pull-requests/). Put the Jira issue number, e.g. {{ site.data.project.jira }}-100 in the pull request title. The tag [WIP] can also be used in the title of pull requests to indicate that you are not ready to merge but want feedback. Remove [WIP] when you are ready for merge. Make sure you document your code and contribute tests along with the code.Read [DEVELOPMENT.md](https://github.com/apache/incubator-quarks/blob/master/DEVELOPMENT.md) at the top of the code tree for details on setting up your development environment.## Web site and documentation source codeThe project website and documentation sources are accessible via the [website source code repository]({{ site.data.project.website_repository }}) which is also mirrored in [GitHub]({{ site.data.project.website_repo
sitory_mirror }}). Contributing changes to the web site and documentation is similar to contributing code. Follow the instructions in the *Source Code* section above, but fork and issue a pull request against the [web site mirror]({{ site.data.project.website_repository_mirror }}). Follow the instructions in the top-level [README.md]({{ site.data.project.website_repository_mirror }}/blob/master/README.md) for details on contributing to the web site and documentation.You will need to use [Markdown](https://daringfireball.net/projects/markdown/) and [Jekyll](http://jekyllrb.com) to develop pages. See:* [Markdown Cheat Sheet](https://github.com/adam-p/markdown-here/wiki/Markdown-Cheatsheet)* [Jekyll on linux and Mac](https://jekyllrb.com/)* [Jekyll on Windows](https://jekyllrb.com/docs/windows/) is not officially supported but people have gotten it to work"
+"body": "Every volunteer project obtains its strength from the people involved in it. We invite you to participate as much or as little as you choose.You can:* Use our project and provide a feedback.* Provide us with the use-cases.* Report bugs and submit patches.* Contribute code, javadocs, documentation.Visit the [Contributing](http://www.apache.org/foundation/getinvolved.html) page for general Apache contribution information. If you plan to make any significant contribution, you will need to have an Individual Contributor License Agreement [\\(ICLA\\)](https://www.apache.org/licenses/icla.txt) on file with Apache.## Mailing listGet help using {{ site.data.project.short_name }} or contribute to the project on our mailing lists:{% if site.data.project.user_list %}* [site.data.project.user_list](mailto:{{ site.data.project.user_list }}) is for usage questions, help, and announcements. [subscribe](mailto:{{ site.data.project.user_list_subscribe }}?subject=send this email to subscribe
), [unsubscribe](mailto:{{ site.data.project.dev_list_unsubscribe }}?subject=send this email to unsubscribe), [archives]({{ site.data.project.user_list_archive_mailarchive }}){% endif %}* [{{ site.data.project.dev_list }}](mailto:{{ site.data.project.dev_list }}) is for people who want to contribute code to {{ site.data.project.short_name }}. [subscribe](mailto:{{ site.data.project.dev_list_subscribe }}?subject=send this email to subscribe), [unsubscribe](mailto:{{ site.data.project.dev_list_unsubscribe }}?subject=send this email to unsubscribe), [Apache archives]({{ site.data.project.dev_list_archive }}), [mail-archive.com archives]({{ site.data.project.dev_list_archive_mailarchive }})* [{{ site.data.project.commits_list }}](mailto:{{ site.data.project.commits_list }}) is for commit messages and patches to {{ site.data.project.short_name }}. [subscribe](mailto:{{ site.data.project.commits_list_subscribe }}?subject=send this email to subscribe), [unsubscribe](mailto:{{ site.data.pro
ject.commits_list_unsubscribe }}?subject=send this email to unsubscribe), [Apache archives]({{ site.data.project.commits_list_archive }}), [mail-archive.com archives]({{ site.data.project.commits_list_archive_mailarchive }})## Issue trackerWe use JIRA here: [https://issues.apache.org/jira/browse/{{ site.data.project.jira }}](https://issues.apache.org/jira/browse/{{ site.data.project.jira }})### Bug reportsFound a bug? Create an issue in [JIRA](https://issues.apache.org/jira/browse/{{ site.data.project.jira }}).Before submitting an issue, please:* Verify that the bug does in fact exist* Search the issue tracker to verify there is no existing issue reporting the bug you've found* Consider tracking down the bug yourself in the {{ site.data.project.short_name }} source and submitting a pull request along with your bug report. This is a great time saver for the {{ site.data.project.short_name }} developers and helps ensure the bug will be fixed quickly.### Feature requestsEnhancement req
uests for new features are also welcome. The more concrete the request is and the better rationale you provide, the greater the chance it will incorporated into future releases. To make a request, create an issue in [JIRA](https://issues.apache.org/jira/browse/{{ site.data.project.jira }}).## Source codeThe project sources are accessible via the [source code repository]({{ site.data.project.source_repository }}) which is also mirrored in [GitHub]({{ site.data.project.source_repository_mirror }}).When you are considering a code contribution, make sure there is a [JIRA issue](https://issues.apache.org/jira/browse/{{ site.data.project.jira }}) that describes your work or the bug you are fixing. For significant contributions, please discuss your proposed changes in the issue so that others can comment on your plans. Someone else may be working on the same functionality, so it's good to communicate early and often. A committer is more likely to accept your change if there is clear inform
ation in the issue.To contribute, [fork](https://help.github.com/articles/fork-a-repo/) the [mirror]({{ site.data.project.source_repository_mirror }}) and issue a [pull request](https://help.github.com/articles/using-pull-requests/). Put the JIRA issue number, e.g. {{ site.data.project.jira }}-100 in the pull request title. The tag [WIP] can also be used in the title of pull requests to indicate that you are not ready to merge but want feedback. Remove [WIP] when you are ready for merge. Make sure you document your code and contribute tests along with the code.Read [DEVELOPMENT.md]({{ site.data.project.source_repository_mirror }}/blob/master/DEVELOPMENT.md) at the top of the code tree for details on setting up your development environment.## Website and documentation source codeThe project website and documentation sources are accessible via the [website source code repository]({{ site.data.project.website_repository }}) which is also mirrored in [GitHub]({{ site.data.project.websit
e_repository_mirror }}). Contributing changes to the website and documentation is similar to contributing code. Follow the instructions in the *Source Code* section above, but fork and issue a pull request against the [website mirror]({{ site.data.project.website_repository_mirror }}). Follow the instructions in the top-level [README.md]({{ site.data.project.website_repository_mirror }}/blob/master/README.md) for details on contributing to the website and documentation.You will need to use [Markdown](https://daringfireball.net/projects/markdown/) and [Jekyll](http://jekyllrb.com) to develop pages. See:* [Markdown Cheat Sheet](https://github.com/adam-p/markdown-here/wiki/Markdown-Cheatsheet)* [Jekyll on Linux and Mac](https://jekyllrb.com/)* [Jekyll on Windows](https://jekyllrb.com/docs/windows/) is not officially supported but people have gotten it to work"
},
@@ -52,7 +52,7 @@
"keywords": "",
"url": "../docs/console",
"summary": "",
-"body": "## Visualizing and monitoring your applicationThe Quarks application console is a web application that enables you to visualize your application topology and monitor the tuples flowing through your application. The kind of oplets used in the topology, as well as the stream tags included in the topology, are also visible in the console.## Adding the console web app to your applicationTo use the console, you must use the Quarks classes that provide the service to access the console web application or directly call the `HttpServer` class itself, start the server and then obtain the console URL.The easiest way to include the console in your application is to use the the `DevelopmentProvider` class. `DevelopmentProvider` is a subclass of `DirectProvider` and adds services such as access to the console web application and counter oplets used to determine tuple counts. You can get the URL for the console from the `DevelopmentProvider` using the `getService` method as shown in a hy
pothetical application shown below:```javaimport java.util.concurrent.TimeUnit;import quarks.console.server.HttpServer;import quarks.providers.development.DevelopmentProvider;import quarks.topology.TStream;import quarks.topology.Topology;public class TempSensorApplication { public static void main(String[] args) throws Exception { TempSensor sensor = new TempSensor(); DevelopmentProvider dp = new DevelopmentProvider(); Topology topology = dp.newTopology(); TStream tempReadings = topology.poll(sensor, 1, TimeUnit.MILLISECONDS); TStream filteredReadings = tempReadings.filter(reading -> reading 80); filteredReadings.print(); System.out.println(dp.getServices().getService(HttpServer.class).getConsoleUrl()); dp.submit(topology); }}```Note that the console URL is being printed to `System.out`. The `filteredReadings` are as well, since `filteredReadings.print()` is being called in the application. You may need to scroll your te
rminal window up to see the output for the console URL.Optionally, you can modify the above code in the application to have a timeout before submitting the topology, which would allow you to see the console URL before any other output is shown. The modification would look like this:```java// Print the console URL and wait for 10 seconds before submitting the topologySystem.out.println(dp.getServices().getService(HttpServer.class).getConsoleUrl());try { TimeUnit.SECONDS.sleep(10);} catch (InterruptedException e) { // Do nothing}dp.submit(topology);```The other way to embed the console in your application is shown in the `HttpServerSample.java` example (on [GitHub](https://github.com/apache/incubator-quarks/blob/master/samples/console/src/main/java/quarks/samples/console/HttpServerSample.java)). It gets the `HttpServer` instance, starts it, and prints out the console URL. Note that it does not submit a job, so when the console is displayed in the browser, there are no running jo
bs and therefore no topology graph. The example is meant to show how to get the `HttpServer` instance, start the console web app and get the URL of the console.## Accessing the consoleThe console URL has the following format:`http://host_name:port_number/console`Once it is obtained from `System.out`, enter it in a browser window.If you cannot access the console at this URL, ensure there is a `console.war` file in the `webapps` directory. If the `console.war` file cannot be found, an exception will be thrown (in `std.out`) indicating `console.war` was not found.## ConsoleWaterDetector sampleTo see the features of the console in action and as a way to demonstrate how to monitor a topology in the console, let's look at the `ConsoleWaterDetector` sample (on [GitHub](https://github.com/apache/incubator-quarks/blob/master/samples/console/src/main/java/quarks/samples/console/ConsoleWaterDetector.java)).Prior to running any console applications, the `console.war` file must be built as menti
oned above. If you are building quarks from a Git repository, go to the top level Quarks directory and run `ant`.Here is an example in my environment:```Susans-MacBook-Pro-247:quarks susancline$ pwd/Users/susancline/git/quarksSusans-MacBook-Pro-247:quarks susancline$ antBuildfile: /Users/susancline/git/quarks/build.xmlsetcommitversion:init:suball:init:project.component:compile:...[javadoc] Constructing Javadoc information...[javadoc] Standard Doclet version 1.8.0_71[javadoc] Building tree for all the packages and classes...[javadoc] Generating /Users/susancline/git/quarks/target/docs/javadoc/quarks/analytics/sensors/package-summary.html...[javadoc] Copying file /Users/susancline/git/quarks/analytics/sensors/src/main/java/quarks/analytics/sensors/doc-files/deadband.png to directory /Users/susancline/git/quarks/target/docs/javadoc/quarks/analytics/sensors/doc-files...[javadoc] Generating /Users/susancline/git/quarks/target/docs/javadoc/quarks/topology/package-summary.html...[javadoc]
Copying file /Users/susancline/git/quarks/api/topology/src/main/java/quarks/topology/doc-files/sources.html to directory /Users/susancline/git/quarks/target/docs/javadoc/quarks/topology/doc-files...[javadoc] Building index for all the packages and classes...[javadoc] Building index for all classes...all:BUILD SUCCESSFULTotal time: 3 seconds```This command will let you know that `console.war` was built and is in the correct place, under the `webapps` directory.```Susans-MacBook-Pro-247:quarks susancline$ find . -name console.war -print./target/java8/console/webapps/console.war```Now we know we have built `console.war`, so we're good to go. To run this sample from the command line:```Susans-MacBook-Pro-247:quarks susancline$ pwd/Users/susancline/git/quarksSusans-MacBook-Pro-247:quarks susancline$ java -cp target/java8/samples/lib/quarks.samples.console.jar:. quarks.samples.console.ConsoleWaterDetector```If everything is successful, you'll start seeing output. You may have to scroll ba
ck up to get the URL of the console:```Susans-MacBook-Pro-247:quarks susancline$ java -cp target/java8/samples/lib/quarks.samples.console.jar:. quarks.samples.console.ConsoleWaterDetectorMar 07, 2016 12:04:52 PM org.eclipse.jetty.util.log.Log initializedINFO: Logging initialized @176msMar 07, 2016 12:04:53 PM org.eclipse.jetty.server.Server doStartINFO: jetty-9.3.6.v20151106Mar 07, 2016 12:04:53 PM org.eclipse.jetty.server.handler.ContextHandler doStartINFO: Started o.e.j.s.ServletContextHandler@614c5515{/jobs,null,AVAILABLE}Mar 07, 2016 12:04:53 PM org.eclipse.jetty.server.handler.ContextHandler doStartINFO: Started o.e.j.s.ServletContextHandler@77b52d12{/metrics,null,AVAILABLE}Mar 07, 2016 12:04:53 PM org.eclipse.jetty.webapp.StandardDescriptorProcessor visitServletINFO: NO JSP Support for /console, did not find org.eclipse.jetty.jsp.JettyJspServletMar 07, 2016 12:04:53 PM org.eclipse.jetty.server.handler.ContextHandler doStartINFO: Started o.e.j.w.WebAppContext@2d554825{/console,
file:///private/var/folders/0c/pb4rznhj7sbc886t30w4vpxh0000gn/T/jetty-0.0.0.0-0-console.war-_console-any-3101338829524954950.dir/webapp/,AVAILABLE}{/console.war}Mar 07, 2016 12:04:53 PM org.eclipse.jetty.server.AbstractConnector doStartINFO: Started ServerConnector@66480dd7{HTTP/1.1,[http/1.1]}{0.0.0.0:57964}Mar 07, 2016 12:04:53 PM org.eclipse.jetty.server.Server doStartINFO: Started @426mshttp://localhost:57964/consoleWell1 alert, ecoli value is 1Well1 alert, temp value is 48Well3 alert, ecoli value is 1```Now point your browser to the URL displayed above in the output from running the Java command to launch the `ConsoleWaterDetector` application. In this case, the URL is `http://localhost:57964/console`.Below is a screen shot of what you should see if everything is working properly:## ConsoleWaterDetector application scenarioThe application is now running in your browser. Let's discuss the scenario for the application.A county agency is responsible for ensuring the safety of resi
dents well water. Each well they monitor has four different sensor types:* Temperature* Acidity* Ecoli* LeadThe sample application topology monitors 3 wells:* For the hypothetical scenario, Well1 and Well3 produce 'unhealthy' values from their sensors on occasion. Well2 always produces 'healthy' values.* Each well that is to be measured is added to the topology. The topology polls each sensor (temp, ecoli, etc.) for each well as a unit. A `TStream` is returned from polling the toplogy and represents a sensor reading. Each sensor reading for the well has a tag added to it with the reading type i.e, \"temp\", and the well id. Once all of the sensor readings are obtained and the tags added, each sensor reading is 'unioned' into a single `TStream`. Look at the `waterDetector` method for details on this.* Now, each well has a single stream with each of the sensors readings as a property with a name and value in the `TStream`. Next the `alertFilter` method is called on the `TStream` repre
senting each well. This method checks the values for each well's sensors to determine if they are 'out of range' for healthy values. The `filter` oplet is used to do this. If any of the sensor's readings are out of the acceptable range the tuple is passed along. Those that are within an acceptable range are discarded.* Next the applications' `splitAlert` method is called on each well's stream that contains the union of all the sensor readings that are out of range. The `splitAlert` method uses the `split` oplet to split the incoming stream into 5 different streams. Only those tuples that are out of range for each stream, which represents each sensor type, will be returned. The object returned from `splitAlert` is a list of `TStream` objects. The `splitAlert` method is shown below: ```java public static List> splitAlert(TStream alertStream, int wellId) { List> allStreams = alertStream.split(5, tuple -> { if (tuple.get(\"temp\") != null) { JsonOb
ject tempObj = new JsonObject(); int temp = tuple.get(\"temp\").getAsInt(); if (temp = TEMP_ALERT_MAX) { tempObj.addProperty(\"temp\", temp); return 0; } else { return -1; } } else if (tuple.get(\"acidity\") != null){ JsonObject acidObj = new JsonObject(); int acid = tuple.get(\"acidity\").getAsInt(); if (acid = ACIDITY_ALERT_MAX) { acidObj.addProperty(\"acidity\", acid); return 1; } else { return -1; } } else if (tuple.get(\"ecoli\") != null) { JsonObject ecoliObj = new JsonObject(); int ecoli = tuple.get(\"ecoli\").getAsInt(); if (ecoli >= ECOLI_ALERT) { ecoliObj.addProperty(\"ecoli\", ecoli); return 2; } else {
return -1; } } else if (tuple.get(\"lead\") != null) { JsonObject leadObj = new JsonObject(); int lead = tuple.get(\"lead\").getAsInt(); if (lead >= LEAD_ALERT_MAX) { leadObj.addProperty(\"lead\", lead); return 3; } else { return -1; } } else { return -1; } }); return allStreams; } ```* Next we want to get the temperature stream from the first well and put a rate meter on it to determine the rate at which the out of range values are flowing in the stream ```java List> individualAlerts1 = splitAlert(filteredReadings1, 1); // Put a rate meter on well1's temperature sensor output Metrics.rateMeter(individualAlerts1.get(0)); ```* Next all the sensors for well 1 have tags added to the stream indicating the stream is out of range for that sensor and
the well id. Next a sink is added, passing the tuple to a `Consumer` that formats a string to `System.out` containing the well id, alert type (sensor type) and value of the sensor. ```java // Put a rate meter on well1's temperature sensor output Metrics.rateMeter(individualAlerts1.get(0)); individualAlerts1.get(0).tag(TEMP_ALERT_TAG, \"well1\").sink(tuple -> System.out.println(\"\\n\" + formatAlertOutput(tuple, \"1\", \"temp\"))); individualAlerts1.get(1).tag(ACIDITY_ALERT_TAG, \"well1\").sink(tuple -> System.out.println(formatAlertOutput(tuple, \"1\", \"acidity\"))); individualAlerts1.get(2).tag(ECOLI_ALERT_TAG, \"well1\").sink(tuple -> System.out.println(formatAlertOutput(tuple, \"1\", \"ecoli\"))); individualAlerts1.get(3).tag(LEAD_ALERT_TAG, \"well1\").sink(tuple -> System.out.println(formatAlertOutput(tuple, \"1\", \"lead\"))); ```Output in the terminal window from the `formatAlertOutput` method will look like this:```Well1 alert, temp value is 86Well3
alert, ecoli value is 2Well1 alert, ecoli value is 1Well3 alert, acidity value is 1Well1 alert, lead value is 12Well1 alert, ecoli value is 2Well3 alert, lead value is 10Well3 alert, acidity value is 10```Notice how only those streams that are out of range for the temperature sensor type show output.## Detecting zero tuple countsAt the end of the `ConsoleWaterDetector` application is this snippet of code, added after the topology has been submitted:```javadp.submit(wellTopology);while (true) { MetricRegistry metricRegistry = dp.getServices().getService(MetricRegistry.class); SortedMap counters = metricRegistry.getCounters(); Set> values = counters.entrySet(); for (Entry e : values) { if (e.getValue().getCount() == 0) { System.out.println(\"Counter Op:\" + e.getKey() + \" tuple count: \" + e.getValue().getCount()); } } Thread.sleep(2000);}```What this does is get all the counters in the `MetricRegistry` class and print out the name of the co
unter oplet they are monitoring along with the tuple count if it is zero. Here is some sample output:```Counter Op:TupleCounter.quarks.oplet.JOB_0.OP_44 has a tuple count of zero!Counter Op:TupleCounter.quarks.oplet.JOB_0.OP_45 has a tuple count of zero!Counter Op:TupleCounter.quarks.oplet.JOB_0.OP_46 has a tuple count of zero!Counter Op:TupleCounter.quarks.oplet.JOB_0.OP_47 has a tuple count of zero!Counter Op:TupleCounter.quarks.oplet.JOB_0.OP_89 has a tuple count of zero!Counter Op:TupleCounter.quarks.oplet.JOB_0.OP_95 has a tuple count of zero!Counter Op:TupleCounter.quarks.oplet.JOB_0.OP_96 has a tuple count of zero!Counter Op:TupleCounter.quarks.oplet.JOB_0.OP_97 has a tuple count of zero!Counter Op:TupleCounter.quarks.oplet.JOB_0.OP_98 has a tuple count of zero!```To summarize what the application is doing:* Unions all sensor type readings for a single well* Filters all sensor type readings for a single well, passing on an object that only contains tuples for the object that
have at least one sensor type with out of range values* Splits the object that contained name/value pairs for sensor type and readings into individual sensor types returning only those streams that contain out of range values* Outputs to the command line the well id, sensor type and value that is out of range* Tags are added at various points in the topology for easier identification of either the well or some out of range condition* The topology contains counters to measure tuple counts since `DevelopmentProvider` was used* Individual rate meters were placed on `well1` and `well3`'s temperature sensors to determine the rate of 'unhealthy' values* Prints out the name of the counter oplets whose tuple counts are zero## Topology graph controlsNow that you have an understanding of what the application is doing, let's look at some of the controls in the console, so we can learn how to monitor the application. Below is a screen shot of the top controls: the controls that affect the Topol
ogy Graph.* **Job**: A drop down to select which job is being displayed in the Topology Graph. An application can contain multiple jobs.* **State**: Hovering over the 'State' icon shows information about the selected job. The current and next states of the job, the job id and the job name.* **View by**: This select is used to change how the topology graph is displayed. The three options for this select are: - Static flow - Tuple count - Oplet kind - Currently it is set to 'Static flow'. This means the oplets (represented as circles in the topology graph) do not change size, nor do the lines or links (representing the edges of the topology graph) change width or position. The graph is not being refreshed when it is in 'Static flow' mode.* **Refresh interval**: Allows the user to select an interval between 3 - 20 seconds to refresh the tuple count values in the graph. Every X seconds the metrics for the topology graph are refreshed. More about metrics a little bit later.* **Pause
graph**: Stops the refresh interval timer. Once the 'Pause graph' button is clicked, the user must push 'Resume graph' for the graph to be updated, and then refreshed at the interval set in the 'Refresh interval' timer. It can be helpful to pause the graph if multiple oplets are occupying the same area on the graph, and their names become unreadable. Once the graph is paused, the user can drag an oplet off of another oplet to better view the name and see the edge(s) that connect them.* **Show tags**: If the checkbox appears in the top controls, it means: - The 'View by' layer is capable of displaying stream tags - The topology currently shown in the topology graph has stream tags associated with it* **Show all tags**: Selecting this checkbox shows all the tags present in the topology. If you want to see only certain tags, uncheck this box and select the button labeled 'Select individual tags ...'. A dialog will appear, and you can select one or all of the tags listed in the dialog
which are present in the topology. The next aspect of the console we'll look at are the popups available when selecting 'View all oplet properties', hovering over an oplet and hovering over an edge (link).The screen shot below shows the output from clicking on the 'View all oplet properties' link directly below the job selector:Looking at the sixth line in the table, where the Name is 'OP_5', we can see that the Oplet kind is a `Map`, a `quarks.oplet.functional.Map`, the Tuple count is 0 (this is because the view is in Static flow mode - the graph does not show the number of tuples flowing in it), the source oplet is 'OP_55', the target oplet is 'OP_60', and there are no stream tags coming from the source or target streams. Relationships for all oplets can be viewed in this manner.Now, looking at the graph, if we want to see the relationships for a single oplet, we can hover over it. The image below shows the hover when we are over 'OP_5'.You can also hover over the edges of the
topology graph to get information. Hover over the edge (link) between 'OP_0' and 'OP_55'. The image shows the name and kind of the oplet as the source, and the name and kind of oplet as the target. Again the tuple count is 0 since this is the 'Static flow' view. The last item of information in the tooltip is the tags on the stream.One or many tags can be added to a stream. In this case we see the tags 'temperature' and 'well1'.The section of the code that adds the tags 'temperature' and 'well1' is in the `waterDetector` method of the `ConsoleWaterDetector` class.```javapublic static TStream waterDetector(Topology topology, int wellId) { Random rNum = new Random(); TStream temp = topology.poll(() -> rNum.nextInt(TEMP_RANDOM_HIGH - TEMP_RANDOM_LOW) + TEMP_RANDOM_LOW, 1, TimeUnit.SECONDS); TStream acidity = topology.poll(() -> rNum.nextInt(ACIDITY_RANDOM_HIGH - ACIDITY_RANDOM_LOW) + ACIDITY_RANDOM_LOW, 1, TimeUnit.SECONDS); TStream ecoli = topology.poll(() -> rNum.nextInt(
ECOLI_RANDOM_HIGH - ECOLI_RANDOM_LOW) + ECOLI_RANDOM_LOW, 1, TimeUnit.SECONDS); TStream lead = topology.poll(() -> rNum.nextInt(LEAD_RANDOM_HIGH - LEAD_RANDOM_LOW) + LEAD_RANDOM_LOW, 1, TimeUnit.SECONDS); TStream id = topology.poll(() -> wellId, 1, TimeUnit.SECONDS); // add tags to each sensor temp.tag(\"temperature\", \"well\" + wellId);```### LegendThe legend(s) that appear in the console depend on the view currently displayed. In the static flow mode, if no stream tags are present, there is no legend. In this example we have stream tags in the topology, so the static flow mode gives us the option to select 'Show tags'. If selected, the result is the addition of the stream tags legend:This legend shows all the tags that have been added to the topology, regardless of whether or not 'Show all tags' is checked or specific tags have been selected from the dialog that appears when the 'Select individual tags ...' button is clicked.### Topology graphNow that we've covered mo
st of the ways to modify the view of the topology graph and discussed the application, let's look at the topology graph as a way to understand our application.When analyzing what is happening in your application, here are some ways you might use the console to help you understand it:* Topology of the application - how the edges and vertices of the graph are related* Tuple flow - tuple counts since the application was started* The affect of filters or maps on the downstream streams* Stream tags - if tags are added dynamically based on a condition, where the streams with tags are displayed in the topologyLet's start with the static flow view of the topology. We can look at the graph, and we can also hover over any of the oplets or streams to better understand the connections. Also, we can click 'View all oplet properties' and see the relationships in a tabular format.The other thing to notice in the static flow view are the tags. Look for any colored edges (the links between the oplet
s). All of the left-most oplets have streams with tags. Most of them have the color that corresponds to 'Multiple tags'. If you hover over the edges, you can see the tags. It's obvious that we have tagged each sensor with the sensor type and the well id.Now, if you look to the far right, you can see more tags on streams coming out of a `split` oplet. They also have multiple tags, and hovering over them you can determine that they represent out of range values for each sensor type for the well. Notice how the `split` oplet, OP_43, has no tags in the streams coming out of it. If you follow that split oplet back, you can determine from the first tags that it is part of the well 2 stream.If you refer back to the `ConsoleWaterDetector` source, you can see that no tags were placed on the streams coming out of `well2`'s split because they contained no out of range values.Let's switch the view to Oplet kind now. It will make more clear which oplets are producing the streams with the tags on
them. Below is an image of how the graph looks after switching to the Oplet kind view.In the Oplet kind view the links are all the same width, but the circles representing the oplets are sized according to tuple flow. Notice how the circles representing OP_10, OP_32 and OP_21 are large in relation to OP_80, OP_88 and OP_89. As a matter of fact, we can't even see the circle representing OP_89. Looking at OP_35 and then the Oplet kind legend, you can see by the color that it is a Filter oplet. This is because the filter that we used against `well2`, which is the stream that OP_35 is part of returned no tuples. This is a bit difficult to see. Let's look at the Tuple count view.The Tuple count view will make it more clear that no tuples are following out of OP_35, which represents the filter for `well2` and only returns out of range values. You may recall that in this example `well2` returned no out of range values. Below is the screen shot of the graph in 'Tuple count' view mode.The t
opology graph oplets can sometimes sit on top of each other. If this is the case, pause the refresh and use your mouse to pull down on the oplets that are in the same position. This will allow you to see their name. Alternately, you can use the 'View all properties' table to see the relationships between oplets.### MetricsIf you scroll the browser window down, you can see a Metrics section. This section appears when the application contains the following:* A `DevelopmentProvider` is used; this automatically inserts counters on the streams of the topology* A `quarks.metrics.Metric.Counter` or `quarks.metrics.Metric.RateMeter` is added to an individual stream## CountersIn the `ConsoleWaterDetector` application we used a `DevelopmentProvider`. Therefore, counters were added to most streams (edges) with the following exceptions (from the [Javadoc]({{ site.docsurl }}/lastest/quarks/metrics/Metrics.html#counter-quarks.topology.TStream-) for `quarks.metrics.Metrics`):*Oplets are only inser
ted upstream from a FanOut oplet.**If a chain of Peek oplets exists between oplets A and B, a Metric oplet is inserted after the last Peek, right upstream from oplet B.**If a chain of Peek oplets is followed by a FanOut, a metric oplet is inserted between the last Peek and the FanOut oplet.The implementation is not idempotent; previously inserted metric oplets are treated as regular graph vertices. Calling the method twice will insert a new set of metric oplets into the graph.*Also, the application inserts counters on `well2`'s streams after the streams from the individual sensors were unioned and then split:```javaList> individualAlerts2 = splitAlert(filteredReadings2, 2);TStream alert0Well2 = individualAlerts2.get(0);alert0Well2 = Metrics.counter(alert0Well2);alert0Well2.tag(\"well2\", \"temp\");TStream alert1Well2 = individualAlerts2.get(1);alert1Well2 = Metrics.counter(alert1Well2);alert1Well2.tag(\"well2\", \"acidity\");TStream alert2Well2 = individualAlerts2.get(2);alert2Wel
l2 = Metrics.counter(alert2Well2);alert2Well2.tag(\"well2\", \"ecoli\");TStream alert3Well2 = individualAlerts2.get(3);alert3Well2 = Metrics.counter(alert3Well2);alert3Well2.tag(\"well2\", \"lead\");```When looking at the select next to the label 'Metrics', make sure the 'Count, oplets OP_37, OP_49 ...' is selected. This select compares all of the counters in the topology visualized as a bar graph. An image is shown below:Hover over individual bars to get the value of the number of tuples flowing through that oplet since the application was started. You can also see the oplet name. You can see that some of the oplets have zero tuples flowing through them.The bars that are the tallest and therefore have the highest tuple count are OP_76, OP_67 and OP_65. If you look back up to the topology graph, in the Tuple count view, you can see that the edges (streams) surrounding these oplets have the color that corresponds to the highest tuple count (in the pictures above that color is br
ight orange in the Tuple count legend).### Rate metersThe other type of metric we can look at are rate meter metrics. In the `ConsoleWaterDetector` application we added two rate meters here with the objective of comparing the rate of out of range readings between `well1` and `well3`:```List> individualAlerts1 = splitAlert(filteredReadings1, 1);// Put a rate meter on well1's temperature sensor outputMetrics.rateMeter(individualAlerts1.get(0));...List> individualAlerts3 = splitAlert(filteredReadings3, 3);// Put a rate meter on well3's temperature sensor outputMetrics.rateMeter(individualAlerts3.get(0));```Rate meters contain the following metrics for each stream they are added to: * Tuple count * The rate of change in the tuple count. The following rates are available for a single stream: - 1 minute rate change - 5 minute rate change - 15 minute rate change - Mean rate changeNow change the Metrics select to the 'MeanRate'. In our example these correspond to oplets OP_37
and OP_49:Hovering over the slightly larger bar, the one to the right, the name is OP_49. Looking at the topology graph and changing the view to 'Static flow', follow the edges back from OP_49 until you can see an edge with a tag on it. You can see that OP_49's source is OP_51, whose source is OP_99. The edge between OP_99 and it's source OP_48 has multiple tags. Hovering over this stream, the tags are 'TEMP out of range' and 'well3'.If a single rate meter is placed on a stream, in addition to plotting a bar chart, a line chart over the last 20 measures can be viewed. For example, if I comment out the addition of the rate meter for `well1` and then rerun the application, the Metrics section will look like the image below. I selected the 'OneMinuteRate' and 'Line chart' for Chart type:## SummaryThe intent of the information on this page is to help you understand the following:* How to add the console application to a Quarks application* How to run the `ConsoleWaterDetector` sample*
The design/architecture in the `ConsoleWaterDetector` application* The controls for the Topology graph are and what they do, including the different views of the graph* The legend for the graph* How to interpret the graph and use the tooltips over the edges and vertices, as well as the 'View all properties' link* How to add counters and rate meters to a topology* How to use the metrics section to understand tuple counters and rate meters* How to correlate values from the metrics section with the topology graphThe Quarks console will continue to evolve and improve. Please open an issue if you see a problem with the existing console, but more importantly add an issue if you have an idea of how to make the console better.The more folks write Quarks applications and view them in the console, the more information we can gather from the community about what is needed in the console. Please consider making a contribution if there is a feature in the console that would really help you and o
thers!"
+"body": "## Visualizing and monitoring your applicationThe Edgent application console is a web application that enables you to visualize your application topology and monitor the tuples flowing through your application. The kind of oplets used in the topology, as well as the stream tags included in the topology, are also visible in the console.## Adding the console web app to your applicationTo use the console, you must use the Edgent classes that provide the service to access the console web application or directly call the `HttpServer` class itself, start the server and then obtain the console URL.The easiest way to include the console in your application is to use the the `DevelopmentProvider` class. `DevelopmentProvider` is a subclass of `DirectProvider` and adds services such as access to the console web application and counter oplets used to determine tuple counts. You can get the URL for the console from the `DevelopmentProvider` using the `getService` method as shown in a hy
pothetical application shown below:```javaimport java.util.concurrent.TimeUnit;import org.apache.edgent.console.server.HttpServer;import org.apache.edgent.providers.development.DevelopmentProvider;import org.apache.edgent.topology.TStream;import org.apache.edgent.topology.Topology;public class TempSensorApplication { public static void main(String[] args) throws Exception { TempSensor sensor = new TempSensor(); DevelopmentProvider dp = new DevelopmentProvider(); Topology topology = dp.newTopology(); TStream tempReadings = topology.poll(sensor, 1, TimeUnit.MILLISECONDS); TStream filteredReadings = tempReadings.filter(reading -> reading 80); filteredReadings.print(); System.out.println(dp.getServices().getService(HttpServer.class).getConsoleUrl()); dp.submit(topology); }}```Note that the console URL is being printed to `System.out`. The `filteredReadings` are as well, since `filteredReadings.print()` is being called in the
application. You may need to scroll your terminal window up to see the output for the console URL.Optionally, you can modify the above code in the application to have a timeout before submitting the topology, which would allow you to see the console URL before any other output is shown. The modification would look like this:```java// Print the console URL and wait for 10 seconds before submitting the topologySystem.out.println(dp.getServices().getService(HttpServer.class).getConsoleUrl());try { TimeUnit.SECONDS.sleep(10);} catch (InterruptedException e) { // Do nothing}dp.submit(topology);```The other way to embed the console in your application is shown in the `HttpServerSample.java` example (on [GitHub]({{ site.data.project.source_repository_mirror }}/blob/master/samples/console/src/main/java/{{ site.data.project.unix_name }}/samples/console/HttpServerSample.java)). It gets the `HttpServer` instance, starts it, and prints out the console URL. Note that it does not submit a
job, so when the console is displayed in the browser, there are no running jobs and therefore no topology graph. The example is meant to show how to get the `HttpServer` instance, start the console web app and get the URL of the console.## Accessing the consoleThe console URL has the following format:`http://host_name:port_number/console`Once it is obtained from `System.out`, enter it in a browser window.If you cannot access the console at this URL, ensure there is a `console.war` file in the `webapps` directory. If the `console.war` file cannot be found, an exception will be thrown (in `std.out`) indicating `console.war` was not found.## ConsoleWaterDetector sampleTo see the features of the console in action and as a way to demonstrate how to monitor a topology in the console, let's look at the `ConsoleWaterDetector` sample (on [GitHub]({{ site.data.project.source_repository_mirror }}/blob/master/samples/console/src/main/java/{{ site.data.project.unix_name }}/samples/console/Consol
eWaterDetector.java)).Prior to running any console applications, the `console.war` file must be built as mentioned above. If you are building Edgent from a Git repository, go to the top level Edgent directory and run `ant`.Here is an example in my environment:```Susans-MacBook-Pro-247:edgent susancline$ pwd/Users/susancline/git/edgentSusans-MacBook-Pro-247:edgent susancline$ antBuildfile: /Users/susancline/git/edgent/build.xmlsetcommitversion:init:suball:init:project.component:compile:...[javadoc] Constructing Javadoc information...[javadoc] Standard Doclet version 1.8.0_71[javadoc] Building tree for all the packages and classes...[javadoc] Generating /Users/susancline/git/edgent/target/docs/javadoc/edgent/analytics/sensors/package-summary.html...[javadoc] Copying file /Users/susancline/git/edgent/analytics/sensors/src/main/java/edgent/analytics/sensors/doc-files/deadband.png to directory /Users/susancline/git/edgent/target/docs/javadoc/edgent/analytics/sensors/doc-files...[javadoc]
Generating /Users/susancline/git/edgent/target/docs/javadoc/edgent/topology/package-summary.html...[javadoc] Copying file /Users/susancline/git/edgent/api/topology/src/main/java/edgent/topology/doc-files/sources.html to directory /Users/susancline/git/edgent/target/docs/javadoc/edgent/topology/doc-files...[javadoc] Building index for all the packages and classes...[javadoc] Building index for all classes...all:BUILD SUCCESSFULTotal time: 3 seconds```This command will let you know that `console.war` was built and is in the correct place, under the `webapps` directory.```Susans-MacBook-Pro-247:edgent susancline$ find . -name console.war -print./target/java8/console/webapps/console.war```Now we know we have built `console.war`, so we're good to go. To run this sample from the command line:```Susans-MacBook-Pro-247:edgent susancline$ pwd/Users/susancline/git/edgentSusans-MacBook-Pro-247:edgent susancline$ java -cp target/java8/samples/lib/edgent.samples.console.jar:. edgent.samples.con
sole.ConsoleWaterDetector```If everything is successful, you'll start seeing output. You may have to scroll back up to get the URL of the console:```Susans-MacBook-Pro-247:edgent susancline$ java -cp target/java8/samples/lib/edgent.samples.console.jar:. edgent.samples.console.ConsoleWaterDetectorMar 07, 2016 12:04:52 PM org.eclipse.jetty.util.log.Log initializedINFO: Logging initialized @176msMar 07, 2016 12:04:53 PM org.eclipse.jetty.server.Server doStartINFO: jetty-9.3.6.v20151106Mar 07, 2016 12:04:53 PM org.eclipse.jetty.server.handler.ContextHandler doStartINFO: Started o.e.j.s.ServletContextHandler@614c5515{/jobs,null,AVAILABLE}Mar 07, 2016 12:04:53 PM org.eclipse.jetty.server.handler.ContextHandler doStartINFO: Started o.e.j.s.ServletContextHandler@77b52d12{/metrics,null,AVAILABLE}Mar 07, 2016 12:04:53 PM org.eclipse.jetty.webapp.StandardDescriptorProcessor visitServletINFO: NO JSP Support for /console, did not find org.eclipse.jetty.jsp.JettyJspServletMar 07, 2016 12:04:53 PM
org.eclipse.jetty.server.handler.ContextHandler doStartINFO: Started o.e.j.w.WebAppContext@2d554825{/console,file:///private/var/folders/0c/pb4rznhj7sbc886t30w4vpxh0000gn/T/jetty-0.0.0.0-0-console.war-_console-any-3101338829524954950.dir/webapp/,AVAILABLE}{/console.war}Mar 07, 2016 12:04:53 PM org.eclipse.jetty.server.AbstractConnector doStartINFO: Started ServerConnector@66480dd7{HTTP/1.1,[http/1.1]}{0.0.0.0:57964}Mar 07, 2016 12:04:53 PM org.eclipse.jetty.server.Server doStartINFO: Started @426mshttp://localhost:57964/consoleWell1 alert, ecoli value is 1Well1 alert, temp value is 48Well3 alert, ecoli value is 1```Now point your browser to the URL displayed above in the output from running the Java command to launch the `ConsoleWaterDetector` application. In this case, the URL is `http://localhost:57964/console`.Below is a screen shot of what you should see if everything is working properly:## ConsoleWaterDetector application scenarioThe application is now running in your browser.
Let's discuss the scenario for the application.A county agency is responsible for ensuring the safety of residents well water. Each well they monitor has four different sensor types:* Temperature* Acidity* Ecoli* LeadThe sample application topology monitors 3 wells:* For the hypothetical scenario, Well1 and Well3 produce 'unhealthy' values from their sensors on occasion. Well2 always produces 'healthy' values.* Each well that is to be measured is added to the topology. The topology polls each sensor (temp, ecoli, etc.) for each well as a unit. A `TStream` is returned from polling the toplogy and represents a sensor reading. Each sensor reading for the well has a tag added to it with the reading type i.e, \"temp\", and the well id. Once all of the sensor readings are obtained and the tags added, each sensor reading is 'unioned' into a single `TStream`. Look at the `waterDetector` method for details on this.* Now, each well has a single stream with each of the sensors readings as a p
roperty with a name and value in the `TStream`. Next the `alertFilter` method is called on the `TStream` representing each well. This method checks the values for each well's sensors to determine if they are 'out of range' for healthy values. The `filter` oplet is used to do this. If any of the sensor's readings are out of the acceptable range the tuple is passed along. Those that are within an acceptable range are discarded.* Next the applications' `splitAlert` method is called on each well's stream that contains the union of all the sensor readings that are out of range. The `splitAlert` method uses the `split` oplet to split the incoming stream into 5 different streams. Only those tuples that are out of range for each stream, which represents each sensor type, will be returned. The object returned from `splitAlert` is a list of `TStream` objects. The `splitAlert` method is shown below: ```java public static List> splitAlert(TStream alertStream, int wellId) { List> al
lStreams = alertStream.split(5, tuple -> { if (tuple.get(\"temp\") != null) { JsonObject tempObj = new JsonObject(); int temp = tuple.get(\"temp\").getAsInt(); if (temp = TEMP_ALERT_MAX) { tempObj.addProperty(\"temp\", temp); return 0; } else { return -1; } } else if (tuple.get(\"acidity\") != null){ JsonObject acidObj = new JsonObject(); int acid = tuple.get(\"acidity\").getAsInt(); if (acid = ACIDITY_ALERT_MAX) { acidObj.addProperty(\"acidity\", acid); return 1; } else { return -1; } } else if (tuple.get(\"ecoli\") != null) { JsonObject ecoliObj = new JsonObject(); int ecoli = tuple.get(\"ecoli\").getAsInt(); if (ecoli >= ECOLI_ALERT) {
ecoliObj.addProperty(\"ecoli\", ecoli); return 2; } else { return -1; } } else if (tuple.get(\"lead\") != null) { JsonObject leadObj = new JsonObject(); int lead = tuple.get(\"lead\").getAsInt(); if (lead >= LEAD_ALERT_MAX) { leadObj.addProperty(\"lead\", lead); return 3; } else { return -1; } } else { return -1; } }); return allStreams; } ```* Next we want to get the temperature stream from the first well and put a rate meter on it to determine the rate at which the out of range values are flowing in the stream ```java List> individualAlerts1 = splitAlert(filteredReadings1, 1); // Put a rate meter on well1's temperature sensor output Metrics.rateMeter(individualAlerts1.get(0)); ```* Next all
the sensors for well 1 have tags added to the stream indicating the stream is out of range for that sensor and the well id. Next a sink is added, passing the tuple to a `Consumer` that formats a string to `System.out` containing the well id, alert type (sensor type) and value of the sensor. ```java // Put a rate meter on well1's temperature sensor output Metrics.rateMeter(individualAlerts1.get(0)); individualAlerts1.get(0).tag(TEMP_ALERT_TAG, \"well1\").sink(tuple -> System.out.println(\"\\n\" + formatAlertOutput(tuple, \"1\", \"temp\"))); individualAlerts1.get(1).tag(ACIDITY_ALERT_TAG, \"well1\").sink(tuple -> System.out.println(formatAlertOutput(tuple, \"1\", \"acidity\"))); individualAlerts1.get(2).tag(ECOLI_ALERT_TAG, \"well1\").sink(tuple -> System.out.println(formatAlertOutput(tuple, \"1\", \"ecoli\"))); individualAlerts1.get(3).tag(LEAD_ALERT_TAG, \"well1\").sink(tuple -> System.out.println(formatAlertOutput(tuple, \"1\", \"lead\"))); ```Output in the
terminal window from the `formatAlertOutput` method will look like this:```Well1 alert, temp value is 86Well3 alert, ecoli value is 2Well1 alert, ecoli value is 1Well3 alert, acidity value is 1Well1 alert, lead value is 12Well1 alert, ecoli value is 2Well3 alert, lead value is 10Well3 alert, acidity value is 10```Notice how only those streams that are out of range for the temperature sensor type show output.## Detecting zero tuple countsAt the end of the `ConsoleWaterDetector` application is this snippet of code, added after the topology has been submitted:```javadp.submit(wellTopology);while (true) { MetricRegistry metricRegistry = dp.getServices().getService(MetricRegistry.class); SortedMap counters = metricRegistry.getCounters(); Set> values = counters.entrySet(); for (Entry e : values) { if (e.getValue().getCount() == 0) { System.out.println(\"Counter Op:\" + e.getKey() + \" tuple count: \" + e.getValue().getCount()); } } Thread.sleep(2
000);}```What this does is get all the counters in the `MetricRegistry` class and print out the name of the counter oplet they are monitoring along with the tuple count if it is zero. Here is some sample output:```Counter Op:TupleCounter.edgent.oplet.JOB_0.OP_44 has a tuple count of zero!Counter Op:TupleCounter.edgent.oplet.JOB_0.OP_45 has a tuple count of zero!Counter Op:TupleCounter.edgent.oplet.JOB_0.OP_46 has a tuple count of zero!Counter Op:TupleCounter.edgent.oplet.JOB_0.OP_47 has a tuple count of zero!Counter Op:TupleCounter.edgent.oplet.JOB_0.OP_89 has a tuple count of zero!Counter Op:TupleCounter.edgent.oplet.JOB_0.OP_95 has a tuple count of zero!Counter Op:TupleCounter.edgent.oplet.JOB_0.OP_96 has a tuple count of zero!Counter Op:TupleCounter.edgent.oplet.JOB_0.OP_97 has a tuple count of zero!Counter Op:TupleCounter.edgent.oplet.JOB_0.OP_98 has a tuple count of zero!```To summarize what the application is doing:* Unions all sensor type readings for a single well* Filters a
ll sensor type readings for a single well, passing on an object that only contains tuples for the object that have at least one sensor type with out of range values* Splits the object that contained name/value pairs for sensor type and readings into individual sensor types returning only those streams that contain out of range values* Outputs to the command line the well id, sensor type and value that is out of range* Tags are added at various points in the topology for easier identification of either the well or some out of range condition* The topology contains counters to measure tuple counts since `DevelopmentProvider` was used* Individual rate meters were placed on `well1` and `well3`'s temperature sensors to determine the rate of 'unhealthy' values* Prints out the name of the counter oplets whose tuple counts are zero## Topology graph controlsNow that you have an understanding of what the application is doing, let's look at some of the controls in the console, so we can learn
how to monitor the application. Below is a screen shot of the top controls: the controls that affect the Topology Graph.* **Job**: A drop down to select which job is being displayed in the Topology Graph. An application can contain multiple jobs.* **State**: Hovering over the 'State' icon shows information about the selected job. The current and next states of the job, the job id and the job name.* **View by**: This select is used to change how the topology graph is displayed. The three options for this select are: - Static flow - Tuple count - Oplet kind - Currently it is set to 'Static flow'. This means the oplets (represented as circles in the topology graph) do not change size, nor do the lines or links (representing the edges of the topology graph) change width or position. The graph is not being refreshed when it is in 'Static flow' mode.* **Refresh interval**: Allows the user to select an interval between 3 - 20 seconds to refresh the tuple count values in the graph. Ever
y X seconds the metrics for the topology graph are refreshed. More about metrics a little bit later.* **Pause graph**: Stops the refresh interval timer. Once the 'Pause graph' button is clicked, the user must push 'Resume graph' for the graph to be updated, and then refreshed at the interval set in the 'Refresh interval' timer. It can be helpful to pause the graph if multiple oplets are occupying the same area on the graph, and their names become unreadable. Once the graph is paused, the user can drag an oplet off of another oplet to better view the name and see the edge(s) that connect them.* **Show tags**: If the checkbox appears in the top controls, it means: - The 'View by' layer is capable of displaying stream tags - The topology currently shown in the topology graph has stream tags associated with it* **Show all tags**: Selecting this checkbox shows all the tags present in the topology. If you want to see only certain tags, uncheck this box and select the button labeled 'Sel
ect individual tags ...'. A dialog will appear, and you can select one or all of the tags listed in the dialog which are present in the topology. The next aspect of the console we'll look at are the popups available when selecting 'View all oplet properties', hovering over an oplet and hovering over an edge (link).The screen shot below shows the output from clicking on the 'View all oplet properties' link directly below the job selector:Looking at the sixth line in the table, where the Name is 'OP_5', we can see that the Oplet kind is a `Map`, a `edgent.oplet.functional.Map`, the Tuple count is 0 (this is because the view is in Static flow mode - the graph does not show the number of tuples flowing in it), the source oplet is 'OP_55', the target oplet is 'OP_60', and there are no stream tags coming from the source or target streams. Relationships for all oplets can be viewed in this manner.Now, looking at the graph, if we want to see the relationships for a single oplet, we can h
over over it. The image below shows the hover when we are over 'OP_5'.You can also hover over the edges of the topology graph to get information. Hover over the edge (link) between 'OP_0' and 'OP_55'. The image shows the name and kind of the oplet as the source, and the name and kind of oplet as the target. Again the tuple count is 0 since this is the 'Static flow' view. The last item of information in the tooltip is the tags on the stream.One or many tags can be added to a stream. In this case we see the tags 'temperature' and 'well1'.The section of the code that adds the tags 'temperature' and 'well1' is in the `waterDetector` method of the `ConsoleWaterDetector` class.```javapublic static TStream waterDetector(Topology topology, int wellId) { Random rNum = new Random(); TStream temp = topology.poll(() -> rNum.nextInt(TEMP_RANDOM_HIGH - TEMP_RANDOM_LOW) + TEMP_RANDOM_LOW, 1, TimeUnit.SECONDS); TStream acidity = topology.poll(() -> rNum.nextInt(ACIDITY_RANDOM_HIGH - ACIDIT
Y_RANDOM_LOW) + ACIDITY_RANDOM_LOW, 1, TimeUnit.SECONDS); TStream ecoli = topology.poll(() -> rNum.nextInt(ECOLI_RANDOM_HIGH - ECOLI_RANDOM_LOW) + ECOLI_RANDOM_LOW, 1, TimeUnit.SECONDS); TStream lead = topology.poll(() -> rNum.nextInt(LEAD_RANDOM_HIGH - LEAD_RANDOM_LOW) + LEAD_RANDOM_LOW, 1, TimeUnit.SECONDS); TStream id = topology.poll(() -> wellId, 1, TimeUnit.SECONDS); // add tags to each sensor temp.tag(\"temperature\", \"well\" + wellId);```### LegendThe legend(s) that appear in the console depend on the view currently displayed. In the static flow mode, if no stream tags are present, there is no legend. In this example we have stream tags in the topology, so the static flow mode gives us the option to select 'Show tags'. If selected, the result is the addition of the stream tags legend:This legend shows all the tags that have been added to the topology, regardless of whether or not 'Show all tags' is checked or specific tags have been selected from the dialog th
at appears when the 'Select individual tags ...' button is clicked.### Topology graphNow that we've covered most of the ways to modify the view of the topology graph and discussed the application, let's look at the topology graph as a way to understand our application.When analyzing what is happening in your application, here are some ways you might use the console to help you understand it:* Topology of the application - how the edges and vertices of the graph are related* Tuple flow - tuple counts since the application was started* The affect of filters or maps on the downstream streams* Stream tags - if tags are added dynamically based on a condition, where the streams with tags are displayed in the topologyLet's start with the static flow view of the topology. We can look at the graph, and we can also hover over any of the oplets or streams to better understand the connections. Also, we can click 'View all oplet properties' and see the relationships in a tabular format.The other
thing to notice in the static flow view are the tags. Look for any colored edges (the links between the oplets). All of the left-most oplets have streams with tags. Most of them have the color that corresponds to 'Multiple tags'. If you hover over the edges, you can see the tags. It's obvious that we have tagged each sensor with the sensor type and the well id.Now, if you look to the far right, you can see more tags on streams coming out of a `split` oplet. They also have multiple tags, and hovering over them you can determine that they represent out of range values for each sensor type for the well. Notice how the `split` oplet, OP_43, has no tags in the streams coming out of it. If you follow that split oplet back, you can determine from the first tags that it is part of the well 2 stream.If you refer back to the `ConsoleWaterDetector` source, you can see that no tags were placed on the streams coming out of `well2`'s split because they contained no out of range values.Let's swit
ch the view to Oplet kind now. It will make more clear which oplets are producing the streams with the tags on them. Below is an image of how the graph looks after switching to the Oplet kind view.In the Oplet kind view the links are all the same width, but the circles representing the oplets are sized according to tuple flow. Notice how the circles representing OP_10, OP_32 and OP_21 are large in relation to OP_80, OP_88 and OP_89. As a matter of fact, we can't even see the circle representing OP_89. Looking at OP_35 and then the Oplet kind legend, you can see by the color that it is a Filter oplet. This is because the filter that we used against `well2`, which is the stream that OP_35 is part of returned no tuples. This is a bit difficult to see. Let's look at the Tuple count view.The Tuple count view will make it more clear that no tuples are following out of OP_35, which represents the filter for `well2` and only returns out of range values. You may recall that in this example `
well2` returned no out of range values. Below is the screen shot of the graph in 'Tuple count' view mode.The topology graph oplets can sometimes sit on top of each other. If this is the case, pause the refresh and use your mouse to pull down on the oplets that are in the same position. This will allow you to see their name. Alternately, you can use the 'View all properties' table to see the relationships between oplets.### MetricsIf you scroll the browser window down, you can see a Metrics section. This section appears when the application contains the following:* A `DevelopmentProvider` is used; this automatically inserts counters on the streams of the topology* A `edgent.metrics.Metric.Counter` or `edgent.metrics.Metric.RateMeter` is added to an individual stream## CountersIn the `ConsoleWaterDetector` application we used a `DevelopmentProvider`. Therefore, counters were added to most streams (edges) with the following exceptions (from the [Javadoc]({{ site.docsurl }}/latest/{{ si
te.data.project.unix_name }}/metrics/Metrics.html#counter-{{ site.data.project.unix_name }}.topology.TStream-) for `edgent.metrics.Metrics`):*Oplets are only inserted upstream from a FanOut oplet.**If a chain of Peek oplets exists between oplets A and B, a Metric oplet is inserted after the last Peek, right upstream from oplet B.**If a chain of Peek oplets is followed by a FanOut, a metric oplet is inserted between the last Peek and the FanOut oplet.The implementation is not idempotent; previously inserted metric oplets are treated as regular graph vertices. Calling the method twice will insert a new set of metric oplets into the graph.*Also, the application inserts counters on `well2`'s streams after the streams from the individual sensors were unioned and then split:```javaList> individualAlerts2 = splitAlert(filteredReadings2, 2);TStream alert0Well2 = individualAlerts2.get(0);alert0Well2 = Metrics.counter(alert0Well2);alert0Well2.tag(\"well2\", \"temp\");TStream alert1Well2 = in
dividualAlerts2.get(1);alert1Well2 = Metrics.counter(alert1Well2);alert1Well2.tag(\"well2\", \"acidity\");TStream alert2Well2 = individualAlerts2.get(2);alert2Well2 = Metrics.counter(alert2Well2);alert2Well2.tag(\"well2\", \"ecoli\");TStream alert3Well2 = individualAlerts2.get(3);alert3Well2 = Metrics.counter(alert3Well2);alert3Well2.tag(\"well2\", \"lead\");```When looking at the select next to the label 'Metrics', make sure the 'Count, oplets OP_37, OP_49 ...' is selected. This select compares all of the counters in the topology visualized as a bar graph. An image is shown below:Hover over individual bars to get the value of the number of tuples flowing through that oplet since the application was started. You can also see the oplet name. You can see that some of the oplets have zero tuples flowing through them.The bars that are the tallest and therefore have the highest tuple count are OP_76, OP_67 and OP_65. If you look back up to the topology graph, in the Tuple count vie
w, you can see that the edges (streams) surrounding these oplets have the color that corresponds to the highest tuple count (in the pictures above that color is bright orange in the Tuple count legend).### Rate metersThe other type of metric we can look at are rate meter metrics. In the `ConsoleWaterDetector` application we added two rate meters here with the objective of comparing the rate of out of range readings between `well1` and `well3`:```List> individualAlerts1 = splitAlert(filteredReadings1, 1);// Put a rate meter on well1's temperature sensor outputMetrics.rateMeter(individualAlerts1.get(0));...List> individualAlerts3 = splitAlert(filteredReadings3, 3);// Put a rate meter on well3's temperature sensor outputMetrics.rateMeter(individualAlerts3.get(0));```Rate meters contain the following metrics for each stream they are added to: * Tuple count * The rate of change in the tuple count. The following rates are available for a single stream: - 1 minute rate change - 5 m
inute rate change - 15 minute rate change - Mean rate changeNow change the Metrics select to the 'MeanRate'. In our example these correspond to oplets OP_37 and OP_49:Hovering over the slightly larger bar, the one to the right, the name is OP_49. Looking at the topology graph and changing the view to 'Static flow', follow the edges back from OP_49 until you can see an edge with a tag on it. You can see that OP_49's source is OP_51, whose source is OP_99. The edge between OP_99 and it's source OP_48 has multiple tags. Hovering over this stream, the tags are 'TEMP out of range' and 'well3'.If a single rate meter is placed on a stream, in addition to plotting a bar chart, a line chart over the last 20 measures can be viewed. For example, if I comment out the addition of the rate meter for `well1` and then rerun the application, the Metrics section will look like the image below. I selected the 'OneMinuteRate' and 'Line chart' for Chart type:## SummaryThe intent of the informatio
n on this page is to help you understand the following:* How to add the console application to an Edgent application* How to run the `ConsoleWaterDetector` sample* The design/architecture in the `ConsoleWaterDetector` application* The controls for the Topology graph are and what they do, including the different views of the graph* The legend for the graph* How to interpret the graph and use the tooltips over the edges and vertices, as well as the 'View all properties' link* How to add counters and rate meters to a topology* How to use the metrics section to understand tuple counters and rate meters* How to correlate values from the metrics section with the topology graphThe Edgent console will continue to evolve and improve. Please open an issue if you see a problem with the existing console, but more importantly add an issue if you have an idea of how to make the console better.The more folks write Edgent applications and view them in the console, the more information we can gather
from the community about what is needed in the console. Please consider making a contribution if there is a feature in the console that would really help you and others!"
},
@@ -60,12 +60,12 @@
{
-"title": "FAQ",
+"title": "Getting started with Apache Edgent",
"tags": "",
"keywords": "",
-"url": "../docs/faq",
+"url": "../docs/edgent-getting-started",
"summary": "",
-"body": "## What is Apache Quarks?Quarks provides APIs and a lightweight runtime to analyze streaming data at the edge.## What do you mean by the edge?The edge includes devices, gateways, equipment, vehicles, systems, appliances and sensors of all kinds as part of the Internet of Things.## How is Apache Quarks used?Quarks can be used at the edge of the Internet of Things, for example, to analyze data on devices, engines, connected cars, etc. Quarks could be on the device itself, or a gateway device collecting data from local devices. You can write an edge application on Quarks and connect it to a Cloud service, such as the IBM Watson IoT Platform. It can also be used for enterprise data collection and analysis; for example log collectors, application data, and data center analytics.## How are applications developed?Applications are developed using a functional flow API to define operations on data streams that are executed as a graph of \"oplets\" in a lightweight embeddable runtime
. The SDK provides capabilities like windowing, aggregation and connectors with an extensible model for the community to expand its capabilities.## What APIs does Apache Quarks support?Currently, Quarks supports APIs for Java and Android. Support for additional languages, such as Python, is likely as more developers get involved. Please consider joining the Quarks open source development community to accelerate the contributions of additional APIs.## What type of analytics can be done with Apache Quarks?Quarks provides windowing, aggregation and simple filtering. It uses Apache Common Math to provide simple analytics aimed at device sensors. Quarks is also extensible, so you can call existing libraries from within your Quarks application. In the future, Quarks will include more analytics, either exposing more functionality from Apache Common Math, other libraries or hand-coded analytics.## What connectors does Apache Quarks support?Quarks supports connectors for MQTT, HTTP, JDBC, Fi
le, Apache Kafka and IBM Watson IoT Platform. Quarks is extensible; you can add the connector of your choice.## What centralized streaming analytic systems does Apache Quarks support?Quarks supports open source technology (such as Apache Spark, Apache Storm, Flink and samza), IBM Streams (on-premises or IBM Streaming Analytics on Bluemix), or any custom application of your choice.## Why do I need Apache Quarks on the edge, rather than my streaming analytic system?Quarks is designed for the edge, rather than a more centralized system. It has a small footprint, suitable for running on devices. Quarks provides simple analytics, allowing a device to analyze data locally and to only send to the centralized system if there is a need, reducing communication costs.## Why do I need Apache Quarks, rather than coding the complete application myself?Quarks is a tool for edge analytics that allows you to be more productive. Quarks provides a consistent data model (streams and windows) and provid
es useful functionality, such as aggregations, joins, etc. Using Quarks lets you to take advantage of this functionality, allowing you to focus on your application needs.## Where can I download Apache Quarks to try it out?Quarks is migrating from github quarks-edge to Apache. You can download the source from Apache and build it yourself [here](https://github.com/apache/incubator-quarks). You can also find already built pre-Apache releases of Quarks for download [here](https://github.com/quarks-edge/quarks/releases/latest). These releases are not associated with Apache.## How do I get started?Getting started is simple. Once you have downloaded Quarks, everything you need to know to get up and running, you will find [here](quarks-getting-started). We suggest you also run the [Quarks sample programs](samples) to familiarize yourselves with the code base.## How can I get involved?We would love to have your help! Visit [Get Involved](community) to learn more about how to get involved.##
How can I contribute code?Just submit a [pull request](https://github.com/apache/incubator-quarks) and wait for a committer to review. For more information, visit our [committer page](committers) and read [DEVELOPMENT.md](https://github.com/apache/incubator-quarks/blob/master/DEVELOPMENT.md) at the top of the code tree.## Can I become a committer?Read about Quarks committers and how to become a committer [here](committers).## Where can I get the code?The source code is available [here](https://github.com/apache/incubator-quarks).## Can I take a copy of the code and fork it for my own use?Yes. Quarks is available under the Apache 2.0 license which allows you to fork the code. We hope you will contribute your changes back to the Quarks community.## How do I suggest new features?Click [Issues](https://issues.apache.org/jira/browse/QUARKS) to submit requests for new features. You may browse or query the Issues database to see what other members of the Quarks community have already reque
sted.## How do I submit bug reports?Click [Issues](https://issues.apache.org/jira/browse/QUARKS) to submit a bug report.## How do I ask questions about Apache Quarks?Use [site.data.project.user_list](mailto:{{ site.data.project.user_list }}) to submit questions to the Quarks community.## Why is Apache Quarks open source?With the growth of the Internet of Things there is a need to execute analytics at the edge. Quarks was developed to address requirements for analytics at the edge for IoT use cases that were not addressed by central analytic solutions. These capabilities will be useful to many organizations and that the diverse nature of edge devices and use cases is best addressed by an open community. Our goal is to develop a vibrant community of developers and users to expand the capabilities and real-world use of Quarks by companies and individuals to enable edge analytics and further innovation for the IoT space."
+"body": "## What is Apache Edgent?Edgent is an open source programming model and runtime for edge devices that enables you to analyze streaming data on your edge devices. When you analyze on the edge, you can:* Reduce the amount of data that you transmit to your analytics server* Reduce the amount of data that you storeFor more information, see the [Edgent overview](home).### Apache Edgent and streaming analyticsThe fundamental building block of an Edgent application is a **stream**: a continuous sequence of tuples (messages, events, sensor readings, and so on).The Edgent API provides the ability to process or analyze each tuple as it appears on a stream, resulting in a derived stream.Source streams are streams that originate data for analysis, such as readings from a device's temperature sensor.Streams are terminated using sink functions that can perform local device control or send information to centralized analytic systems through a message hub.Edgent's primary API is functional
where streams are sourced, transformed, analyzed or sinked though functions, typically represented as lambda expressions, such as `reading -> reading 80` to filter temperature readings in Fahrenheit.### Downloading Apache EdgentTo use Edgent, access the source code and build it. You can read more about building Edgent [here]({{ site.data.project.source_repository_mirror }}/blob/master/DEVELOPMENT.md).After you build the Edgent package, you can set up your environment.### Setting up your environmentEnsure that you are running a supported environment. For more information, see the [Edgent overview](home). This guide assumes you're running Java 8.The Edgent Java 8 JAR files are located in the `edgent/java8/lib` directory.1. Create a new Java project in Eclipse, and specify Java 8 as the execution environment JRE: 2. Modify the Java build path to include all of the JAR files in the `edgent\\java8\\lib` directory: Your environment is set up! You can start writing your first Edgen
t application.## Creating a simple applicationIf you're new to Edgent or to writing streaming applications, the best way to get started is to write a simple program.Edgent is a framework that pushes data analytics and machine learning to *edge devices*. (Edge devices include things like routers, gateways, machines, equipment, sensors, appliances, or vehicles that are connected to a network.) Edgent enables you to process data locally—such as, in a car engine, on an Android phone, or Raspberry Pi—before you send data over a network.For example, if your device takes temperature readings from a sensor 1,000 times per second, it is more efficient to process th
<TRUNCATED>