Unable to test Analysis model using thingworx analytics manager

[ahawaldar-2]

Hi,

I am following this tutorial https://support.ptc.com/help/thingworx_hc/thingworx_analytics_52/#page/thingworx_analytics_7%2Fthingpredictor%2Fthingpredictor-deployment.html%23

When I publish the model available in thingworx analytics to thingworx analytics manager, and provide the data in Test and press​ Submit job in thingworx analytics manager, I am getting the below error

Job was not found in runningJobs, could cause processing to stop

Please do let me know if I am missing something here.

Thanks,

Azim

[posipova]

Neel Sampat​ do we have a separate category on the community for analytics?

[cmorfin]

Hi Azim

Do you have the ThingPredictor agent running ?

It is a standalone java application that needs to be started

Could you give the following info too:

- <ThignWorxStorage>/logs/ApplicationLog.log

-<thingpredictor>/logs/ folder

It woudl be also a good idea to try the demo scenario at PTC

this demo provide a pmml file for a machine learning model that you can upload so you can make sure that the setup is ok before starting using your own model.

Hope this helps

Regrards

Christophe

[ahawaldar-2]

Hi Christophe Morfin​,

ThingPredictor Agent is running. Below are the logs generated when I hit submit button in thingworx analytics manager in thingworx.

2017-09-22 10:46:19.556+0530 [L: ERROR] [O: c.p.a.c.s.BasicSDKConnector] [T: WSExecutionProcessor-11]

                 Job was not found in runningJobs, could cause processing to stop

2017-09-22 10:46:21.579+0530 [L: ERROR] [O: c.p.a.c.s.BasicSDKConnector] [T: WSExecutionProcessor-12]

                 Job was not found in runningJobs, could cause processing to stop

2017-09-22 10:47:53.918+0530 [L: ERROR] [O: c.p.a.c.s.BasicSDKConnector] [T: WSExecutionProcessor-13]

                 Job was not found in runningJobs, could cause processing to stop

2017-09-22 10:47:55.972+0530 [L: ERROR] [O: c.p.a.c.s.BasicSDKConnector] [T: WSExecutionProcessor-14]

                 Job was not found in runningJobs, could cause processing to stop

Below are the logs generated in thingworx predictor log folder

INFO  [2017-09-22 01:13:20,128] org.eclipse.jetty.server.handler.ContextHandler: Started i.d.j.MutableServletContextHandler@784d9bc{/,null,AVAILABLE}

INFO  [2017-09-22 01:13:20,138] org.eclipse.jetty.server.ServerConnector: Started application@65eabaab{HTTP/1.1}{0.0.0.0:9090}

INFO  [2017-09-22 01:13:20,141] org.eclipse.jetty.server.ServerConnector: Started admin@7123be6c{HTTP/1.1}{0.0.0.0:9091}

INFO  [2017-09-22 01:13:20,141] org.eclipse.jetty.server.Server: Started @23333ms

INFO  [2017-09-22 01:16:18,617] com.thingworx.analytics.ccc.data.datasources.JsonLocalDatasetSource: Received [1] rows to score

INFO  [2017-09-22 01:16:18,959] com.thingworx.analytics.scoring.core.predictive.PredictiveScoringJob: The model is not a time series model.

INFO  [2017-09-22 01:16:19,239] com.thingworx.analysis.TWAnalysisAgentAdapter: Completed synchronous scoring for [1] rows

INFO  [2017-09-22 01:16:19,239] com.thingworx.analysis.TWAnalysisAgentAdapter: Result of predictive scoring for the request PredictiveScoringRequest [modelUri=file:/D:/software/thingworx/MED-61217-CD-074_F000_ThingWorx-Analytics-ThingPredictor-7-4-0/ThingWorx-Analytics-ThingPredictor/./AgentWorkDir/TW.AnalysisServices.ThingPredictorConnector/Models/bearing_rul_Condition_368.xml, importantFieldCount=3, causalTechnique=FULL_RANGE, dataSource=com.thingworx.analytics.ccc.data.datasources.BodyDatasetRef@4b6bd1e6, goalName=condition] :

[[predictiveScore={condition=0.8478829789902185, condition_Max=0.9459829789902184, condition_Normalized=0.8478829789902185, condition_Result=0.8478829789902185, condition_ConfidenceInterval=0.0981, condition_Min=0.7497829789902185}, importantFields={time=0.9189298025133084, magnitude=0.08107019748669166}, errorMessage=null]]

INFO  [2017-09-22 01:17:53,607] com.thingworx.analytics.ccc.data.datasources.JsonLocalDatasetSource: Received [1] rows to score

INFO  [2017-09-22 01:17:53,607] com.thingworx.analytics.scoring.core.predictive.PredictiveScoringJob: The model is not a time series model.

INFO  [2017-09-22 01:17:53,613] com.thingworx.analysis.TWAnalysisAgentAdapter: Completed synchronous scoring for [1] rows

INFO  [2017-09-22 01:17:53,613] com.thingworx.analysis.TWAnalysisAgentAdapter: Result of predictive scoring for the request PredictiveScoringRequest [modelUri=file:/D:/software/thingworx/MED-61217-CD-074_F000_ThingWorx-Analytics-ThingPredictor-7-4-0/ThingWorx-Analytics-ThingPredictor/./AgentWorkDir/TW.AnalysisServices.ThingPredictorConnector/Models/bearing_rul_Condition_368.xml, importantFieldCount=3, causalTechnique=FULL_RANGE, dataSource=com.thingworx.analytics.ccc.data.datasources.BodyDatasetRef@54835f2, goalName=condition] :

[[predictiveScore={condition=0.8478829789902185, condition_Max=0.9459829789902184, condition_Normalized=0.8478829789902185, condition_Result=0.8478829789902185, condition_ConfidenceInterval=0.0981, condition_Min=0.7497829789902185}, importantFields={time=0.9189298025133084, magnitude=0.08107019748669166}, errorMessage=null]]

INFO  [2017-09-22 01:21:16,648] com.thingworx.analytics.ccc.data.datasources.JsonLocalDatasetSource: Received [1] rows to score

INFO  [2017-09-22 01:21:16,649] com.thingworx.analytics.scoring.core.predictive.PredictiveScoringJob: The model is not a time series model.

INFO  [2017-09-22 01:21:16,654] com.thingworx.analysis.TWAnalysisAgentAdapter: Completed synchronous scoring for [1] rows

INFO  [2017-09-22 01:21:16,655] com.thingworx.analysis.TWAnalysisAgentAdapter: Result of predictive scoring for the request PredictiveScoringRequest [modelUri=file:/D:/software/thingworx/MED-61217-CD-074_F000_ThingWorx-Analytics-ThingPredictor-7-4-0/ThingWorx-Analytics-ThingPredictor/./AgentWorkDir/TW.AnalysisServices.ThingPredictorConnector/Models/bearing_rul_Condition_368.xml, importantFieldCount=3, causalTechnique=FULL_RANGE, dataSource=com.thingworx.analytics.ccc.data.datasources.BodyDatasetRef@60694bd, goalName=condition] :

[[predictiveScore={condition=0.8478829789902185, condition_Max=0.9459829789902184, condition_Normalized=0.8478829789902185, condition_Result=0.8478829789902185, condition_ConfidenceInterval=0.0981, condition_Min=0.7497829789902185}, importantFields={time=0.9189298025133084, magnitude=0.08107019748669166}, errorMessage=null]]

But I do not see the result of Predictive scoring below the Submit job button. I have attached the screen shot of expected(captured from tutorial video) and actual UI. 

Expected

Actual

[cmorfin · Accepted solution]

Hi Azim

From the log it seems that the job did execute and some prediction have been done.

So you might have a UI issue.

You have to note the following:

- the Test UI is not meant to be a fully robust one, it is only there for quick testing.

- in case it does not work you shoudl be able to get some result by populating the property of a Thign which are more what is expected to be needed in a real case scenario. See example at Video Link : 1817  around time 3:58.

- you can also try different browser to check if the display can be better

Kind regards

Chrisotphe

[ahawaldar-2]

Hi Christophe Morfin​,

Thank you, I will go through the above mentioned URL. Just curious to know is there any API which will provide me the model equation which is built by thingworx analytics server by using training data.

Appreciate your help on this.

Thanks,

Azim

[cmorfin]

Hi Azim

There is no API to retrieve the model equation: https://www.ptc.com/en/support/article?n=CS250636

Kind regards

Christophe

[ahawaldar-2]

Hello Christophe Morfin​,

I was going through the video you had suggested earlier i.e. Analytics manager 7.4: Part 2 of 2. Configure Analysis Event, Real-Time Scoring . I am stuck at one point I was not able to find CreateOrUpdateThingTemplateForModel

services under TW.AnalysisServices.ModelManagementServicesAPI resource, though I have this resource TW.AnalysisServices.ModelManagementServicesAPI  imported from extension ThingWorx-Analytics-Extensions-7-4-0.

Please do let me know where is this service available?

Thanks,

Azim

[ahawaldar-2]

Hello Christophe Morfin​,

Awaiting for your response.

Thanks,

Azim

[cmorfin]

Hi Azim

Note that there is no activity over the week end so if you post something on the Saturday you won't have an aanswer until at least the following Monday.

Also the Community is a place for people to share their experience and issues so that others can profit from it or assist. There is however no guarantee in getting a response. If you need to be sure to get an answer I would advise you open a case to Technical Support.

Regaridng your last point, it is a bit odd, you should have this service.
Could you please access http://<yourServer>/Thingworx/Resources/TW.AnalysisServices.ModelManagementServicesAPI/ServiceDefinitions/

and save the result page in a html file, then post this html file here.

Thank you

Kind rgeards

Christophe

[ahawaldar-2]

Hi Christophe Morfin​,

First of all I want to apologize for asking the response over the weekend. I am sorry again.  I was now able to find the service CreateOrUpdateThingTemplateForModel and I followed the video till end. I was able to create a new Analysis Event, Map the input and Result. Post this navigated to the thing created from service CreateOrUpdateThingTemplateForModel and added some values for "data" property. This made a job to be executed within analysis job. I can see the values inside analysis jobs listing page,  I am able to view the predicted score values in "dataCollection" property. How do fetch this value and show it on any mashup? Is there any service which can help me with this?

Thanks,

Azim

[cmorfin]

Hi Azim

Good to hear it works for you.

I think the followign video should help for showing result in a mashup: Video Link : 1884

Note also that it seems the original issue has been fixed so it woudl be good if you can mark this as answered and, in case  you have further question, open a new thread.

This is to avoid getting very long thread that deals with multiple issue which then make it difficult to follow and lose its ability to help others.

Thank you

Kind regards

Christophe

[ahawaldar-2]

Hello Christophe Morfin​,

To continue on above discussion, I have started a new thread. Please have a look and do let me know whats wrong I have done.

Thanks,

Azim