Dear forum
In a PoC we successfully configured SN Predictive Intelligence to provide Field Recommendations to Incidents (Service, Assignment Group etc.). But now the decision was taken to integrate the customers AI Engine via REST and to not use OOB Predictive Intelligence.
We are now running a new PoC but were so far not able to integrate a scripted resource generator.
On a Recommendation (sn_nb_action_recommended_action) there is the possibility to reference a custom Resource generator (sn_nb_action_resource_generator) of type "Scripting". There are a few example OOB Resource generators but they seem all to be used for "Similarity (Action Type: Guidance (sn_nb_action_type_definition))" and not for "Categorization (Action Type: Field recommendation (sn_nb_action_type_definition)" .
We tried nevertheless to create a script which is based on the OOB Examples but did not succeed so far.
There are no recommendations shown and the log shows the following error:
*** Script [ScriptingGeneratorTypeHandler]: Script include 13e4e42187b5ee1c2f880f230cbb355d is invalid: no thrown error
Apparently the Script Include is not returning a value as expected. We were not able to find any documentation about it.
The script is for the time being only trying to recommend a single service (hard coded) - just to verify if it is possible to script our own recommendations. The final script would invoke a REST call to retrieve the results.
The code is as follows:
var ai_incident_suggestServiceGenerator = Class.create();
ai_incident_suggestServiceGenerator.prototype = Object.extendsObject(sn_nb_action.ScriptingGeneratorHandlerBase, {
category: 'ra_scripting_generator', // DO NOT REMOVE THIS LINE!
getId: function() {
return "sn_sow_inc.ai_incident_suggestServiceGenerator";
},
getOutputSchema: function() {
var schema = [{
'name': 'predictedRecord',
'label': 'Predicted Record',
'type': 'reference',
'referenceTable': 'cmdb_ci_service'
}];
return {
'status': 'success',
'schema': schema,
};
},
getOutputs: function(param) {
try {
//var nonMLRecommendationsHelper = new sn_sow_inc.NonMLRecommendationsHelper();
//var similarOpenIncidentsSysIds = nonMLRecommendationsHelper.getSimilarOpenIncidentsWithFilters(param.contextRecord);
var response = {
'status': 'success',
outputs: []
};
response.outputs.push({predictedRecord: "3ca79da0ad24b0106d379e95d3892905"}); // SV-JIRA
return response;
} catch (e) {
return {
'status': 'error',
'errorCode': 40001,
'errorMessage': 'Failed to get Non-ML recommendations for Similar Open Problem records' // this will be logged
};
}
},
type: 'ai_incident_suggestServiceGenerator'
});
Any help is appreciated.
Thanks and regards
Alan
