IBM Watson NLU Text Analysis

UiPath.Cognitive.Activities.Text.Analysis.IbmWatsonNluTextAnalysis

Important:

Extracts key information from a specified text, and returns multiple IBM Watson NLU specific data types as well as a raw result in .json format, stored in a String variable. You can find more information about the Natural Language Understanding service by reading the official documentation here.

Properties

Common

• DisplayName - The display name of the activity.

Input

• Key - The API key used to provide you access to the IBM Watson Text Analysis tool. This field supports String variables.
• Text - The text that is to be analyzed. This field supports strings or String variables.
• ServiceUrl - The custom web address of the local or regional IBM Watson instance used to connect to the text analysis service. If the field remains empty at runtime, the default value for this field is "https://gateway.watsonplatform.net/natural-language-understanding/api/v1/analyze". This field supports only strings and String variables.
• Service Version - The Natural Language Understanding service version that is to be used. The default value for this field is "2017-02-27". This field supports only strings and String variables.
• Password - The password used to login to the Natural Language Understanding service. This field supports only strings and String variables.
• Username - The username used to login to the Natural Language Understanding service. This field supports only strings and String variables.

Misc

• Private - If selected, the values of variables and arguments are no longer logged at Verbose level.

Options

• AnalysisType - Select the type of analysis to be performed on the specified text. The following types are available: Categories, Concepts, Entities, Keywords, and Sentiment. Depending on the analysis types selected, the output properties may differ.

Output

• Categories - The category which the specified text belongs to, stored in a list of <KeyValuePair<String,Double>> variable. Each item in the list contains the category name and the categorization score.
  Note:

  For more info on the categories hierarchy, view the official IBM documentation here.

• Concepts - The concepts extracted from the specified text, stored in a list of <KeyValuePair<String,Double>> variables. Each item in the list contains the concept and the concept score.
• Entities - The entities extracted from the specified text, stored in a list of IBMEntity variables. The IBMEntity variable contains the following information for each entity:
  • Type - The type of the entity.
  • Sentiment - The sentiment score and label of the entity.
  • Relevance - The relevance of the entity in the specified text.
  • Emotion - The scores for all the defined emotions.
  • Disambiguation - Other concepts that are present in the specified text and are linked to the entity.
  • Count - The number of times the entity is mentioned in the specified text.

Here is an example of a .json text that shows the IBMEntity information and structure:

{  
         "type":"Location",
         "text":"Europe",
         "sentiment":{  
            "score":0.0,
            "label":"neutral"
         },
         "relevance":0.677612,
         "emotion":{  
            "sadness":0.336228,
            "joy":0.044658,
            "fear":0.075551,
            "disgust":0.355227,
            "anger":0.076605
         },
         "disambiguation":{  
            "subtype":[  
               "MusicalGroup",
               "BroadcastArtist",
               "FilmMusicContributor",
               "Lyricist",
               "MusicalArtist",
               "RecordProducer",
               "Continent"
            ],
            "name":"Europe",
            "dbpedia_resource":"http://dbpedia.org/resource/Europe"
         },
         "count":2
}

• Error - Extracts error messages sent by the IBM Watson NLU server, if any, stored in String variables.
• Keywords - The keywords extracted from the specified text, stored in an IBMKeyword variable. The IBMKeyword variable contains the following information for each keyword:
  • Sentiment - The sentiment score and label of the keyword.
  • Relevance - The relevance of the keyword in the specified text.
  • Emotion - The scores for all the defined emotions.

Here is an example of a .json text that shows the IBMKeyword information and structure:

{  
         "text":"Occidental",
         "sentiment":{  
            "score":0.0,
            "label":"neutral"
         },
         "relevance":0.377729,
         "emotion":{  
            "sadness":0.093656,
            "joy":0.56355,
            "fear":0.037183,
            "disgust":0.058567,
            "anger":0.23449
         }
 }

• Language - Extracts the language the given text is in and stores it in a String variable.
• RawResult - Enables you to extract the information from the specified text, in a .JSON format, as a String variable. The data obtained differs depending on the Analysis Type selected.
• Sentiment - The sentiment score and label extracted from the specified text, stored in a KeyValuePair<Double,String> variable.