[Text Analysis] Text Analysis Component - Documentation

Forge Component
Published on 13 May (9 days ago) by Rui Parente
1 vote
Published on 13 May (9 days ago) by Rui Parente

How to use the Text Analysis Component in your OutSystems application

Use the Text Analysis Component to carry out different text analysis operations such as key phrase detection, sentiment analysis and so on. 

Key phrase detection

Drag and drop the KeyPhrasesHighlight web block to your screen. This block has two mandatory input parameters:

  • Text - The text to be analyzed for detecting key phrases

  • Language - A two letter code representing the language that follows the ISO 639-1 standard of language codes

The results of the detection will be shown by changing the style of the detected key phrases to bold with a yellow background color. The GetKeyPhrasesInText action analyzes the text using the Text Analytics API KeyPhrases action from the Azure Cognitive Services Connector.

Sentiment analysis

  1. Drag and drop the SentimentDetection web block to your screen. The web block has three placeholders and the action that performs the sentiment detection (DetectSentimentInText Server Action)

  2. Customize these placeholders (using icons, images and so on) to show the result of the sentiment detection of a text. 

Each placeholder is related to an interval of a scale that ranges from 0% to 100% where:

Negative: 0% - 39% Neutral: 40% - 69% Positive: 70% - 100%

The percentage shows the score of the detected sentiment and is an indicator of how the sentiment of the text is positioned in each scale interval.

By default, the following emojis are indicators of the text sentiment.

Language Detection

The GetLanguageFromText server action receives an input parameter, Text, and retrieves the language code detected in the text. The returned code is the two letter code that follows the ISO 639-1 standard.

Exception handling

The possible exceptions that can happen in the actions of the UI Web Blocks are handled directly by showing a Feedback Message when an exception occurs. 

You need to handle any exceptions raised inside the server actions within the component.

Nice component, easy to implement!