top of page

Can UiPath integrate with natural language processing (NLP) tools?

Updated: Mar 29, 2023

Certainly! Integrating UiPath with natural language processing (NLP) tools can bring about intelligent automation of text-based tasks that require natural language understanding. This includes tasks like sentiment analysis, entity extraction, language detection, and more. UiPath offers a range of activities and custom activities that allow developers to connect to NLP APIs and process text data within UiPath workflows.


To integrate UiPath with NLP tools, you need to follow these steps:

  1. Choose an NLP tool that fits your needs and create an account to access its API. Some popular options are Google Cloud Natural Language, IBM Watson, and Amazon Comprehend. You can also use open-source libraries like spaCy, NLTK, and Stanford CoreNLP.

  2. Install the relevant packages or libraries for the chosen NLP tool. For example, to use Google Cloud Natural Language, you need to install the Google.Cloud.Language.V1 package.

  3. Use UiPath's HTTP Request activity or custom activities to send text data to the NLP API and retrieve the results.

For example, if you want to perform sentiment analysis on a text string using Google Cloud Natural Language, you can use the following workflow code:


// Set up the API request
UiPath: Assign activity: apiKey = "your_api_key"
UiPath: Assign activity: text = "your_text_string"
UiPath: Assign activity: requestUrl = "https://language.googleapis.com/v1/documents:analyzeSentiment?key=" + apiKey
UiPath: Invoke Code activity:
    arguments:
    - in_text: text
    - out_response: new System.Net.Http.HttpResponseMessage
    - out_content: new Newtonsoft.Json.Linq.JObject
    - out_error: new System.Exception
    code:
        // This is C#. Send the API request and parse the response
    var client = new System.Net.Http.HttpClient()
        var requestBody = new {
            document = new {
                type = "PLAIN_TEXT",
                content = in_text
            }
        }
        var requestContent = new Newtonsoft.Json.Linq.JObject(requestBody)
        var response = client.PostAsync(requestUrl, new System.Net.Http.StringContent(requestContent.ToString(), System.Text.Encoding.UTF8, "application/json")).Resultout_response = responseout_content = Newtonsoft.Json.Linq.JObject.Parse(response.Content.ReadAsStringAsync().Result)

// Process the API response
UiPath: Assign activity: sentimentScore = out_content("documentSentiment")("score").ToString
UiPath: Assign activity: sentimentMagnitude = out_content("documentSentiment")("magnitude").ToString
UiPath: Log Message activity: "Sentiment score: " + sentimentScore
UiPath: Log Message activity: "Sentiment magnitude: " + sentimentMagnitude

This code sends a POST request to the Google Cloud Natural Language API's analyzeSentiment method with the text data, retrieves the sentiment score and magnitude from the API response, and logs them to the UiPath output panel.

By integrating UiPath with NLP tools, you can automate text-based tasks that require natural language understanding. For example, you can automate the processing of customer feedback to identify areas of improvement, extract product names from reviews, or identify language-specific content for translation. This can help organizations streamline their processes and improve their efficiency.


The above code given in the example is UiPath workflow code, which is used in the UiPath Studio software to automate repetitive tasks. The code uses UiPath's activities and custom activities to connect to an NLP API and process text data, in this case, performing sentiment analysis using the Google Cloud Natural Language API. It also uses C# programming language syntax to send HTTP requests to the API, parse the response, and extract the sentiment score and magnitude, which are then logged to the UiPath output panel. Please see notes in code above. And remember, there are always more way than one!


Now, let's see how we can implement this in UiPath using the Invoke Python Method activity.

26 views0 comments

Comments

Rated 0 out of 5 stars.
No ratings yet

Add a rating
bottom of page