Recognizers-Text/.NET/Samples at master · microsoft/Recognizers-Text

The best way to learn how to use something is through examples. The .NET solution contains a simple sample to get you started.

Simple Console (source)

This sample demonstrate the combination of all Recognizers to extract possible values from the user's input.

The important piece is the Microsoft.Recognizers.Text NuGet package, which you'll need to add using:

Install-Package Microsoft.Recognizers.Text

Then, the sample uses the recognize methods of each available Recognizer. We need to do so by passing the Culture code we'll want to detect. E.g.: en-us.

So far, the available recognition methods are:

using Microsoft.Recognizers.Text;
using Microsoft.Recognizers.Text.DateTime;
using Microsoft.Recognizers.Text.Number;
using Microsoft.Recognizers.Text.NumberWithUnit;
using Microsoft.Recognizers.Text.Sequence;

// Use English for the Recognizers culture
var culture = Culture.English;

// Number recognizer will find any number from the input
// E.g "I have two apples" will return "2".
NumberRecognizer.RecognizeNumber(query, culture);

// Ordinal number recognizer will find any ordinal number
// E.g "eleventh" will return "11".
NumberRecognizer.RecognizeOrdinal(query, culture);

// Percentage recognizer will find any number presented as percentage
// E.g "one hundred percents" will return "100%"
NumberRecognizer.RecognizePercentage(query, culture);

// Number Range recognizer will find any cardinal or ordinal number range
// E.g. "between 2 and 5" will return "(2,5)"
NumberRecognizer.RecognizeNumberRange(query, culture);

// Age recognizer will find any age number presented
// E.g "After ninety five years of age, perspectives change" will return "95 Year"
NumberWithUnitRecognizer.RecognizeAge(query, culture);

// Currency recognizer will find any currency presented
// E.g "Interest expense in the 1988 third quarter was $ 75.3 million" will return "75300000 Dollar"
NumberWithUnitRecognizer.RecognizeCurrency(query, culture);

// Dimension recognizer will find any dimension presented
// E.g "The six-mile trip to my airport hotel that had taken 20 minutes earlier in the day took more than three hours." will return "6 Mile"
NumberWithUnitRecognizer.RecognizeDimension(query, culture);

// Temperature recognizer will find any temperature presented
// E.g "Set the temperature to 30 degrees celsius" will return "30 C"
NumberWithUnitRecognizer.RecognizeTemperature(query, culture);

// Datetime recognizer This model will find any Date even if its write in coloquial language 
// E.g "I'll go back 8pm today" will return "2017-10-04 20:00:00"
DateTimeRecognizer.RecognizeDateTime(query, culture);

// PhoneNumber recognizer will find any phone number presented
// E.g "My phone number is ( 19 ) 38294427."
SequenceRecognizer.RecognizePhoneNumber(query, culture);

//IP recognizer will find any Ipv4/Ipv6 presented
// E.g "My Ip is 8.8.8.8"
SequenceRecognizer.RecognizeIpAddress(query, culture);

//Boolean recognizer will find yes/no like responses, including emoji -
// E.g "yup, I need that" will return "True"
ChoiceRecognizer.RecognizeBoolean(query, culture);

All these methods accept an input string and culture, and returns an IEnumerable of ModelResult:

Alternatively, you can obtain model instances that can be re-used for recognizing multiple inputs:

// Number model
var recognizer = new NumberRecognizer("en-us");
var model = recognizer.GetNumberModel();

// Parse input using Number model
var result = model.Parse("I have twenty apples");

// result is:
// [
// 	{
// 		"start": 7,
// 		"end": 12,
// 		"resolution": {
// 			"value": "20"
// 		},
// 		"text": "twenty",
// 		"typeName": "number"
// 	}
// ]

BotBuilder Sample (source)

This sample demonstrate how the Recognizers can be used with a BotBuilder Bot to parse user input. The bot provides a basic experience for ordering roses, it starts by asking the amount of roses and then asks for a delivery date and time.

To test the sample:

Launch the BotFramework Emulator.
File -> Open bot and navigate to <your_project_folder>/Samples/BotBuilder folder.
Select BotBuilderRecognizerBot.bot file.

Once connected, the bot will send a welcome message. You can start the order flow by sending any message to the bot.

In order to validate user input, Recognizers-Text are used: Recognizers-Text.Number and Recognizers-Text.DateTime.

This is a copy of the DeliveryDialog ValidateQuantity method which validates that the user input is an Integer, between 1 and 100 and not a decimal number:

	private async Task<bool> ValidateQuantity(PromptValidatorContext<string> promptContext, CancellationToken cancellationToken)
        {
            var result = promptContext.Recognized.Value ?? string.Empty;
            var results = NumberRecognizer.RecognizeNumber(result, culture);

            if (results.Count == 0)
            {
                await promptContext.Context.SendActivityAsync(InvalidQuantityErrorMessage);
                return false;
            }

            if (results.First().TypeName == "number" && double.TryParse(results.First().Resolution["value"].ToString(), out double value))
            {
                // Validate number
                if ((value < 1) || (value % 1 != 0))
                {
                    await promptContext.Context.SendActivityAsync(InvalidQuantityErrorMessage);
                    return false;
                }

                if (value > 100)
                {
                    await promptContext.Context.SendActivityAsync(InvalidOverQuantityErrorMessage);
                    return false;
                }

                var quantityRoses = Convert.ToInt32(results.First().Resolution["value"]);
                var quantityMessage = quantityRoses == 1
                ? "I'll send just one rose."
                : $"I'll send {quantityRoses} roses.";
                promptContext.Recognized.Value = quantityRoses.ToString();
                await promptContext.Context.SendActivityAsync(quantityMessage);
                return true;
            }
            else
            {
                await promptContext.Context.SendActivityAsync(InvalidQuantityErrorMessage);
                return false;
            }
        }

The Prompts are added in a waterfall Dialog:

	public DeliveryDialog(IStatePropertyAccessor<DeliveryState> userProfileStateAccessor, ILoggerFactory loggerFactory)
            : base(nameof(DeliveryDialog))
        {
            UserProfileAccessor = userProfileStateAccessor ?? throw new ArgumentNullException(nameof(userProfileStateAccessor));
            // Add control flow dialogs
            var waterfallSteps = new WaterfallStep[]
            {
                    InitializeStateStepAsync,
                    PromptForQuantityStepAsync,
                    PromptForDateStepAsync,
                    DisplayDeliveryStateStepAsync,
            };
            AddDialog(new WaterfallDialog(ProfileDialog, waterfallSteps));
            AddDialog(new TextPrompt(QuantityPrompt, ValidateQuantity));
            AddDialog(new TextPrompt(DatePrompt, ValidateDate));
        }

Asking the user for a specific delivery time may require special parsing, like extracting both date and time from the user input, or even obtain a range of dates and times.

The DeliveryDialog ValidateDate method does exactly that. It will prompt the user for a possible delivery time, parse the user's input and extract, at least, one of these available return values using the DateTime Recognizer:

date
daterange
datetime
datetimerange

(These are the DateTime Recognizer types that contains date information)

NOTE: The DateTime Recognizer uses LUIS datetimeV2 subtypes. For a full list, please visit LUIS prebuilt entities - Subtypes of datetimeV2.

Recognizer Function (source)

This sample is a variant of the SimpleConsole, which you can deploy as a Web API using Azure Functions (serverless). It is a combination of all Recognizers to extract possible values from the user's input.

You can pass the text (required) and culture (optional) properties to the WebAPI using GET or POST. The response will be a JSON array containing all entities.

Example URL:

https://[app-name].azurewebsites.net/api/entities?text=[INPUT TEXT]&culture=[CULTURE CODE]