AccessibleAI.Bots.Blobs 0.0.16-dev

AccessibleAI.Bots.Blobs

This package contains utility classes related to bot management using Azure Storage Blobs.

Conversation Logging

This project comes with a ConversationBlobStorage class that can be used to log all conversations to Azure Blob Storage in a human-readable format like the following:

singularity-sally 
Hi, I'm Singularity Sally and I'm here to help you learn data science!

singularity-sally 
Please tell me what you'd like to learn about today.

You @ 9/17/2022 11:38:09 PM 
I want to learn about AI!

singularity-sally 

Artificial Intelligence, or AI, refers to the capability for computers to emulate the decision-making processes of creatures (including humans). AI includes everything in machine learning and deep learning along with conversational AI, game AI, decision trees, and other forms of AI.

Usage

In your Startup.cs file add a using statement to get access to the AccessibleAI.Bots.Blobs namespace:

using AccessibleAI.Bots.Blobs;

Next we'll need to add or expand the existing constructor as follows to create a ConversationBlobStorage class:

private readonly ITranscriptStore _conversationStorage;

public Startup(IConfiguration config)
{
    _conversationStorage = new ConversationBlobStorage(config.GetValue<string>("StorageConnStr"), config.GetValue<string>("ConversationContainerName"));
}

Note that StorageConnStr and ConversationContainerName refer to settings in your appsettings.json file. Customize these values with whatever settings you're using.

Also note that the storage container must already exist. You can create it using Azure Storage Explorer or the Azure Portal.

Next we'll need to add the following to the ConfigureServices method:

services.AddSingleton(_conversationStorage);
services.AddSingleton<IMiddleware, TranscriptLoggerMiddleware>(_ => new TranscriptLoggerMiddleware(_conversationStorage));

Next, in your bot's AdapterWithErrorHandler or similar CloudAdapter constructor, you'll need to add an ITranscriptStore transcriptLogger parameter to the parameter list.

After that, you should add the following line:

Use(new TranscriptLoggerMiddleware(transcriptLogger));

Once you deploy your app, new conversation logs will appear in your Azure Storage container.

Note that you may want to delete old months and years and only store the most recent conversations. Right now there is no provided mechanism in this library for doing that.