I have a bot built using nodejs v3 sdk and I am trying to adaptivecards to make UI experience richer. Is there a goto place I can start with ? Some example will be good to for starters. [https://adaptivecards.io/][1]
The Samples section on the Adaptive Cards Website is a good place to start to see the art of the possible with cards. If your using JavaScript for your Bot like I did the Microsoft Documentation for adding Rich Card attachments to the Botframework is useful and this Article has some good examples of showing ways to insert data into fields of Adaptive Cards, which we find useful with loading in answers from QnA Maker.
Hope this helps!
Related
We developed a bot using nodejs(backend) and react webchat(frontend) via directline.In the backend we used adaptive cards with version 1.0 .....now we want to use the latest features of adaptive cards like isRequired parameter here which is not supporting in v1.0 of ac. What my problem is I am using botframework webchat v4.7.0 in which we had so many customizations inside node-modules of botframe-work. If I update this everything will gonna change my entire ui and getting unexpected behaviour and look of my bot.We tried by using v 1.3 but adaptive cards are not even rendering. Can anyone help me in reaching out this plss..it means a lot to me. Thank you.
I have an FAQ document (MS word) that I want to convert into a QnA knowledge base to support a QnA bot. I can use QnA Maker to import text from the document but all images get stripped out. Without the images the answers are nonsensical. I know QnA maker supports markdown so I could add the images back with image markdown tags. This is time consuming and requires the images be stored somewhere that can be accessed by QnA maker (eg Azure blog storage). The other issue with this is that my bot uses Hero Cards for follow on questions (I believe this is the recommended approach for handling follow on questions). However, Hero Cards don't support markdown. I could use adaptive cards instead of hero cards. Adaptive cards do support markdown but not the image tag :(. Also, markdown is not exactly user friendly. It's fine if someone technical is managing the KB. But if this is the responsibility of a non-technical person then markdown isn't ideal.
The fact that QnA Maker doesn't effectively support images seems like a significant oversight. I get that bot conversations should involve short exchanges between bot and user. But (apologies for the cliche) a picture is worth a 1000 words sometimes.
Griping aside, does anyone have suggestions on how to effectively build a KB with images? I'm also open to alternative suggestions on how to use a bot as a front end to an image rich FAQ. I should also add that my bot is accessed via MS Teams.
Thanks
Not a complete answer but I just noticed (or QnA interface was recently added) that you can edit QnA answers without the need to write markdown directly (hurray!). You can even add images though you still need to store/organize them yourself
(screen shot of edit bar). You can also leave your answer as MD if you'd prefer to edit them directly.
Does it make sense to use bot framework for an app on google assistant? is there some documentation around it if it makes sense?
Thanks a lot
We haved discussed the matter in depth here:
https://github.com/Microsoft/BotBuilder/issues/4255
In short, there are no plans for integrating Google Assistant with bot framework as a channel right now but there are third parties solutions on top of Direct Line that allows you to use Bot Framework with actions in Google Assistant.
You can use this open source bot adapter (bot framework V4):
https://github.com/BotBuilderCommunity/botbuilder-community-dotnet/tree/develop/libraries/Bot.Builder.Community.Adapters.Google
We are using it in combination with the Virtual Assistant template (not the previous bot enterprise template for V4). It works, but we did not include it as a Nuget package, we hauled in the project and made some changes as it was flawed here and there (at the time of writing). Converting Google messages into Bot Framework messages (and the other way around) can be quite tricky.
Amazingly, I can't find any documentation for this, and Google tells me nothing. Yet tantalizingly, it is used in the many sample projects supplied by Microsoft to showcase Bot Framework. What is a 'fiber' and what should I use it for in a bot context?
Is it possible to use wit.ai with Microsoft Bot framework, as it supports more languages than LUIS for NLP?if so any sample demo or document available?
Not sure if you are still looking for this but will be helpful to anyone coming here.
There is a very nice npm module
https://github.com/sebsylvester/botbuilder-wit that supports using wit.ai with Microsoft Bot Framework.
To my knowledge, there is no Wit.ai support on the Bot Framework (for now).
Nervertheless, Microsoft plans to do it (https://github.com/Microsoft/BotBuilder/issues/144).
Since the Bot Framework is an open source project, you can do it by yourself but it's quite time consuming.
I am at the early stages, but you can have a look at this.
https://github.com/q3blend/BotBuilder-Wit.ai
I hope it can help you