watson-personal-assistant / developer Goto Github PK

Document how the audio gateway can be configured to use another Speech To Text Service besides the embedded and included supported Speech to text service that comes with Watson Assistant Solutions. Document how to check which Speech to Text Service your Tenant ID and Audio Gateway service is currently using.

Document what languages the audio gateway supports. For now use the Languages that are supported by Watson STT and TTS:

• English
• Spanish
• French
• Japanese
• Korean
• Portuguese

Document how the customer can switch their configuration to use their own Google STT service and corresponding API key.

https://watson-personal-assistant.github.io/developer/skill/knoweldge-store

When trying to launch the Watson Assistant service located in Germany, I get the following error message:

404 Not Found: Requested route ('watson-assistant.eu-de.bluemix.net') does not exist.

Launching a service located in UK works fine. I have also created a ticket as I know this might not be the correct place to report the issue.

Cards enable multi modal input  voice / text and selection. They can be used in any skill to improve the experience. IE  Rather than answer back by saying  or typing back   "Sheraton at park plaza NYC  you can pick from a list of card options that also include graphics for example pictures of the hotel.   Where the user clicks on the one they want.   That example is for the POI skill     How they render is left up to the client.   For example for the weather skill Kaon wanted to show clouds or sun or partly cloudy based on images returned from Weather service in the weather skill.
 
Skills SDK/REST and JSON structure format needs to describe how to return for example images, audio and a list of choosable options. Additionally we need to do some thinking regarding specifying how to represent display / sizes impact the returned data.

What skills

1. By Solution/ Industry
2. GA, Alpha, Beta
3. Description
4. Skill-Set - Built-In
5. Availability by Geo/Datacenter - Germany versus Dallas

Skills are listed here examples: Slide 16 ( remove reminder from the list/product)
https://ibm.box.com/s/pm48vi3w55g48483joljoxm1hfq7bgrq

Tracks: https://github.ibm.com/ConsumerIoT/SaganCore/issues/705

When explaining the architecture and the evaluate request sent by Watson Assistant Solutions to the skill we need to include the JSON structure sent and the JSON Structure that is required to be returned. A skill developer will never be able to create a skill without this. The APIs need to be documented in the Swagger.

https://watson-personal-assistant.github.io/developer/understand-service/how-it-works/

JSON for 1, 2, 3, 4 and 5.

Here are examples: Slide 18
https://ibm.box.com/s/pm48vi3w55g48483joljoxm1hfq7bgrq

Need to document / provide examples of the REST interfaces a skill has like hello world. These are different than the REST interfaces we link to on the CORE on the console home page.

The docs are currently missing information about how you authorize yourself if you try to make an API request against Watson Assistant. This is especially confusing since the Swagger examples to not have the API key when you execute them, and the error talks about an authorization, even though auth happens using a GET parameter.

Do change this I'd suggest that a new paragraph is added, e. g. to the converse page, which explains how you can authorize external requests to the service.

The using wcs tutorial contains outdated information on how to integrate WCS. The evaluation should happen completely in skill, using SkillBoilerplate's "evaluate" action.
Basically, the handler.converse in 'unhandled' action is unnecessary, and Step 6 - adding credentials to skill can also be omitted.

The documentation should state what languages Watson Assistant Solutions supports.

Watson Assistant Solutions

Core Service ( routing text )
The core service is able to handle multi-language conversations.

Audio Gateway Service ( routing audio )
We support multiple languages aligns with Watson Speech to Text - STT and TTS for See https://console.bluemix.net/docs/services/speech-to-text/input.html#models

Customers can separately contract a third party like Google for their own own STT / TTS service if they need more Audio Gateway language support. See doc on how to client configure you API key.

Skills
What are the set of languages a skill has been translated to and that it supports.

Out of the box Skills
English only. Currently Watson Assistant Solutions For Industry on supports English conversations. The skills haven't been translated to other languages.

Custom Skills
Can support multi-language if you created, translated and tested in those languages. See languages here: https://console.bluemix.net/docs/services/conversation/lang-support.html#supported-languages

Skill Sets

Updates required to the Accessing your Watson Assistant logs.

The doc states: Step 5 "Please use the value under Tenant Id (it looks similar to a02ad488f-a456-423c-3drsg348hcv8) to send the following in an email to your Watson Assistant support representative." There is no link to the email or way that I can see to send this information in to IBM.

We need to document what Datacenters we support

'Assistant' is spelled 'Assitant' in the diagram.

@luaithrenn

What is the git_url for the hello world skill boilerplate in the page https://watson-personal-assistant.github.io/developer/skill/build-skill/