ybot1122 / deburgger

Deburgger is an AI powered chatbot who will help debug and troubleshoot your burger.

Deburgger

Deburgger is a progressive web application that uses Azure and artifical intelligence to help you debug your burger.

Available Scripts

In the project directory, you can run:

npm start

Runs the app in the development mode.
Open http://localhost:3000 to view it in the browser.

The page will reload if you make edits.
You will also see any lint errors in the console.

npm test

Runs unit tests

npm run build

Builds the app for production to the build folder.
It correctly bundles React in production mode and optimizes the build for the best performance.

The build is minified and the filenames include the hashes.
Your app is ready to be deployed!

deburgger's Issues

Manually Curated Data Source: Burger Troubleshooting.

This is the original motivation of this project. We will need a manually curated data source for troubleshooting a homecooked burger. This might be done via the Language Studio UI, or done by a text source.

Examples include:

"What should I do if my bun gets soggy?" | "Try toasting the buns, and make sure the toppings are dry before putting them on".
"How do I get grill marks on my patty?" | "The simplest answer is to cook it on a grill."
"My condiments keep falling off." | "Try using less condiments. Otherwise try spreading them before closing the bun."

NOTE: This might end up using the a script similar to 'known_burgers'.

Integrate With Azure CLU

We need to produce a dataset that can train a model to learn intents and entities. And then integrate to it on the clientside.

Intent tells us what a user wants to do, and entities can highlight specific parameters. We'll need an initial dataset to just do some programming.

In the long term, we will simply produce sentences and outsource the data labeling.

Example:

[
    {
        "intent": "Quote",
        "language": "en-us",
        "text": "I left my heart in Alexandria Egypt",
        "entities": [
            {
                "category": "CityName",
                "offset": 19,
                "length": 10
            }
        ]
    },
    {
        "language": "en-us",
        "intent": "BookFlight",
        "text": "Book me a flight to Alexandria Egypt",
        "entities": [
            {
                "category": "CityName",
                "offset": 20,
                "length": 10
            },
            {
                "category": "CountryName",
                "offset": 31,
                "length": 5
            }
        ]
    }
]

Implement Clientside Input Validation

We should always reject any user input that contains:

• URLs/hyperlinks
• PII
• Certain special characters

Use Generic Analysis Endpoint to reduce Number of Calls

We can use generic analysis endpoint to return multiple actions with a single API Call: https://github.com/Azure/azure-sdk-for-js/blob/main/sdk/textanalytics/ai-text-analytics/samples/v5/typescript/src/beginAnalyzeActions.ts

Linked Entity Recognition + Summarize Text

Use Linked Entity Recognition to get text from a link, and then use the Summarize API to return a summary of the text to users.

Summarization API: https://learn.microsoft.com/en-us/azure/cognitive-services/language-service/summarization/quickstart?pivots=rest-api&tabs=document-summarization%2Cwindows

Write Script to Generate Known_Burgers.xlsx Data Source

The script should take a mapping of Restaurant -> MenuItem[] and an array of prompts: Prompt[]

The script should generate an xlsx where the A column is a question and the B column is an answer.

{
    'Burger King': ['Whopper', 'Whopper Jr', 'Bacon King']
}

['Where can I find the ${item}?', 'Which place has the ${item}?', 'Who makes the ${item}?']

results in

Where can I find the Whopper? | Burger King
Which place has the Whopper? | Burger King
Who makes the Whopper? | Burger King
Where can I find the Whopper Jr? | Burger King
Which place has the Whopper Jr? | Burger King
Who makes the Whopper Jr? | Burger King
Where can I find the Bacon King? | Burger King
Which place has the Bacon King? | Burger King
Who makes the Bacon King? | Burger King

Create Q&A Data Sources

https://learn.microsoft.com/en-us/azure/cognitive-services/qnamaker/concepts/data-sources-and-content

Move Cognitive API Calls to Serverside

We need to setup a node.js server to handle the API calls to Azure because it requires an API secret.