Partner Bot

A React front end for the Gen AI MLT Partner Bot course. A user picks a company, filing year, and period, asks a question about the company's quarterly (10-Q) filing, and sees an LLM-generated answer.

Built with Vite + React + TypeScript, Amplify UI components, and an Amplify Gen 2 backend (Cognito auth). Hosted on AWS Amplify.

Status

• Chat form — done. Four inputs (company, year, period, question) compose a request body that conforms to the Lambda Contract and render the answer.
• Handler — currently a stub in src/App.tsx (submitQuery) that logs the request body and returns a fake response. The live API Gateway + Lambda integration replaces it in a later module.
• Auth — Cognito User Pool is defined in amplify/auth/resource.ts, wired into the UI in a later module.

Develop

npm install
npm run dev            # Vite dev server on http://localhost:5173

The app runs against the stub handler with no backend required. Open the browser console and submit the form to see the contract-shaped request body logged.

Amplify backend (sandbox)

npx ampx sandbox       # deploys a personal cloud sandbox, writes amplify_outputs.json

amplify_outputs.json is generated and gitignored. src/main.tsx configures Amplify from it automatically when present, and falls back to the stub when it is absent (so the app builds and runs without a deployed backend).

Contract

Request body sent to the inference Lambda:

{
  "question": "What were the key revenue drivers this quarter?",
  "ticker": "MSFT",
  "year": 2024,
  "period": "Q2"
}

Response shape:

{ "answer": "...", "meta": {} }

Company display names map to tickers in COMPANIES (src/App.tsx); period is one of Q1 Q2 Q3 Q4 FY; year is submitted as an integer.