• add 100 to mortimierliu0
• add 100 to mortimerliu1

How to select the 200

• enough volume?
• enough volatility?

Refactor the directory structure

Goal: Enable running codes outside of the directory

• remove useless code
• refactor codes

update the on_discounted event handler to handle ConnectionRefusedError and try to reconnect until success.

• For Trader, only process the events for today
• Run topic clean up every time Trader restart
• For app.py, also only process the events for today

Currently ReaTimeTrading and app.py only support processing events from today.

Issue: invalid username or password every morning.

the issue seems gone for now after creating separate paper trading username and password.

Currently need to remove all intraday high and low events (by deleting kafka topic) which is not ideal

• make the consecutive high or low sign symbol solid
• change time to local in websocket messages (although not nessssory as js can handle it automatically)

Automatically reset the states of trader, intrady_ticker between days

• RawTicker: read and write messages

Solution: define a new class UTCDatetime representing utc time and always be timezone aware. use this class internally everywhere. In this way, we can gaurantee that the datetime is actually in UTC and tz aware

Show top 4 increasing stocks and top 4 descreasing of the day at the top of the page

• Kafka topic names: top-high, top-low
• data structure: TopSymbol, TopNSymbols

Design

1. IntradayTicker is responsible for calculating each ticker's largest high and low (current)
2. RealTimeTrading then ranks all tickers accordingly and emit events
   a. when to rank: every message
   b. when to emit: only when the top symbols changed (including pricing changes)

• Apply a filter to the RawTicker events. Only cosume events within the core trading hours, exlcuding optional bake in/out time.
• The core trading hours could be obtained from exchange_calendars
• In this implementation, the core trading hours are determined based on NASDAQ exchange