Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Invest Alchemy is a trading assistant focused on ETF portfolios.
| Date | Stars |
|---|---|
| 2026-07-24 | 769 |
| 2026-07-25 | 768 |
| 2026-07-28 | 768 |
| 2026-07-30 | 768 |
| 2026-08-06 | 768 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.00%/day
# Invest Alchemy
Invest Alchemy is a trading assistant focused on ETF portfolios. For more context, see this [post](https://www.bmpi.dev/money/invest-alchemy/).
## TODO
- Feature
- Data
- [x] [Tushare](https://tushare.pro/)
- [ ] [AKShare](https://akshare.xyz/)
- Other
- [理杏仁](https://www.lixinger.com/)
- [Yahoo](https://finance.yahoo.com/)
- [ETF.com](https://www.etf.com/)
- [ETFDB.com](https://etfdb.com/)
- Strategy
- Trade strategies
- [x] Double MA
- [ ] Turtle
- [ ] Backtrace MA250
- [ ] Breakthrough Platform
- [ ] Keep Increasing
- [ ] Low Atr
- Trade backtesting
- [ ] Spike [backtrader](https://github.com/mementum/backtrader), maybe we can use it to do the strategy backtesting before we go to implement the strategy signal.
- [ ] ETF similarity calculation, can filter similar ETF
- https://www.etf.com/etfanalytics/etf-comparison-tool
- Market
- [ ] Support USA share ETF
- [ ] Index Historical P/E Ratio
- [x] Base database store baseline (10%) and popular indexs performance (000905/000300/399006/HSI/IXIC/INX)
- Portfolio
- [x] Transaction/Funding/Holding/NetValue/Performance Ledger
- [ ] Add portfolio fund utilization performance (holding assets / total assets)
- [x] Add popular indexs net value compare Ledger
- [x] Support trade A share ETF/LOF funds
- [x] Support trade A share stock
- [ ] Support trade A public funds (not include ETF/LOF)
- Trader
- User
- [x] User portfolio calculation support
- Robot
- Double MA (11/22)
- [x] Long ETF strategy robot trader
- [x] All In strategy robot trader
- [ ] Periodic Payment strategy robot trader
- Web UI
- [x] Re-Design web UI page
- [x] Portfolio Risk Monitor
- [x] Portfolio holding/transaction/funding history
- Notify
- [x] Email
- [x] Telegram bot/channel
- [ ] Slack bot/channel
- [ ] Show different strategy signal (daily)
- [ ] Show different trade robot portfolio performance report (weekly)
- [ ] Can subscribe trade event of the special portfolio by email/rss
- Arch Improve
- Storage
- [x] Add database schema migration tool
- [x] SQLite DB backup to AWS S3
- [x] Trade data store in SQLite DB
- [x] Base Database migrate from SQLite to Supabase(PostgreSQL)
- Bug Fix
- [x] Trading signals error when double ma price are same.
- [ ] Portfolio performance ledger max_days_of_continuous_loss calculation error, max is 5, because the days_of_continuous_loss will set to 0 when on non-trade days.
## Note
### Strategy
- Trade strategy
- [Sequoia](https://github.com/sngyai/Sequoia)
- Issue
- **Split-adjusted share prices**: Use `qfq` adjust price to generate trade signals.
- Note
- Why the trade strategy may not good at trading stock?
- Compare index, stocks have the liquidity problem and suspension risk. Liquidity can lead to a huge gap between simulated trading and actual trading results, suspension will result in inability to trade, producing results that are completely different from simulated trading.
- **About strategy backtesting overfitting**: Backtesting overfitting may be a killer in trade stragety, so we not pursue the profit maximization by overfitting, just using the simple trade strategy to overcome the market volatility.
### Portfolio
- **Split-adjusted share prices**: Not processed in the trading ledger (only robots, **user trader need add a transaction when split-adjusted happens**), but processed in the holding ledger calculation.
- [x] 🐛 Robot user still can not process, because split-adjusted check on the portfolio net value ledger calculation, but sell transaction amount do not equal to holding amount because sell transaction generated on the transaction ledger calculation, so it needs to process on portfolio net value calculation when user is a robot type.
- ~~Fix this issue by: robotExcerpt of 6,354 characters
Read on GitHub
Dawei Ma · @i365dev · China
203
Would you bet a product on this? Bounded 0–100 and slow moving.
matched fp:1da5aa7fdce1a4c3, topic:serverless