StockGraber

English | 繁體中文

A desktop stock-charting app built with finplot embedded in a PySide6 (Qt) window, styled as a dark "trading terminal". It pulls real daily OHLC data from Yahoo Finance (via yfinance), caches it locally in SQLite, and renders candlesticks with volume, moving-average crossovers, RSI/MACD sub-panels, and a percent-change index-comparison view. It launches full screen.

⚠️ Not financial advice. StockGraber is a charting/learning tool. The indicators and signals it draws (golden/death crosses, RSI, MACD, etc.) are common technical-analysis conventions, not buy/sell recommendations. No indicator reliably predicts price.

---

Features

• Dark terminal UI — title bar, command bar, a stacked chart column (price / volume / RSI / MACD) and a sidebar of cards (Quote / Indicators / Comparison / Legend / Date Range). Launches full screen; Exit button or Esc to quit.
• Bilingual UI — one-click toggle between English and Traditional Chinese (繁體中文) for the whole interface.
• Recent symbols — every symbol you successfully chart is remembered in the local database and listed as clickable chips for one-click reuse.
• Real data, cached — daily OHLC from Yahoo Finance, stored in a local SQLite database so repeat loads are instant and only missing dates are downloaded.
• Candlesticks + volume — price panel with a volume sub-panel, colored by up/down day.
• Moving-average crossovers — preset fast/slow pairs (9/21, 20/50, 50/200, 100/200) shown as Fast/Slow value tiles; toggle SMA ↔ EMA.
• Golden / death cross markers — bright-yellow-ringed ▲ (bullish) / ▼ (bearish) arrows where the fast MA crosses the slow MA.
• RSI & MACD sub-panels — RSI (14) with a 30–70 band, MACD (12,26,9) with a green/red histogram; Volume / RSI / MACD panels can each be shown/hidden.
• Index comparison (percent mode) — pick a market index (US / Canada / Hong Kong / global) and the price panel switches to a rebased %-change view: stock as a filled area, index as a line, MAs in %, on a percent axis — rebased to the start of the visible window and re-fit (~90% height) as you pan/zoom.
• Hover readout — a boxed O/H/L/C panel with the day's change vs. the previous close (color-coded ▲/▼).
• Navigation — drag to pan, mouse-wheel or +/−/Reset buttons to zoom, a timeframe segmented control (1M…MAX), and a log/linear toggle. In candle mode the highest visible bar is kept ~80% up the panel.
• Responsive UI — all network fetches run on a background thread, so the window never freezes on a cold symbol.

---

Requirements

• Python 3.10+ (developed on 3.12)
• Windows / macOS / Linux (developed on Windows 10)
• Packages: PySide6, finplot, pandas, numpy, yfinance (pyqtgraph is pulled in by finplot)
• An internet connection for the first fetch of each symbol/date range

---

Installation

pip install PySide6 finplot pandas numpy yfinance

Qt binding note: finplot installs PyQt6, while this app uses PySide6. The app pins PYQTGRAPH_QT_LIB=PySide6 at startup so both halves agree — you don't need to set anything yourself.

Helper scripts

On Windows double-click the batch files; on macOS/Linux run the shell scripts:

---

Usage

python StockGraber.py

(or double-click run.bat on Windows / run ./run.sh on macOS & Linux)

It opens full screen with AAPL loaded. Type a ticker and press Enter (or click Load) to chart another symbol. Quit with the ✕ Exit button (top-right) or the Esc key.

Ticker symbols

Use the Yahoo Finance symbol format:

---

Controls

Command bar (top)

Sidebar — Indicators card

Sidebar — other cards

Mouse

• Drag — pan left/right through history
• Wheel — zoom; Ctrl+Wheel — vertical zoom
• Right-drag — box zoom
• Hover — OHLC + change readout box (top-left of the price panel)

---

Indicators & signals (quick reference)

MA-pair character: 50/200 = long-term, fewest/most-trusted signals; 20/50 = swing trading; 10/20 & 5/10 = short-term, more signals and more noise (whipsaws); 9/21 & 12/26 = popular short–medium pairings, traditionally used as EMAs.

---

Index comparison list

• United States — S&P 500 (^GSPC), Dow (^DJI), Nasdaq Composite (^IXIC), Nasdaq 100 (^NDX), Russell 2000 (^RUT), VIX (^VIX)
• Canada — S&P/TSX Composite (^GSPTSE)
• Hong Kong — Hang Seng (^HSI), HS China Enterprises (^HSCE)
• Global / Other — FTSE 100 (^FTSE), DAX (^GDAXI), CAC 40 (^FCHI), Nikkei 225 (^N225), Euro Stoxx 50 (^STOXX50E), ASX 200 (^AXJO), Nifty 50 (^NSEI)

Selecting an index switches the price panel to percent-change mode: both the stock and the index are rebased to 0% at the start of the visible window, so where the stock line ends above the index line, the stock outperformed.

---

Building a standalone app

On Windows, double-click build.bat to produce dist\StockGraber.exe. On macOS, run ./build.sh to produce dist/StockGraber.app (plus a plain dist/StockGraber UNIX binary).

Both are single-file apps with the StockGraber icon that run without a Python install. Under the hood the build forces the PySide6 Qt binding (PyQt5/PyQt6 are excluded, since PyInstaller can't bundle multiple Qt bindings):

pyinstaller --noconfirm --windowed --onefile --name StockGraber \
  --icon StockGraber.ico --collect-all finplot --collect-all pyqtgraph \
  --exclude-module PyQt5 --exclude-module PyQt6 StockGraber.py

The Windows exe writes its SQLite cache next to itself; the macOS app writes it to ~/Library/Application Support/StockGraber/ (the .app bundle itself can be read-only under Gatekeeper). The build output (build/, dist/) is git-ignored — distribute the app via a GitHub Release rather than committing it.

macOS note: the built app is unsigned, so on another Mac you may need to right-click → Open the first time (or xattr -dr com.apple.quarantine StockGraber.app) to get past Gatekeeper.

---

Project layout

StockGraber/
├── StockGraber.py     # the app: PySide6 window, chart panels, controls
├── data.py            # data layer: yfinance fetch, SQLite cache, indicators
├── StockGraber.ico    # app icon (built from StockGraber.png)
├── setup / run / build.bat   # Windows helper scripts
├── setup / run / build.sh    # macOS & Linux helper scripts
├── stockgraber.db     # SQLite cache (auto-created on first run, git-ignored)
└── README.md

data.py — data layer

• get_ohlc(symbol, start, end, force_refresh=False) — daily OHLC for a date range. Reads from cache and downloads only the missing dates; never re-downloads data already stored (a per-symbol coverage window is tracked so even empty spans, like pre-IPO dates, aren't re-requested).
• moving_average(series, period, method) — SMA or EMA.
• ma_cross_markers(df, fast, slow, method) — golden/death cross points.
• rsi(close, period=14) — Wilder's RSI.
• macd(close, fast=12, slow=26, signal=9) — MACD line, signal, histogram.
• record_symbol(symbol) / recent_symbols(limit) / forget_symbol(symbol) — the reuse history behind the Recent symbols card.

The database has three SQLite tables: ohlc (the candles, keyed by symbol, date), coverage (the date span already fetched per symbol), and history (symbols you've charted, with use-count and last-used time).

StockGraber.py — the app

A single SpikeWindow (QMainWindow) hosting four x-linked finplot panels (price / volume / RSI / MACD) plus the dark control chrome and sidebar cards. Network fetches run on a QThread worker and deliver results back to the UI thread via Qt signals. Chart items are split into volume, panel (RSI+MACD) and price (candles or percent lines) groups, so changing MA settings re-renders only the price panel without refetching or losing your zoom.

Tunable constants on SpikeWindow:

data.py constant DEFAULT_START ("2015-01-01") sets how far back a symbol is first fetched when no start date is given.

---

Notes & limitations

• Prices are not split/dividend-adjusted (auto_adjust=False). Cached rows are only overwritten when re-fetched, so old bars keep their raw values after a split until that range is re-fetched.
• Index comparison is rebased to the start of the visible window and re-fits as you pan/zoom; it replaces the candles with a percent-change view while active.
• MA periods are preset-only (4 pairs); RSI/MACD periods are fixed (14, and 12/26/9).
• Data quality depends on Yahoo Finance; occasional gaps or glitches are possible.

---

Disclaimer

StockGraber is for educational and informational purposes only and is not financial advice. Do your own research and consult a qualified professional before making investment decisions. Market data is provided by Yahoo Finance and may be delayed or inaccurate.