Skip to content

DESKTOP_APP.md

Latest commit

 

History

History
142 lines (105 loc) · 5.54 KB

DESKTOP_APP.md

File metadata and controls

142 lines (105 loc) · 5.54 KB

CryptVault Desktop Application

Overview

CryptVault ships with a native desktop GUI built on Tkinter and Matplotlib. It provides candlestick charting, pattern detection overlays, and ML price predictions in a dark-themed trading terminal layout.

Launching

python launch_desktop.py

Or via the CLI:

python cryptvault_cli.py --desktop

Layout

┌────────────────────────────────────────────────────────────────┐
│ CryptVault v6.0   Symbol: [BTC-USD]  [Analyze]  1D 5D 1M ...  │  <- Topbar
├──────────┬─────────────────────────────────┬───────────────────┤
│ NAVIGATE │                                 │  ML Prediction    │
│          │      Candlestick Chart          │  ─────────────    │
│ Chart    │      Bollinger Bands            │  Detected         │
│ Analysis │      Pattern overlays           │  Patterns         │
│          │      Volume                     │  (scrollable)     │
│ ─────── │      RSI                        │                   │
│ LAST     │                                 │                   │
│ ANALYSIS │                                 │                   │
└──────────┴─────────────────────────────────┴───────────────────┘
│ Status bar                                                      │
└────────────────────────────────────────────────────────────────┘

Topbar Controls

Control Description
Symbol entry Type any ticker (BTC-USD, ETH-USD, AAPL) and press Enter or Analyze
Quick-pick buttons BTC, ETH, SOL, BNB, XRP one-click analysis
Timeframe buttons 1D, 5D, 1M, 3M, 6M, 1Y, 2Y
Analyze button Fetches data, detects patterns, runs ML prediction

Chart Panel

The chart panel renders a professional candlestick chart with three sub-plots:

  1. Price chart (top 60%) — candlesticks, Bollinger Bands, pattern overlays
  2. Volume (middle 20%) — color-coded volume bars (green = up, red = down)
  3. RSI (bottom 20%) — 14-period RSI with overbought (70) and oversold (30) zones

Toolbar Toggles

  • BB — Bollinger Bands (20-period, 2 standard deviations)
  • Volume — volume subplot
  • RSI — RSI subplot
  • Patterns — pattern overlays on price chart
  • Reset Zoom — restores full date range

Scroll to Zoom

Scroll the mouse wheel over the chart to zoom in/out horizontally.

Pattern Overlays

Patterns are drawn directly on the candlestick chart:

Pattern Type Visual
Candlestick patterns Triangle marker (^ bullish, v bearish) at pattern bar
Double Top / Bottom Lines connecting the two peaks/troughs + dashed neckline
Head & Shoulders Lines connecting LS → Head → RS + dashed neckline
Triple Top / Bottom Lines connecting all three peaks/troughs
Triangles (Sym/Asc/Desc) Upper and lower trendlines over the pattern window
Rising / Falling Wedge Upper and lower trendlines over the pattern window
All patterns with target Dotted horizontal target price line

Green = bullish pattern. Red = bearish pattern.

Analysis Panel

The right panel displays two cards:

ML Prediction Card

  • Predicted Price — next-bar price estimate
  • Direction — UP / DOWN / NEUTRAL with arrow indicator
  • Confidence gauge — 0–100% horizontal bar
  • Metadata — MAPE, horizon, model name

The system tries ProductionPredictor (full ensemble) first, falling back to a momentum-based estimate if the full predictor is unavailable.

Detected Patterns Card

Scrollable list of all patterns found in the current dataset. Each row shows:

  • BULL / BEAR badge
  • Pattern name and description
  • Category (CANDLESTICK / REVERSAL / CONTINUATION / etc.)
  • Strength percentage

Summary line shows total pattern count and bullish/bearish breakdown.

Sidebar Stats

After each analysis, the sidebar updates with:

  • Symbol
  • Last price
  • Period change %
  • Total patterns detected
  • Overall signal (Bullish / Bearish / Neutral)

Architecture

cryptvault/desktop/
    app.py              entry point — creates and runs MainWindow
    main_window.py      main window, topbar, sidebar, analysis thread
    theme.py            dark color palette, fonts, ttk style definitions
    panels/
        chart_panel.py      Matplotlib candlestick chart + pattern drawing
        analysis_panel.py   ML prediction card + pattern list panel

Data Flow

  1. User enters symbol and clicks Analyze
  2. MainWindow._run_analysis() launches a background thread
  3. Thread fetches OHLCV data via yfinance
  4. ComprehensivePatternDetector.detect_all() runs all pattern detectors
  5. ProductionPredictor.predict() runs the ML ensemble
  6. Results are queued back to the main thread via root.after()
  7. ChartPanel.update_chart() and AnalysisPanel.update() redraw the UI

Supported Symbols

Any symbol supported by Yahoo Finance:

  • Crypto: BTC-USD, ETH-USD, SOL-USD, BNB-USD, XRP-USD
  • Stocks: AAPL, TSLA, GOOGL, MSFT
  • Indices: ^GSPC, ^DJI