Docs: rewrite README.md for complete PySide6 transition

This commit is contained in:
lukas 2026-03-10 16:56:01 +01:00
parent 551f5a6b8f
commit 635b65b7e1
1 changed files with 29 additions and 23 deletions

View File

@ -1,24 +1,29 @@
<div style="display:flex; gap:16px; align-items:center;">
<img src="app/assets/logo.png" alt="ICRA" width="140"/>
<div>
<strong>Interactive Color Range Analyzer</strong> is being reimagined with a <em>PySide6</em> user interface.<br/>
This branch focuses on building a native desktop shell with modern window controls before porting the color-analysis features.
<strong>Interactive Color Range Analyzer (ICRA)</strong><br/>
A fully-featured, PySide6-powered desktop application for precise color matching, image analysis, and statistics generation.
</div>
</div>
## Current prototype
- Custom frameless window with minimise / maximise / close controls that hook into Windows natively
- Dark themed layout and basic image preview powered by Qt
- “Open image” workflow that displays the selected asset scaled to the viewport
> ⚠️ Legacy Tk features (sliders, exclusions, folder navigation, stats) are not wired up yet. The goal here is to validate the PySide6 shell first.
## Features
- **High-Performance Image Processing:** Native, vectorized NumPy operations for lightning-fast HSV conversion and color matching.
- **Batch Folder Processing:** Load a folder of images and instantly export an aggregated `icra_stats.csv` with localized formatting (dot or comma decimals depending on your language).
- **Eyedropper Tool:** Quickly pick target matching colors directly from the image canvas.
- **Advanced Selection:** Support for exclusion zones (rectangles and free-draw polygons) overlaid on the image, all dynamically rendered via `QGraphicsView`.
- **Modern UI & UX:**
- Drag-and-drop support for files and folders.
- Custom dark and light themes with native-feeling borderless titlebars and a categorized menu bar.
- Window size and position persistence between launches.
- Keyboard shortcuts for rapid workflow.
- **Configurable:** Uses `config.toml` to drive everything from overlay matching colors to default sliders and application language (`en`/`de`).
## Requirements
- Python 3.11+
- [uv](https://github.com/astral-sh/uv) for dependency management
- Windows 10/11 recommended (PySide6 build included; Linux/macOS should work but are untested in this branch)
- Works across Windows, macOS, and Linux (requires PySide6 and numpy)
## Setup with uv (PowerShell example)
## Setup with uv
```bash
git clone https://git.lukasmahler.de/lm/ICRA.git
cd ICRA
@ -28,24 +33,25 @@ uv pip install .
uv run icra
```
The app launches directly as a PySide6 GUI—no browser or local web server involved. Use the “Open Image…” button to load a file and test resizing/snap behaviour.
## Running the Test Suite
The core image processing logic and UI data models are rigorously tested using `pytest`.
```bash
uv run pytest tests/ -v
```
## Roadmap (branch scope)
1. Port hue/saturation/value controls to Qt widgets
2. Re-implement exclusion drawing using QPainter overlays
3. Integrate existing image-processing logic (`app/logic`) with the new UI
## Project layout
## Project Layout
```
app/
assets/ # Shared branding
gui/, logic/ # Legacy Tk code kept for reference
qt/ # New PySide6 implementation (main_window, app bootstrap)
config.toml # Historical defaults (unused in the prototype)
lang/ # Localization strings (TOML format)
logic/ # Configuration loading and application constants
qt/ # PySide6 implementation (main_window, custom UI widgets, vectorized image processing)
tests/ # Pytest unit tests
config.toml # Configurable overlay colors, slider defaults
main.py # Entry point -> PySide6 launcher
```
## Development notes
## Development Notes
- The legacy `Tkinter` codebase has been completely removed in favor of `PySide6`.
- Quick syntax check: `uv run python -m compileall app main.py`
- Uploaded images are not persisted; the preview uses Qt pixmaps only.
- Contributions welcome—please target this branch with PySide6-specific improvements.
- Contributions welcome!