Installation

Installation & Usage in QGIS

LIRiAP can be installed and used in two ways in QGIS.

Option 1: As Script Folder (Quick Testing)

This method is useful for quick testing and experiments without enabling a full plugin.

Steps

1. Locate your QGIS scripts folder:

   • Windows: C:\Users\<username>\AppData\Roaming\QGIS\QGIS3\profiles\default\processing\scripts\
   • Linux: ~/.local/share/QGIS/QGIS3/profiles/default/processing/scripts/
   • macOS: ~/Library/Application Support/QGIS/QGIS3/profiles/default/processing/scripts/

2. Copy the algorithm pack:

   • Copy the LIRiAP_pack folder into the scripts folder

3. Restart QGIS (if already open)

4. Use the algorithms:

   • Open Processing Toolbox (Processing → Toolbox)
   • Search for "LIRiAP"
   • Algorithms appear under Scripts → LIRiAP

Notes

• This method doesn't require enabling a plugin
• Changes to scripts require QGIS restart to take effect
• Good for testing individual algorithms

---

Option 2: As Plugin Provider (Recommended)

This is the recommended method for regular use, as it provides better integration with QGIS.

Steps

1. Locate your QGIS plugins folder:

   • Windows: C:\Users\<username>\AppData\Roaming\QGIS\QGIS3\profiles\default\python\plugins\LIRiAP\
   • Linux: ~/.local/share/QGIS/QGIS3/profiles/default/python/plugins/LIRiAP/
   • macOS: ~/Library/Application Support/QGIS/QGIS3/profiles/default/python/plugins/LIRiAP/

2. Copy the plugin files: Create the LIRiAP folder and copy:

   • The LiRiAP_provider/ folder
   • The LIRiAP_pack/ folder

   Your folder structure should look like:

   LIRiAP/
   ├── LiRiAP_provider/
   │   ├── __init__.py
   │   ├── LIRiAP_plugin.py
   │   └── algorithms/
   ├── LIRiAP_pack/
   │   ├── *.py (algorithm and worker files)
   

3. Enable the plugin:

   • Open QGIS
   • Go to Plugins → Manage and Install Plugins
   • Find "LIRiAP" in the list
   • Check the box to enable it

4. Verify installation:

   • Open Processing Toolbox
   • Look for "LIRiAP" group in the algorithm list

---

Running an Algorithm

1. Open Processing Toolbox: Processing → Toolbox (or press Ctrl+Alt+T / Cmd+Alt+T)

2. Find the algorithm:

   • Expand the LIRiAP group, or
   • Search for specific algorithm name in the search box

3. Configure and run:

   • Double-click an algorithm (e.g., "Approximation Standard")
   • Select your Input layer (polygon layer)
   • Adjust parameters as needed:
     • GRID_COARSE / GRID_FINE: Search resolution (higher = more accurate, slower)
     • ANGLE_STEP: Fallback angle sweep step
     • MAX_RATIO: Maximum aspect ratio (0 = unlimited)
     • ALWAYS_RETURN: Enable best-effort fallback
     • USE_BUFFER: Apply containment margin
   • Click Run

4. View results:

   • Output layer is created in memory (or saved to file if specified)
   • Output fields include: area, angle, ratio, and algorithm-specific diagnostics

---

Dependencies

Required

• NumPy
• SciPy
• Shapely

These are typically included with QGIS or can be installed via:

pip install numpy scipy shapely

Optional: Numba (Recommended)

Numba provides JIT compilation that significantly speeds up computations.

Auto-install (in algorithm settings):

• Check AUTO_INSTALL_NUMBA parameter

Manual install:

pip install numba

Numba is optional but recommended for better performance, especially on larger datasets.

---

Troubleshooting

Algorithms not appearing in Processing Toolbox

Script folder method:

• Verify LIRiAP_pack folder is in the correct scripts location
• Check for Python errors in the log (View → Panels → Log Messages → Python)

Plugin method:

• Ensure the plugin is enabled in Plugins → Manage and Install Plugins
• Check Plugins → Plugin Manager → Settings → Show also experimental plugins
• Check for errors in Plugins → Install from ZIP is NOT needed for local installation

"Module not found" errors

Ensure all dependencies are installed:

pip install numpy scipy shapely numba

Performance issues

• Enable AUTO_INSTALL_NUMBA or install Numba manually
• Reduce GRID_FINE for faster (less accurate) results
• Use Approximation family for fastest results
• Use single worker (N_WORKERS=1) for BCRS family