Documentation of QGIS Python plugin brdrQ: FeatureAligner
Instructions

- Choose the settings for the alignment (reference layer,…)
- Select the thematic layer you want to align.
- Select a feature from the list or via ‘Select feature(s) on map’.
- View the prediction(s) for this aligned feature.
Additionally, you can:
- Switch between multiple predictions (via list, slider, or spinbox).
- Plot: Request a ‘plot’ of these predictions (relevant distance (m) vs. change (m²)).
- Visualize: Visualize the predicted relevant distances side by side.
- Save Geometry: Adjust the original geometry to the chosen prediction.
- Reset Geometry: Reset the original geometry (only within a feature session, so if no other feature is selected).
Parameter Guide
Reference layer
- Definition: Target dataset used as geometric truth.
- Why use it: Determines what ?aligned? means.
- Choices: Local/project reference or on-the-fly sources.
- Impact: Reference quality controls final alignment quality.
Open Domain Strategy
- Definition: Handling rule for geometry parts outside reference coverage.
- Why use it: Reflects policy on non-covered area retention.
- Choices: Exclude/keep/snap variants.
- Impact: Changes whether external areas are removed, kept, or reshaped.
Processor
- Definition: Algorithm backend selector.
- Why use it: Optimizes runtime by geometry type.
- Choices: AlignerGeometryProcessor recommended.
- Impact: Better defaults improve speed and consistency.
Threshold value (%)
- Definition: Fallback threshold in uncertain relevance cases.
- Why use it: Prevents unstable candidate acceptance.
- Choices: usually around 50, adapt per dataset.
- Impact: Higher means stricter matching.
Maximal relevant distance (m)
- Definition: Maximum deviation searched for predictions.
- Why use it: Controls correction aggressiveness.
- Choices: low/medium/high according to data drift.
- Impact: Larger values increase candidate range and ambiguity.
Add brdr_metadata?
- Definition: Adds rdr_metadata to results.
- Why use it: Preserves lineage for downstream updates/audits.
- Choices: enabled/disabled.
- Impact: Better traceability with small storage overhead.
Full Reference Strategy
- Definition: Preference for full-overlap candidates.
- Why use it: Raise confidence in selected predictions.
- Choices: ONLY_FULL_REFERENCE, PREFER_FULL_REFERENCE, NO_FULL_REFERENCE.
- Impact: Stricter mode improves certainty but may reduce alternatives.
Snap Strategy
- Definition: Snap strictness to reference vertices (mainly line/point workflows).
- Why use it: Controls structural precision at vertices/endpoints.
- Choices: NO_PREFERENCE, PREFER_VERTICES, PREFER_ENDS_AND_ANGLES, ONLY_VERTICES.
- Impact: Stricter snapping improves topological control but can reduce feasible matches.
Recommended Presets
- Conservative Editing: low distance + PREFER_FULL_REFERENCE.
- Balanced Daily Use: medium distance + default processor + PREFER_VERTICES.
- Strong Recovery for Rough Data: higher distance + permissive full-reference mode.
- Strict Network Snapping: ONLY_VERTICES + low distance.