Compass Python Lib

Conversion commands:

# Install in dev mod
pip install -e ".[dev,test]"

# Install latest stable version
pip install compass_lib

# run some commands
compass convert --input_file=./tests/artifacts/fulford.dat  --output_file=fulford.json --format=json --overwrite
compass convert --input_file=./tests/artifacts/random.dat  --output_file=random.json --format=json --overwrite

GeoJSON section colors

GeoJSON survey legs include properties.color, an opaque CSS #rrggbb color. Every shot in a DAT survey block receives the same generated section color. Sections sharing a station have different colors, including across MAK-linked DAT files. Names alone do not identify sections or station connections: the existing MAK station-scoping rules remain authoritative. Geometric crossings without a shared station do not create a color constraint.

The palette lives in compass_lib/constants.py as SURVEY_COLORS: vivid red, orange, yellow, lime, green, mint, cyan, blue, purple, and hot pink. Both palette colors and generated overflow colors have HSV saturation of at least 80% and value of at least 90%, excluding black, white, gray, and muted shades.

Assignment uses ordinary random choices in at most 16 iterative passes, without hashing, a fixed seed, or algorithm versioning. It stops on completion or no improvement. Remaining sections receive distinct vibrant colors using a finite fallback budget sized to the input, so reaching the pass limit does not leave touching sections with the same viewer fallback color. Colors can change when GeoJSON is regenerated and remain fixed in each stored artifact.

See GeoJSON section coloring for pass semantics, fallback bounds, the color-space limit, export behavior, and test coverage.

CompassSurvey.color is derived metadata assigned during coordinate computation; it is excluded from source JSON and DAT/MAK serialization. SurveyLeg.color carries it through coordinate propagation and adjustment. Passage polygons and clipped passage fragments inherit their leg's color. Station/anchor and misclosure styling retain their existing semantics.

The color property is independent of color_by_origin and CLI --no-colors, which still control origin-based simplestyle stroke/marker-color styling. Clients can use color for shot rendering and fall back to the survey color for legacy data without it. Existing saved GeoJSON must be regenerated to gain these properties; no source files are rewritten. Palette changes also require regenerating stored GeoJSON; existing artifacts retain the colors recorded when they were exported.

Portable regression coverage lives in tests/test_section_colors.py, including duplicate names, linked/scoped stations, crowded junctions, solver paths, round trips, CLI/API exports, vibrant colors, and safe budget exhaustion. Private fixtures remain additional coverage and are never required for these guarantees.