Home

fungeom

Functional geometry as an immutable, decidable resolver graph.

fungeom lets you describe geometry — points, vectors, frames, transforms, time-signals, marker clouds, regions — as a lazy, immutable graph you can reason about before you compute it. You compose a result, ask whether it can be resolved, and when it can't you get back a reason, not an exception or a silent NaN.

The one big idea: partiality is first-class. A geometric question with no answer (a point in a frame that was never placed, a direction from a zero-length vector, a marker occluded mid-capture) is an honest Unresolvable carrying an explanation — and that explanation propagates through everything built on top of it.

from fungeom import Point3, Frame, Resolvable, Unresolvable

gripper = Frame.detached("gripper")          # a sub-assembly, not yet placed
tip = Point3.at(0, 0, 0.1, frame=gripper)    # built lazily

match tip.decide():
    case Resolvable(point):  print(point.coord)
    case Unresolvable(why):  print(why)      # "frame 'gripper' is not grounded to the world"

Start here

• Core Concepts — resolvers, decide() vs resolve(), Resolvable / Unresolvable, and why everything (even a scalar) is a node.
• Decidability & Partiality — the thesis: proving resolvability, carrying reasons, and how partiality propagates.
• The Primitive Layers — a map of the surface: geometry, logic, time, signals, collections, regions.
• Guides — task-oriented walkthroughs (signals over time, marker clouds, regions & contact) and the runnable examples.
• Adding a Primitive — the contributor's procedure and the definition of done.

Reference

The exhaustive combinator table (every constructor, every op, and its exact partiality), the architecture, and the design rationale live in docs/reference.md. Deep design docs for individual layers are in docs/.