CSG Algorithm Prototyping¶
“Production” code¶
Production code meaning that it is used from elsewhere, and thus cannot be changed without regard to its users.
- csg.py
- python CSG description, with serialization that is deserialized by npy/NCSG, used by tboolean-
- glm.py
- translation into python of some GLM matrix transform manipulations
- GParts.py
- indexed access to the partlist of a primitive for oldstyle CGS_FLAGPARTLIST
Development Exploration code¶
Ulyanov Excursion into Heart of Darkness¶
Failed attempt to implement pseudocode from a computer science paper that purported to implement iterative CSG.
- boolean.py
- bugged boolean tables
- csgtree.py
- heart of darkness
Potentially Useful machinery developed whilst debugging in the Heart of Darkness¶
- intersect.py
- numpy based CPU side handling of lots of intersects, some potentially useful techniques that came out of the Heart of Darkness
- intersectTest.py
- 2d matplotlib plotting of intersects
- ray.py
- numpy based collections of rays, used for CPU side debugging of CSG algorithms (2d intersections used with matplotlib plotting of intersects)
- nodeRenderer.py
- matplotlib based rendering of geometry in 2d, used for intersect debugging
Learning : binary tree gymnastics¶
- stacktrav.py
- basics of binary tree traversal in preorder/postorder/inorder/levelorder
- bintree.py
- roundtrip binary tree serialization/deserialization testing, and multi binary tree concatenation/de-concatentation testing
- node.py
- binary tree gymnastics and serialization dev
- postOrderIterative.py
- alt iterative approach
- postOrderTraversalWithoutRecursion.py
- exploring how to traverse binary trees without using recursion
Learning : Morton magic bit twiddling¶
- morton.py
- playing with morton 2d codes
- zorder.py
- quadtree exploration, 2d morton codes, multires z-order curve plotting
Learning : Converting Recursive to Iterative¶
- factorial.py
- very simple example of converting recursive into iterative
- hilbert.py
- more realistic example of converting recursive into iterative, with Hilbert curve plotting with matplotlib
GPU CSG Design¶
- iterativeExpression.py
- Provided the seed for GPU CSG implementation, with the realization that externalizing the postorder traversal makes binary tree evaluation very simple and tractable without recursion. Also the direct parallels between CSG node tree evaluation to binary expression evaluation was realized.
- iterativeTreeEvaluation.py
- starting point for the design of GPU CSG, handling the node traversal and reiteration techniques, lots of design notes : Heart of Design
- ctrl.py
- enum used by iterativeTreeEvaluation*
- iterativeTreeEvaluationFake.py
- comparing iterative and recursive imps
- iterativeTreeEvaluationCSG.py
- applying the nascent algorithm
- postorder.py
devise bit twiddling postorder for binary trees, inspired by 1d Morton codes, postorder_next allows doing bare naked postorder traverses of serialized binary trees
- without reconstructing the node tree
- without using recursion
- without storing postorder sequence indices
- slavish.py
- slavish translation of CUDA CSG ray trace back to python, used for debugging