dsparrow27 / slither

Crude implementation of a parallel execution engine in python but without dirty propagation , meant to be used via commandline. Currently we lock up maya when we use the gui due to maya thinking we want a separate shell which boots mayapy separately, will need to rework this to be single process when running in the GUI but works fine via commandline, yet to be tested over a farm system.

:note:: This Repo is still WIP

TODO

1. unittest
2. maya processing nodes(I/O)
3. 3dsmax nodes
4. on demand dirty propagation
5. documentation
6. error handling within nodes
7. Skip attribute and node initialization if a node fails to initialize
8. up front graph expansion since by default compounds we lazily expand only when we hit a compound not up front.
9. example of creating a graph on the fly.
10. deserialize a graph from json

Environment variables

These variables store the node and data type library, you can add your own directory paths once slither is initialized SLITHER_NODE_LIB SLITHER_TYPE_LIB

Supports

Compound nodes eg. nesting nodes within nodes parallel execution array attributes compound attributes node and datatype library extendability Graph mutation

DEPENDENCIES

Blinker ffmpeg-python ffmpeg

EXAMPLES

import os
import sys
from slither import api
from slither.core import executor

app = api.initialize()
root = app.root

testSum = app.createNode("testNode", type_="Sum")

executor.StandardExecutor().execute(root)
# pprint the graph
pprint.pprint(root.serialize())
# get the final result for the graph
print root.execution.value()