pyhapi

Object-Oriented Python Wrapper for Houdini Engine's C API

Documentation

Please visit pyhapi documentation.

Install

You could use either (a) pip or (b) clone and install.

option a.1 setup PATH

• For Windows, run
  powershell ./sethoupath.ps1
  it should add houdini's dsolib and bin directory to PATH.
• For Linux
  Add /opt/hfs<version>/dsolib/ to LD_LIBRARY_PATH
  Example:
  export LD_LIBRARY_PATH=${LD_LIBRARY_PATH}:/opt/hfs17.5/dsolib/

option a.2 install through PyPI

$ pip install pyhapi

option b

Clone this repo and run python setup.py install

Supported Platforms

• Windows
• Linux

Dependency

• Python: >3.6
• Numpy: >= 1.15
• Houdini >= 17.5
  • License supported:
    • Houdini FX
    • Houdini Core
    • Houdini Engine
  • License NOT supported:
    • Houdini Engine Indie
    • Houdini Indie
    • Houdini Apprentice

Features

• Instantiate node/HDA
• Node connect operation
• Node parameter get/set
• Node async cooking
• Marshall in/out curve
• Marshall in/out mesh
• Marshall in/out heightfield
• SessionPool and task-based processing

Following feature in Houdini Engine is NOT supported yet:

• Marshall in/out volume
• PDG execution

Example Usage

Please see documentation for detailed tutorial. ****

Create a session

import pyhapi as ph
session = ph.HSessionManager.get_or_create_default_session()

See more in
Example: node_networks_operations

Instantiate a HDA

hda_asset = ph.HAsset(session, "hda/FourShapes.hda")
asset_node = hda_asset.instantiate(node_name="TestObject").cook()

Set parameter of node

asset_node.set_param_value("seed", 1.0)
asset_node.set_param_value("foo_attrib", "foo_str")
asset_node.press_button("foo_execute")

See more in
Example: hda_params_getset

Save HIP file

session.save_hip("debug.hip")

Marshall data

see more in
Example: curve_marshall_input
Example: curve_marshall_output
Example: mesh_marshall_input
Example: mesh_marshall_output
Example: heightfield_marshal_input
Example: heightfield_marshal_output

Session Pool

Producer/Consumer

session_pool = ph.HSessionManager.get_or_create_session_pool()

session_pool.run_on_task_producer(producer)

Example: multisession_producer

Batch Run

session_pool = ph.HSessionManager.get_or_create_session_pool()
    
session_pool.enqueue_task(session_task, i, j)
    
session_pool.run_all_tasks()

Example: multisession_run

Threaded Task Producer

session_pool = ph.HSessionManager.get_or_create_session_pool()
    
session_pool.run_task_consumer_on_background()

executor = ThreadPoolExecutor(max_workers=4)
for i in range(0,4):
    executor.submit(producer, i)

Example: multisession_run

Flask Server Demo

Example: multisession_run