pyhapi

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

Note that this is 3rd party binding, not official.

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