Importing Processing libraries

[tabreturn]

How would I go about using Processing Java Libraries?

[hx2A]

Which Processing library would you like to use? Not all of them can be compatible, and not all of them are a good idea to use. For example, there are plenty of good Python physics libraries that can do what Box2D can do. Also, the Processing libraries will not be able to make arbitrary calls to your Python code. Libraries that need to do that won't work.

Here's an example of how you can use the Processing library Camera3D:

https://github.com/hx2A/py5examples/blob/main/examples/Using%20Camera3D.ipynb

The basic idea is to first put the Jars in a jars subdirectory (relative to Python's current working directory). If the Jar file is located somewhere else, use py5_tools to add your jar file's location to the classpath. py5_tools provides tools for getting the classpath or making changes to it. Note that when you import py5 it will also start the JVM. You need to make changes to the classpath before importing it.

Next you need to use JClass to make the Java class available to you in Python, as you see in the example. You can create an instance and pass it the Sketch object like this:

camera3D = Camera3D(py5.get_current_sketch())

Then you can continue on to use the Processing library to enhance your Sketch.

So far I know that PeasyCam, ColorBlindness, and Camera3D all work well with py5. (I created the latter 2, so I made sure they worked)

[tabreturn]

I had to install libffi6 on my system (the latest Ubuntu-based Mint distro) using an old (deb) package from 2018. The 0.4 release was fine using libffi7. Anyhow, I've got both libffi6 and libffi7 on my system now, but I suspect that many Linux users won't be able to install libffi6 directly from their system repos.

[hx2A]

Can you tell me what kind of Python environment you are using? Anaconda? What error or problem arose that prompted you to install the older library?

I did fiddle with the package dependencies because some imports were missing. I didn't downgrade anything though. I'm not sure what the problem is.

[tabreturn]

Environment: Python 3.9.5 (which is built into Thonny 3.3.7). Error message --

%Run /home/nuc/Apps/thonny_py5_0.5/thonny/lib/python3.9/site-packages/py5_tools/tools/run_sketch.py py5.py
Traceback (most recent call last):
  File "/home/nuc/Apps/thonny_py5_0.5/thonny/lib/python3.9/site-packages/py5_tools/tools/run_sketch.py", line 22, in <module>
    from py5_tools import imported
  File "/home/nuc/Apps/thonny_py5_0.5/thonny/lib/python3.9/site-packages/py5_tools/__init__.py", line 23, in <module>
    from .imported import set_imported_mode, get_imported_mode  # noqa
  File "/home/nuc/Apps/thonny_py5_0.5/thonny/lib/python3.9/site-packages/py5_tools/imported.py", line 28, in <module>
    from .py5bot import py5bot
  File "/home/nuc/Apps/thonny_py5_0.5/thonny/lib/python3.9/site-packages/py5_tools/py5bot/__init__.py", line 20, in <module>
    from .kernel import Py5BotKernel  # noqa
  File "/home/nuc/Apps/thonny_py5_0.5/thonny/lib/python3.9/site-packages/py5_tools/py5bot/kernel.py", line 22, in <module>
    from ipykernel.zmqshell import ZMQInteractiveShell
  File "/home/nuc/Apps/thonny_py5_0.5/thonny/lib/python3.9/site-packages/ipykernel/__init__.py", line 2, in <module>
    from .connect import *
  File "/home/nuc/Apps/thonny_py5_0.5/thonny/lib/python3.9/site-packages/ipykernel/connect.py", line 12, in <module>
    import jupyter_client
  File "/home/nuc/Apps/thonny_py5_0.5/thonny/lib/python3.9/site-packages/jupyter_client/__init__.py", line 4, in <module>
    from .connect import *
  File "/home/nuc/Apps/thonny_py5_0.5/thonny/lib/python3.9/site-packages/jupyter_client/connect.py", line 21, in <module>
    import zmq
  File "/home/nuc/Apps/thonny_py5_0.5/thonny/lib/python3.9/site-packages/zmq/__init__.py", line 61, in <module>
    _load_libzmq()
  File "/home/nuc/Apps/thonny_py5_0.5/thonny/lib/python3.9/site-packages/zmq/__init__.py", line 23, in _load_libzmq
    import ctypes
  File "/home/nuc/Apps/thonny_py5_0.5/thonny/lib/python3.9/ctypes/__init__.py", line 8, in <module>
    from _ctypes import Union, Structure, Array
ImportError: libffi.so.6: cannot open shared object file: No such file or directory

I installed libffi6 (an older package for Ubuntu 2018.04) on my system and that solved the problem.

[hx2A]

Well, this is going to be a puzzle to solve.

Can you export your environment so I can see all of the package versions that are installed? ipykernel is one of the ones I bumped. Actually, I forgot to add it as a dependency before this weekend, but when I added it I first required >= 5.5. That caused problems on a Mac computer so I dropped it to >= 5.3 and then it seemed to be OK.

[tabreturn]

Cool, here you go -- a list for my v0.5 environment:

argon2-cffi         20.1.0
astroid             2.5.6
asttokens           2.0.5
async-generator     1.10
attrs               21.2.0
backcall            0.2.0
bcrypt              3.2.0
bitstring           3.1.7
bleach              3.3.1
certifi             2020.12.5
cffi                1.14.5
charset-normalizer  2.0.3
cryptography        3.4.7
debugpy             1.3.0
decorator           5.0.9
defusedxml          0.7.1
distro              1.5.0
docutils            0.17.1
ecdsa               0.16.1
entrypoints         0.3
esptool             3.0
idna                3.2
ipykernel           6.0.2
ipython             7.25.0
ipython-genutils    0.2.0
ipywidgets          7.6.3
isort               5.8.0
jedi                0.18.0
Jinja2              3.0.1
JPype1              1.3.0
jsonschema          3.2.0
jupyter-client      6.1.12
jupyter-core        4.7.1
jupyterlab-pygments 0.1.2
jupyterlab-widgets  1.0.0
lazy-object-proxy   1.6.0
line-profiler       3.3.0
MarkupSafe          2.0.1
matplotlib-inline   0.1.2
mccabe              0.6.1
mistune             0.8.4
mypy                0.812
mypy-extensions     0.4.3
nbclient            0.5.3
nbconvert           6.1.0
nbformat            5.1.3
nest-asyncio        1.5.1
noise               1.2.2
notebook            6.4.0
nptyping            1.4.2
numpy               1.21.1
packaging           21.0
pandas              1.3.0
pandocfilters       1.4.3
paramiko            2.7.2
parso               0.8.2
pexpect             4.8.0
pickleshare         0.7.5
Pillow              8.3.1
pip                 21.1.1
prometheus-client   0.11.0
prompt-toolkit      3.0.19
ptyprocess          0.7.0
py5                 0.5a0
pycparser           2.20
Pygments            2.9.0
pylint              2.8.2
PyNaCl              1.4.0
pyparsing           2.4.7
pyrsistent          0.18.0
pyserial            3.5
python-dateutil     2.8.2
pytz                2021.1
pyzmq               22.1.0
reedsolo            1.5.4
requests            2.26.0
Send2Trash          1.5.0
setuptools          56.0.0
six                 1.15.0
stackprinter        0.2.5
terminado           0.10.1
testpath            0.5.0
thonny              3.3.7
toml                0.10.2
tornado             6.1
traitlets           5.0.5
typed-ast           1.4.3
typing-extensions   3.10.0.0
typish              1.9.2
urllib3             1.26.6
wcwidth             0.2.5
webencodings        0.5.1
websockets          9.0.1
wheel               0.36.2
widgetsnbextension  3.5.1
wrapt               1.12.1

But then I noticed that my py5 v0.4 environment doesn't include the Jupyter packages in v0.5, which I suspect are looking for libffi.so.6 --

astroid            2.5.6
asttokens          2.0.5
backcall           0.2.0
bcrypt             3.2.0
bitstring          3.1.7
certifi            2020.12.5
cffi               1.14.5
charset-normalizer 2.0.3
cryptography       3.4.7
decorator          5.0.9
distro             1.5.0
docutils           0.17.1
ecdsa              0.16.1
esptool            3.0
idna               3.2
ipython            7.25.0
ipython-genutils   0.2.0
isort              5.8.0
jedi               0.18.0
JPype1             1.3.0
lazy-object-proxy  1.6.0
line-profiler      3.3.0
matplotlib-inline  0.1.2
mccabe             0.6.1
mypy               0.812
mypy-extensions    0.4.3
noise              1.2.2
nptyping           1.4.2
numpy              1.21.1
pandas             1.3.0
paramiko           2.7.2
parso              0.8.2
pexpect            4.8.0
pickleshare        0.7.5
Pillow             8.3.1
pip                21.1.1
prompt-toolkit     3.0.19
ptyprocess         0.7.0
py5                0.4a2
pycparser          2.20
Pygments           2.9.0
pylint             2.8.2
PyNaCl             1.4.0
pyserial           3.5
python-dateutil    2.8.2
pytz               2021.1
reedsolo           1.5.4
requests           2.26.0
Send2Trash         1.5.0
setuptools         56.0.0
six                1.15.0
stackprinter       0.2.5
thonny             3.3.7
toml               0.10.2
traitlets          5.0.5
typed-ast          1.4.3
typing-extensions  3.10.0.0
typish             1.9.2
urllib3            1.26.6
wcwidth            0.2.5
websockets         9.0.1
wheel              0.36.2
wrapt              1.12.1

[hx2A]

Here are the relevant differences:

jupyter-core        4.7.1
bleach              3.3.1
jupyterlab-pygments 0.1.2
widgetsnbextension  3.5.1
jsonschema          3.2.0
entrypoints         0.3
Jinja2              3.0.1
jupyterlab-widgets  1.0.0
pyparsing           2.4.7
terminado           0.10.1
pandocfilters       1.4.3
argon2-cffi         20.1.0
ipywidgets          7.6.3
nbformat            5.1.3
tornado             6.1
defusedxml          0.7.1
jupyter-client      6.1.12
nbclient            0.5.3
MarkupSafe          2.0.1
ipykernel           6.0.2
nbconvert           6.1.0
async-generator     1.10
packaging           21.0
mistune             0.8.4
notebook            6.4.0
pyrsistent          0.18.0
attrs               21.2.0
pyzmq               22.1.0
debugpy             1.3.0
testpath            0.5.0
nest-asyncio        1.5.1
prometheus-client   0.11.0
webencodings        0.5.1

Did you not try to use the py5 kernel in the previous release? I added the Jupyter related dependencies for the 0.5 release because the two Jupyter kernels both depend on them. The py5 kernel wouldn't have worked in your 0.4 setup.

I think perhaps that your 0.4 setup didn't care about libffi.so at all but the 0.5 setup does, and something was compiled against the older version of the library. I did some searching online and it seems this issue is something other libraries are dealing with also.

Also, you are using a virttual environment here, right? Not Anaconda? I believe an Anaconda environment would make it easy to install the right version of that shared library.

[hx2A]

I just checked my Anaconda environments and I see my py5 environments have libffi.so.7 in them but not libffi.so.6. Quite a few of the environments for other projects have both, however. So Anaconda seems to take care of this issue for users.

[tabreturn]

Yes, I'm not using Anaconda. I suspect this is something that'll get resolved over time, that won't affect all users, and we have a solution for now 👍

[hx2A]

OK, cool.

Since I did a bad job on the release earlier this week I'd like to do another one fairly soon. I believe I can get the hex color thing done relatively quickly (and correctly), so I'll include it in the next release.

[hx2A]

And separately, I noticed you installed a bad version of jedi:

jedi 0.18.0

For that one the tab completion does not work correctly from Jupyter. I recommend installing 0.17.2 instead.

[jeremydouglass]

Re: using

Camera3D = py5.JClass('camera3D.Camera3D')

to import a Java class -> Python class -- have you considered using JImport instead? It appears to support the syntax from <java class> import <Class> as <Classname>

https://jpype.readthedocs.io/en/devel/imports.html

...and it looks like the behavior is customizable (although I haven't used JPype, I'm just looking through the docs.

[hx2A]

Hi @jeremydouglass

Thanks for pointing out this feature. It does look like it can further simplify the bridge between Python and Java. It must be a new feature because it isn't available in 1.2.1, which is the version I have in front of me right now. The 1.3 version came out somewhat recently and I haven't upgraded my dev environment to use it yet.

Jpype is an awesome library with lots of great features, and I am doing my best to leverage them to add value to py5.

[hx2A]

@jeremydouglass this will be in the next release. I just had to add import jpype.imports to the py5 library, and now instead of writing this

Camera3D = py5.JClass('camera3D.Camera3D')

I can write this:

from camera3D import Camera3D

Much cleaner.

[villares]

I love PeasyCam, so I have to take the time to figure out the p5_tools.add_jars() thing, I'm following the Camera3D example, but no success yet...

[hx2A]

I love PeasyCam, so I have to take the time to figure out the p5_tools.add_jars() thing, I'm following the Camera3D example, but no success yet...

I love PeasyCam also!

Are you using Thonny?

Try putting the PeasyCam jar in a jars subdirectory. That is easier than using add_jars(), but that is also a good option if the jars subdirectory is not working for you. Then use get_classpath to see what the classpath is. It will show you the jars that it knows about. Also, note that if you are calling add_jars() you must do so before import py5. Since Thonny uses run_sketch.py, which does the import for you before any of your code gets executed, you can't use add_jars(). The jars subdirectory must be relative to the current working directory (os.getcwd()).