This PR tries to resolve issues described in: https://github.com/nmwsharp/polyscope/issues/133, https://github.com/nmwsharp/polyscope/issues/135.

add a callback function to the render::engine and its derivatives to allow external ImGui font setting when using polyscope as a library.

One issue that still exists is the globalFontAtlas variable cannot be set from the external function, though it looks that variable does not affect anything. But an additional step was added to make sure that variable is set after the font setting.

It looks like the ribbon shader has been disabled.. Checking/unchecking the "ribbon" checkbox in the interface does nothing (vectors are still displayed with the "arrow" geometry).

Resolves #52 Culling and identical rendering for backfaces was already implemented, previously the "different" option would just render a slightly darker version of the same material, I changed that to a color option that defaults to the inverse of the first material color.

Hi,

I just tried polyscope within python on Windows using the ipython terminal, jupyter notebook and jupyter lab. In all cases, polyscope works well and is quite impressive. But when I try to close the polyscope window, it always crashes python. Are there any settings I can change to address this? I am assuming it is related to the opengl backend, yes?

Cheers, Michael

I've tried the basic point cloud visualization example from the docs for python:

import numpy as np
import polyscope as ps
ps.init()

points = np.random.rand(100, 3)
ps_cloud = ps.register_point_cloud("my points", points)
ps.show()

However, the window that is created looks like this. I've also tried to adjust the radius, but this doesn't change anything.

On the console I get the following output:

[polyscope] Backend: openGL3_glfw -- Loaded openGL version: 3.3 (Core Profile) Mesa 19.2.8

I've installed via the conda command for python 3.6.

When using Polyscope within Python on OSX, the window pops up behind the currently used windows and has to be focused in order to see it. Add and use a focusWindow method so that it shows at the front when the windows is created.

• add a focusWindow method
• use focusWindow when a window is created

Also, there were some minor clang-format changes when editing these files. I figured they should be included since the changes were made based on the config .clang-format file in this repository.

Hi,

I wrote a minimal script to describe the issue I am experiencing. From a high level, if I iterate through a series of objects to render, some type of meta-data is stored, and the results are different depending on what order the objects are rendered in. Even though I set ps.set_use_prefs_file(False) and ps.set_always_redraw(True) for good measure, the results still vary.

For example, below, I show how setting the 'z_up' direction once, and then setting it to 'y_up' will result in the same output. If you change the order, the outputs will be different than previously, but still each render is identical. Any clues as to why this is happening?

import numpy as np


import polyscope as ps
ps.set_use_prefs_file(False)
ps.set_always_redraw(True)
ps.init()


np.random.seed(1)
points = np.random.rand(4, 3)
vals = np.random.rand(4, 3)


ps.set_up_dir('z_up')
ps_cloud = ps.register_point_cloud('points', points, enabled=True, radius=0.08)
ps_cloud.add_color_quantity("rand colors", vals, enabled=True)
ps.set_navigation_style("planar")
ps.screenshot(filename='/home/rana/code/sandbox/points_z.png')

ps.set_up_dir('y_up')
ps_cloud = ps.register_point_cloud('points', points, enabled=True, radius=0.08)
ps_cloud.add_color_quantity("rand colors", vals, enabled=True)
ps.set_navigation_style("planar")
ps.screenshot(filename='/home/rana/code/sandbox/points_y.png')

Resolves #106 Addes a checkbox under Camera Parameters which swaps the perspective matrix with an orthogonal one. Also remaps scroll zoom to scale the camera bounds.

this is a quick hack to be able to use a scalar attribute on a point set to define the radii of the points. This was usefull to me to visualize sets of balls. In its current state it is not possible to use another scalar attribute for colors.

It seems like some machines which claim to support openGL 3.3 core are erroring out on the multisample textures. Perhaps we accidentally used something which is not valid in core 3.3?

First of all, thanks for the great framework!

I noticed that the version of the ImGui dependency is rather dated (April 2020) and incompatible with some awesome libraries like ImPlot (https://github.com/epezent/implot).

So, this PR updates the ImGui submodule from v1.76 to v1.86.

These are the "major" changes:

1. ImGui backends are now located under imgui/backends instead of imgui/examples
2. ImGuiCol_ModalWindowDarkening was renamed to ImGuiCol_ModalWindowDimBg
3. The Columns API is deprecated but still functional in the file imgui_tables.cpp
4. The power argument to ImGui::SliderFloat/ImGui::DragFloat has been replaced by a flags argument (see https://github.com/ocornut/imgui/issues/3361). I replaced all appearances of power > 1 with the flags ImGuiSliderFlags_Logarithmic | ImGuiSliderFlags_NoRoundToFormat, which I think results in natural behavior for the logarithmic sliders. Feel free to double check.

All tests are passing on my machine.

This change will ensure proper behavior is exhibited when the CMAKE_INSTALL_PREFIX is specified by the user during build configuration so that artifacts are installed to their desired location.

Desired workflow: The user programmatically configures several "views" of their data, each of which corresponds to a different collection of settings (stuff which is enabled, etc). The user wants to quickly switch between these views without clicking a bunch.

How to get there: We could get this fairly easily by leveraging the persistent value caches. Work items would be:

• define a serialization format for persistent values (in-memory for now)
• add a storeConfiguration() function or something, which saves the state of all persistent values to an internal list
• add a loadConfiguration() function which restores persistent values
• buttons in the GUI, etc

This has significant overlap with things we would want to do regardless to enable serialization more broadly.

I'm trying to use Polyscope on a Virtual Machine (Ubuntu 22.04 running inside an Oracle Virtualbox on Windows 10) However on calling it from Jupyter-lab I get the following problem:

Kernel Restarting

The kernel for Documents/ROTATOR/geodesic/Barmesh-scratch.ipynb appears to have died. It will restart automatically. So I tried doing a simple test file to just load a mesh and open a Polyscope window to show it, running from command line I get this:

(base) timbo@timbo-VirtualBox:~/Documents/ROTATOR/geodesic$ python3 test.py GLFW emitted error: GLX: Failed to create context: GLXBadFBConfig GLFW emitted error: Cannot set swap interval without a current OpenGL or OpenGL ES context Segmentation fault (core dumped) I guess it's some issue with the display being passed through the VM, any idea how to fix this?

Hi, I would like to have a "drap-n-drop to open file" feature and was wondering if this is sth that should be embedded into the viewer.

On ImGui's page, I found this: https://github.com/ocornut/imgui/issues/2602

But I'm not sure how this should be integrated into polyscope to be multi-platform compatibable.

Right now, the only way to implement an animation is via userCallback. However, this requires you to put your whole program control flow inside of the callback.

We should offer another option where a user calls a function like polyscope::mainLoopIteration() within their own control flow. The user is responsible for calling it often enough to keep the UI responsive.