Voxelium 0.0.5

Voxelium - a powerful tool for volumetric processing with CUDA support.

Voxelium Alpha Testing

This repository currently contains the alpha version of the Voxelium library.

Installation

After you've cloned the repo and cd into the project directory you first need to set up the proper Conda environment. Use the environment.yml file to create a new environment called 'voxelium' with the right module installed, by running:

conda env create -f environment.yml

Visualization-Only Installation

If you only need to visualize reconstruction results (e.g. on you local computer) you can skip the building of the torch extensions. These are only needed on the computational nodes. First activate the new Conda environment:

conda activate voxelium

You can now install the voxelium library from inside the project directory by running:

VOXELIUM_SKIP_EXT=TRUE pip3 install .

In the above, VOXELIUM_SKIP_EXT will skip installation of the torch extensions.

Full Installation

If you need to run reconstruction (e.g. on a computational node), you need to build and install the torch extensions. You will need to have a CUDA toolkit installed for this that matches the pytorch version installed. Once you have that ready you can just run:

pip3 install .

3D Spectral Heterogeneity Analysis (SHA)

Activate the voxelium conda environment. Then run voxelium -h to see a list of modules. To run the analysis, the sha3D module can be run as follows:

voxelium SHA3D <input_star_data> <log_directory> --gpu 0

Here, <input_star_data> is an input STAR-file containing all the particles with CTF and pose parameters set. <log_directory> will contain the results of the job.

NOTE: Adding --preload speeds things up considerably, assuming the dataset fits in memory.

NOTE: You need to install extension for this, see above.

SHA3D Visualization

To visualize the results run:

voxelium sha3D_viwer <log_directory>

In the above, <log_directory> is the path to the directory containing the results of the SHA3D analysis, see above.

Troubleshoot

If there's an issue with the CUDA environment it can help preventing PYPI building in an isolated environment by running

pip install --no-build-isolation .

🤝 Contributor covenant code of conduct

This project adheres to the Contributor Covenant code of conduct. By participating, you are expected to uphold this code. Please report unacceptable behavior to [e-mail].

Responsible Use: We are committed to advancing the responsible development and use of artificial intelligence. Please follow our Acceptable Use Policy when engaging with the model.

🔒 Security

If you believe you have found a security issue, please responsibly disclose by contacting us at [e-mail].