mlnanoshaper.jl
Science Score: 54.0%
This score indicates how likely this project is to be science-related based on various indicators:
-
✓CITATION.cff file
Found CITATION.cff file -
✓codemeta.json file
Found codemeta.json file -
✓.zenodo.json file
Found .zenodo.json file -
○DOI references
-
✓Academic publication links
Links to: zenodo.org -
○Academic email domains
-
○Institutional organization owner
-
○JOSS paper metadata
-
○Scientific vocabulary similarity
Low similarity (11.0%) to scientific vocabulary
Repository
Basic Info
- Host: GitHub
- Owner: concept-lab
- License: mit
- Language: Julia
- Default Branch: main
- Size: 6.04 MB
Statistics
- Stars: 0
- Watchers: 1
- Forks: 1
- Open Issues: 1
- Releases: 0
Metadata Files
README.md
MLNanoShaper
MLnanoshaper is an experimental sofware designed to compute proteins surface like the Nanoshaper program but with less computations using Machine Learning. Instead of having a probe (a virtual water molecule) that rolls over the surface, like in Nanoshaper, we will have a machine learning algorithme will generate a scalar "energy" field from the atoms. We will then define the surface as an implicit surface ie $\{x/ f(x)= 0\}$ that will be triangulated using marching cubes.
To use this project
- install julia
julia --project MLNanoShaperRunner/build/build.jl - clone this repository:
git clone https://github.com/concept-lab/MLNanoShaper.jl - clone the submodule
cd MLNanoShaper.jl git submodule init git submodule update - download dependency:
julia -e 'using Pkg; Pkg.activate(".");Pkg.instantiate()'# make dirsmkdir ~/datasets mkdir ~/datasets/models mkdir ~/datasets/logs mkdir ~/datasets/pqr# install datasetcurl https://zenodo.org/records/14886938/files/shrec.tar.gz --output ~/datasets/shrec.tar.gz tar -xzf ~/datasets/shrec.tar.gz -C ~/datasets/pqr
To train
install parallel and run commands
julia --project ./build/build.jl install
./scripts/training.bash
By default, logs and weights should be stored in ~/datasets/logs/ and ~/datasets/models
Notebooks
The ./scripts/ folder contains some Pluto Notebooks. Theses are for creating figures and for debugging. To execute these notebooks, you should install pluto and launch server with
julia -e "using Pluto; Pluto.run()"
Each notebooks will use the environement in scripts.
Interfaceing with C interface
We have an example of code in MLNanoShaperRunner/example. Here is the commands to compile the example.
Compiling the C interface.
julia --project MLNanoShaperRunner/build/build.jl
Compiling the C code
clang MLNanoShaperRunner/examples/dummy_example.c \
-I MLNanoShaperRunner/build/lib/include \
-L MLNanoShaperRunner/build/lib/lib/ \
-l MLNanoShaperRunner \
-o MLNanoShaperRunner/examples/test
Run the code
in the MLNanoShaperRunner/examples/directory
LD_LIBRARY_PATH=$(pwd)/../build/lib/lib ./test
`
Weights
weights are on the https://zenodo.org/records/16601766
Owner
- Name: CONCEPT Lab @ IIT
- Login: concept-lab
- Kind: organization
- Location: Genova, Italy
- Website: https://concept.iit.it/
- Repositories: 5
- Profile: https://github.com/concept-lab
Citation (CITATION.bib)
@misc{MLNanoShaper.jl,
author = {tristan <tristan.hacquard@polytechnique.org> and contributors},
title = {MLNanoShaper.jl},
url = {https://github.com/hack-hard/MLNanoShaper.jl},
version = {v0.1.0},
year = {2024},
month = {4}
}
GitHub Events
Total
- Release event: 1
- Delete event: 2
- Issue comment event: 1
- Push event: 104
- Pull request event: 4
- Create event: 4
Last Year
- Release event: 1
- Delete event: 2
- Issue comment event: 1
- Push event: 104
- Pull request event: 4
- Create event: 4
Dependencies
- actions/checkout v4 composite
- julia-actions/cache v1 composite
- julia-actions/julia-buildpkg v1 composite
- julia-actions/julia-docdeploy v1 composite
- julia-actions/julia-runtest v1 composite
- julia-actions/setup-julia v2 composite
- JuliaRegistries/TagBot v1 composite
- actions/checkout v4 composite
- julia-actions/setup-julia v2 composite