2021-01-14 Benchmarking Testing Party Meeting notes

# (if starting from docker)
docker pull continuumio/miniconda
docker run -it continuumio/miniconda
groupadd -r app
useradd -r -g app -d /home/app -s /bin/bash -c "Docker image user" app
mkdir /home/app
chown app /home/app
su app
cd
wget "https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh"
bash Miniconda3-latest-Linux-x86_64.sh -b -p miniconda3
. miniconda3/etc/profile.d/conda.sh 
conda activate base

# (Otherwise just start here)
conda install -y anaconda-client -n base
conda env create openforcefield/openff-benchmark-optimization
conda activate openff-benchmark-optimization
conda install -c omnia/label/benchmark openforcefield
git clone https://github.com/openforcefield/openff-benchmark.git
cd openff-benchmark
pip install -e .

# To make a simple single-conf molecule from SMILES
from openforcefield.topology import Molecule
mol = Molecule.from_smiles('COCOCOCOCO')
mol.generate_conformers()
mol.to_file('input.sdf')

# preprocess
openff-benchmark preprocess validate -g BBB data/*sdf
openff-benchmark preprocess generate-conformers 1-validate_and_assign/
openff-benchmark preprocess coverage-report 2-generate_conformers openff_unconstrained-1.3.0.offxml

# optimize
## qm calculations
openff-benchmark optimize execute --season "1:1" -o 3-export-compute-qm 2-generate_conformers/
## mm calculations
openff-benchmark optimize execute --season "1:2" -o 3-export-compute-mm 3-export-compute-qm/b3lyp-d3bj/dzvp/

# analysis 
openff-benchmark report compare-forcefields --input-path 3-export-compute-qm/ --input-path 3-export-compute-mm/openff-1.3.0/ --ref_method b3lyp-d3bj --output_directory 4-compare_forcefields

# plots
openff-benchmark report plots --input-path 4-compare_conformers --ref-method b3lyp-d3bj