CCSNet

Installation of CCSNet environments

Omicron Environment

conda create -n ccsnet-datagen python=3.9 -y
conda install -c conda-forge omicron=4.0.0 -y
pip install poetry==1.7.1
cd ~/CCSNet/libs/ccsnet
poetry install

Training, Validation, and Testing Environment

conda create -n ccsnet-train python=3.9 -y
pip install poetry==1.7.1
cd ~/CCSNet/apps/train
poetry install
cd ~/CCSNet/libs/ccsnet
poetry install

Run CCSNet

To run CCSNet, you would have to provide the project name and path of the code base, data, and output.

export PROJECT="Run_CCSNet/My_first_run" 
export CODE_BASE=path/to/ccsnet
export DATA_PATH=path/to/data_dir
export RESULT_DIR=path/to/output
python <ccsnet-application>

To make life easy, you can create a .env file like this.

# .env

CODE_BASE=path/to/ccsnet
DATA_PATH=path/to/data_dir
RESULT_DIR=path/to/output

Then you can run the training application like this.

PROJECT="Run_CCSNet/My_first_run" python path/to/ccsnet/apps/train/trainer/trainer.py -e $path/to/.env

To run all the CCSNet applications at once you can create a ccsnet.sh file like this.

# ccsnet.sh

PROJECT=Run_CCSNet/My_First_Run
CODE_BASE=/path/to/ccsnet

# Data generation
PROJECT=$PROJECT python $CODE_BASE/apps/datagen/collections/run_omicron.py -e $CODE_BASE/.env # (This line needs to be run separately.)
PROJECT=$PROJECT python $CODE_BASE/apps/datagen/collections/get_strain.py -e $CODE_BASE/.env

# Run Training
PROJECT=$PROJECT python $CODE_BASE/apps/train/trainer/trainer.py -e $CODE_BASE/.env

# Run Test datagen and testing 
for i in {1..3}
do
    PROJECT=$PROJECT python $CODE_BASE/apps/datagen/test_data/background_sampler.py -e $CODE_BASE/.env -s $i
    PROJECT=$PROJECT python $CODE_BASE/apps/tests/test_ccsnet.py -e $CODE_BASE/.env -s $i -r run_01
done

and run bash ccsnet.sh to get a happy life.