dataset_rxn.sh

#!/bin/bash

#SBATCH --job-name=condCL_dataset_rxn
#SBATCH --mail-type=END,FAIL
#SBATCH --mail-user=matthew.ball3@astrazeneca.com
#SBATCH --nodes=1
#SBATCH --cpus-per-task=1
#SBATCH --mem=64G
#SBATCH --time=1-23:59:59
#SBATCH --output=condCL_dataset_rxn-%j.log
#SBATCH --error=condCL_dataset_rxn-%j.log
#SBATCH -p medium

# Load conda environment
#source ~/init_conda.sh
#conda activate condition_learning

# Define paths
SCRIPT_PATH="src/condition_learning/data/scripts/rxns"
CONFIG_PATH="config/data_curation/reactions"

# Function to run Python scripts
run_python_script() {
    local script=$1
    shift
    python "$SCRIPT_PATH/$script" "$@"
}

# 1. Extract reactions from reaction database
# run_python_script "extract_rxns.py" \
#     --config "$CONFIG_PATH/extract_rxns.yaml"

# 2. Clean the reaction conditions
for reaction_type in "amide_coupling" "buchwald_hartwig"; do
    run_python_script "clean_rxn_conditions.py" \
        --config "$CONFIG_PATH/clean_rxn_conditions/${reaction_type}.yaml"
done

# 3. Filter the reactions:
for reaction_type in "amide_coupling" "buchwald_hartwig"; do
    run_python_script "filter_rxns.py" \
        --config "$CONFIG_PATH/filter_rxns/${reaction_type}.yaml"
done

# 4. Add SMARTS-RX information:
for reaction_type in "amide_coupling" "buchwald_hartwig"; do
    python src/condition_learning/smartsrx.py \
        --config "$CONFIG_PATH/smartsrx/${reaction_type}.yaml"
done

# 5. De-duplicate reactions:
for reaction_type in "amide_coupling" "buchwald_hartwig"; do
    run_python_script "deduplicate.py" \
        --config "$CONFIG_PATH/deduplicate/${reaction_type}.yaml"
done