OneScience-Group/CodonTransformer

1#!/usr/bin/env bash2set -euo pipefail3 4# Convert a user-provided finetuning CSV into the JSONL format required by finetuning.5# Run this before scripts/slurm/run_finetune.sh.6# This script does not request SLURM resources.7 8PROJECT_DIR="${PROJECT_DIR:-/public/home/scnb9biwet/jiangqq/CodonTransformer-main}"9CONDA_ENV="${CONDA_ENV:-struct-evo}"10 11INPUT_CSV="${INPUT_CSV:-${PROJECT_DIR}/scripts/data/raw/your_data.csv}"12OUTPUT_JSON="${OUTPUT_JSON:-${PROJECT_DIR}/scripts/data/processed/finetune_data.json}"13 14cd "${PROJECT_DIR}"15mkdir -p "$(dirname "${OUTPUT_JSON}")"16 17export PYTHONPATH="${PROJECT_DIR}/model:${PYTHONPATH:-}"18export INPUT_CSV19export OUTPUT_JSON20export PYTHONFAULTHANDLER=121 22if [[ -n "${CONDA_ENV}" ]] && command -v conda >/dev/null 2>&1; then23    # shellcheck disable=SC109124    source "$(conda info --base)/etc/profile.d/conda.sh"25    conda activate "${CONDA_ENV}"26fi27 28if [[ ! -f "${INPUT_CSV}" ]]; then29    echo "Missing INPUT_CSV: ${INPUT_CSV}" >&230    exit 131fi32 33python - <<'PY'34import os35 36from CodonTransformer.CodonData import prepare_training_data37 38input_csv = os.environ["INPUT_CSV"]39output_json = os.environ["OUTPUT_JSON"]40 41print(f"Input CSV: {input_csv}")42print(f"Output JSONL: {output_json}")43 44prepare_training_data(input_csv, output_json)45PY46