edit-quant

#!/usr/bin/env bash
# Runner for the edit_quant pipeline (amplicon mapping + CRISPRessoPooled).
# Assumes this is being called from within this directory (main/); the
# nextflow workdir defaults to one level up.

set -e

#-----------------------------------------------------------------------
# Function definitions

# Creates a string <date>_<hourminute>
timestamp_folder_name() {
  date '+%Y%m%d_%H%M'
}

# Checks if a path is writable
check_writable_path() {
    local path="$1"
    # Check if path exists
    if [ -e "$path" ]; then
        # Path exists, check if writable
        if [ -w "$path" ]; then
            return 0
        else
            echo "❌ Error: Path '$path' exists but is not writable"
            return 1
        fi
    else
        # Path doesn't exist, check if parent directory is writable
        local parent_dir=$(dirname "$path")
        if [ -w "$parent_dir" ]; then
            return 0
        else
            echo "❌ Error: Cannot create directory at '$path'. Parent directory not writable"
            return 1
        fi
    fi
}

# Print usage
usage() {
    echo -e "Usage: $0 [-c <file.config>] [-p <profile>] [-l] [-w <workdir>] [-q <queue>] [-t <hours>] [-g <group>] -- [nextflow pipeline args]"
    echo -e "-c\t\t\t\t<path/to/run.config> extra Nextflow config file for the run (e.g. sample-specific overrides)"
    echo -e "-p\t\t\t\t<profile> Nextflow profile for the run [local|farm] (default: farm)"
    echo -e "-l\t\t\t\tRun nextflow locally instead of submitting to LSF (shorthand for -p local)"
    echo -e "-w\t\t\t\tSet the nextflow work directory (default: ../workdir)"
    echo -e "-q\t\t\t\tSet the LSF queue for the nextflow driver job (default: normal)"
    echo -e "-t\t\t\t\tSet the time limit for the nextflow driver job (hours) (default: 72)"
    echo -e "-g\t\t\t\tSet the LSF group for the nextflow driver job (required unless -l)"
    echo -e "-- The rest is passed straight through to nextflow -- this is how you set"
    echo -e "   the pipeline's own params (--amplicon_manifest, --fastq_manifest,"
    echo -e "   --ref_fasta, --gene_bed/--genes_gtf, ...), and it overrides anything in -c"
    echo -e ""
    echo -e "Examples:"
    echo -e "  ./edit_quant -l -- --amplicon_manifest manifest.tsv --fastq_manifest fastq.tsv \\"
    echo -e "      --ref_fasta genome.fa --genes_gtf genes.gtf.gz --outdir results"
    echo -e "  ./edit_quant -p farm -g myteam -q normal -- --amplicon_manifest manifest.tsv \\"
    echo -e "      --fastq_manifest fastq.tsv --ref_fasta genome.fa --gene_bed genome.genes.bed"
    exit 1
}

#-----------------------------------------------------------------------
#                 Update these for your installation
#-----------------------------------------------------------------------
# Set defaults
SUBMIT=1
WORKDIR="../workdir"

# These control the parameters for the Nextflow driver job (not the
# per-process resources -- those live in conf/base.config / conf/farm.config)
QUEUE="oversubscribed"
RUNTIME_LIMIT="120"
GROUP="teamtrynka"
PROFILE="lsf"

SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
NF_FILE="${SCRIPT_DIR}/main.nf"

prefix="$(timestamp_folder_name)"
mkdir -p logs/${prefix}

# Source your nextflow installation here (no-op off the farm, e.g. running
# this locally for testing where `module` isn't available)
command -v module >/dev/null 2>&1 && module load HGI/common/nextflow/26.04

# Needed for conda (bioconda/conda-forge) to reach the internet from the farm
# when Nextflow auto-builds the environment.yml env
export HTTP_PROXY='http://wwwcache.sanger.ac.uk:3128'
export HTTPS_PROXY='http://wwwcache.sanger.ac.uk:3128'
export NXF_OPTS="-Xms4G -Xmx4G"

# Set this to true to enable a more fancy logging.
# This fancy logging is a bit funky and designed for interactive sessions!
# Alternatively: tail -f <logfile>  or watch -n 1 tail -10 <logfile>
# if not using an interactive session
export NXF_ANSI_LOG=true

# Nextflow version, change depending on installation you use
export NXF_VER=26.04.6

#-----------------------------------------------------------------
while getopts ":c:p:hlw:q:t:g:" opt; do
    case $opt in
        c)
            CONFIG_FILE="$OPTARG"
            ;;
        p)
            PROFILE="$OPTARG"
            ;;
        l)
            SUBMIT=0
            ;;
        w)
            WORKDIR="$OPTARG"
            ;;
        q)
            QUEUE="$OPTARG"
            ;;
        t)
            RUNTIME_LIMIT="$OPTARG"
            ;;
        g)
            GROUP="$OPTARG"
            ;;
        h)
            usage
            ;;
        \?)
            echo "❌ ERROR: invalid option -$OPTARG"
            usage
            ;;
        :)
            echo "❌ ERROR: option -$OPTARG requires an argument"
            usage
            ;;
    esac
done

# Drop all parsed options so "$@" is just the pass-through nextflow args
shift $((OPTIND - 1))

if [ "$SUBMIT" -eq 1 ] && [ -z "$GROUP" ]; then
    echo "❌ ERROR: -g <LSF group> is required when submitting to LSF (pass -l to run locally instead)"
    exit 1
fi

# Check .config extension
if [ -n "$CONFIG_FILE" ] && [[ ! "$CONFIG_FILE" =~ \.config$ ]]; then
    echo "❌ WARN: -c file '$CONFIG_FILE' does not have a .config extension"
fi

# Check if workdir is writable
if ! check_writable_path "$WORKDIR"; then
    exit 1
fi

# Just to make sure it gets logged correctly
if [ "$SUBMIT" -eq 0 ]; then
    QUEUE="Running locally"
    RUNTIME_LIMIT="NA"
fi

echo "📄 Pipeline:       $NF_FILE"
echo "📄 Config file:    ${CONFIG_FILE:-<none>}"
echo "📒 Logs:           logs/${prefix}"
echo "📂 Workdir:        $WORKDIR"
echo "🔄 Profile/Queue:  $PROFILE / $QUEUE"
echo "⏱️  Runtime limit:  $RUNTIME_LIMIT hours"
echo "🟠 Arguments:      $@"

# Build the nextflow command. Pipeline params (--amplicon_manifest etc) come
# from whatever's left in "$@" after option parsing, and/or from -c's config
# file -- CLI args here override the config file, same as any nextflow run.
CMD="nextflow \
-log logs/${prefix}/edit_quant.nextflow.log \
run ${NF_FILE} \
-profile ${PROFILE} \
-w ${WORKDIR} \
-resume \
-with-report logs/${prefix}/edit_quant.nextflow.html \
-with-trace logs/${prefix}/edit_quant.nextflow.trace"

# Add the config if given
if [ -n "$CONFIG_FILE" ]; then
    CMD="$CMD -c ${CONFIG_FILE}"
fi

# Add the remaining arguments
CMD="$CMD $@"

echo "--------------------------------------------------------------------"
echo $CMD
echo "--------------------------------------------------------------------"
if [ "$SUBMIT" -eq 0 ]; then
    # This runs the nextflow job locally
    eval $CMD
else
    #-----------------------------------------------------------------------
    #                 Update this for your installation
    #-----------------------------------------------------------------------
    # This submits the nextflow runner (the driver, not the individual
    # pipeline processes -- those go through conf/farm.config's LSF executor)
    # to the cluster. Memory/queue here are for the lightweight driver only.
    CMD="bsub -n 1 \
    -G ${GROUP} \
    -M 4000 \
    -q ${QUEUE} \
    -W ${RUNTIME_LIMIT}:00 \
    -R 'span[hosts=1] select[mem>4000] rusage[mem=4000]' \
    -o logs/${prefix}/edit_quant-%J.out \
    -e logs/${prefix}/edit_quant-%J.err \
    $CMD"

    eval $CMD
fi