ftdmp-qa-all

#!/bin/bash

function print_help_and_exit
{
cat >&2 << 'EOF'

'ftdmp-qa-all' scores and ranks multimeric structures of proteins or nucleic acids

Options:
    --workdir                         string  *  path to directory for caching and writing detailed results
    --subselect-contacts              string     query to subselect inter-chain contacts for scoring, default is '[]'
    --constraints-required            string     query to check required inter-chain contacts, default is ''
    --constraints-banned              string     query to check banned inter-chain contacts, default is ''
    --constraint-clashes              number     max allowed clash score, default is ''
    --subselect-atoms-global          string     query to subselect atoms for global scores, default is '[]'
    --reference                       string     input structure file to compute CAD-score with, default is ''
    --ftdmp-root                      string     ftdmp root path, default is '' (autodetected from the calling command)
    --conda-path                      string     conda installation path, default is ''
    --conda-early                     string     flag to activate conda as early as possible
    --conda-env                       string     conda main environment name, default is ''
    --processors                      number     number of processes to run when scoring, default is 8
    --sbatch                          string     sbatch parameters to run scoring in parallel, default is ''
    --score-symmetry                  string     flag to score symmetry, default is 'false'
    --external-scores                 string     path to input file with external scores table, default is ''
    --remap-cadscore                  string     flag to use optimal chains remapping for CAD-score, default is 'false'
    --crude-cadscore                  string     flag to use faster but crude mode for CAD-score
    --keep-top-fast                   number     number of top complexes to keep after full scoring stage, default is 9999999
    --keep-top-slow                   number     number of top complexes to keep before slow full scoring stage, default is 9999999
    --limit-voromqa-light             number     minimal allowed VoroMQA-light whole-stricture score, default is ''
    --rank-names                      string     rank names to use, or name of a standard set of rank names, default is 'protein_protein_voromqa_and_global_and_gnn_no_sr'
    --ranks-top                       number     number of top complexes to consider for each ranking, default is 300
    --jury-slices                     string     slice sizes sequence definition for ranks jury scoring, default is '5 20'
    --jury-cluster                    number     clustering threshold for ranks jury scoring, default is 0.9
    --jury-maxs                       number     number of max values to use for ranks jury scoring, default is 1
    --output-redundancy-threshold     number     minimal ordered redundancy value to accept, default is 0.9
    --plot-jury-scores                string     file path to output plot of jury scores, default is ''
    --plot-jury-diagnostics           string     flag to plot jury diagnostics, default is 'false'
    --write-pdb-file                  string     file path template to output scores in PDB files, default is ''
    --write-pdb-mode                  string     mode for PDB scores output ('voromqa_dark' or 'voromqa_dark_and_gnn'), default is 'voromqa_dark_and_gnn'
    --write-pdb-num                   number     number of top PDB files with scores to write, default is 5
    --write-full-table                string     file path to output full table, default is ''
    --help | -h                                  flag to display help message and exit

Standard input:
    input file paths

Standard output:
    space-separated table of scores

Examples:

    ls ./*.pdb | ftdmp-qa-all --conda-path ~/miniconda3 --workdir './tmp/works' --rank-names protein_protein_voromqa_and_global_and_gnn_no_sr
    
    ls ./*.pdb | ftdmp-qa-all --workdir './tmp/works' --rank-names protein_protein_voromqa_no_sr
    
    ls ./*.pdb | ftdmp-qa-all --conda-path ~/miniconda3 --workdir './tmp/works' --rank-names protein_protein_voromqa_and_global_and_gnn_no_sr \
        --write-pdb-file './output/scored_-RANK-_-BASENAME-' --write-pdb-mode 'voromqa_dark_and_gnn' --write-pdb-num 5

Named collections of rank names, to be provided as a single string to '--rank-names':

    protein_protein_voromqa_and_global_and_gnn_no_sr
    protein_protein_voromqa_and_global_and_gnn_with_sr
    protein_protein_voromqa_and_gnn_no_sr
    protein_protein_voromqa_and_gnn_with_sr
    protein_protein_voromqa_no_sr
    protein_protein_voromqa_with_sr
    protein_protein_simplest_voromqa
    generalized_voromqa

EOF
exit 1
}

################################################################################

if [ -z "$1" ]
then
	print_help_and_exit
fi

WORKDIR=""
INTERFACE_SUBSELECTION="[]"
CONSTRAINTS_REQUIRED=""
CONSTRAINTS_BANNED=""
CONSTRAINT_CLASHES=""
SUBSELECT_ATOMS_FOR_GLOBAL="[]"
REFERENCE_STRUCTURE_FILE_FOR_COMPARISON=""
FTDMP_ROOT=""
CONDA_PATH=""
CONDA_EARLY="false"
CONDA_ENV=""
SCORING_PROCESSORS="8"
SCORING_SBATCH=""
SCORE_SYMMETRY="false"
EXTERNAL_SCORES_FILE=""
REMAP_CADSCORE="false"
CRUDE_CADSCORE="false"
KEEP_TOP_FAST="9999999"
KEEP_TOP_SLOW="9999999"
LIMIT_VOROMQA_LIGHT=""
SCORING_RANKS="protein_protein_voromqa_and_global_and_gnn_no_sr"
SCORING_RANKS_TOP="300"
SCORING_RANKS_JURY_SLICES="5 20" 
SCORING_RANKS_JURY_CLUSTER="0.9" 
SCORING_RANKS_JURY_MAX_VALUES="1" 
REDUNDANCY_THRESHOLD="0.9"
PLOT_JURY_SCORES=""
PLOT_JURY_DIAGNOSTICS="false"
WRITE_SCORES_PDB_FILE=""
WRITE_SCORES_PDB_MODE="voromqa_dark_and_gnn"
WRITE_SCORES_PDB_NUM="5"
WRITE_FULL_TABLE=""
HELP_MODE="false"

while [[ $# > 0 ]]
do
	OPTION="$1"
	OPTARG="$2"
	shift
	case $OPTION in
	--workdir)
		WORKDIR="$OPTARG"
		shift
		;;
	--subselect-contacts)
		INTERFACE_SUBSELECTION="$OPTARG"
		shift
		;;
	--constraints-required)
		CONSTRAINTS_REQUIRED="$OPTARG"
		shift
		;;
	--constraints-banned)
		CONSTRAINTS_BANNED="$OPTARG"
		shift
		;;
	--constraint-clashes)
		CONSTRAINT_CLASHES="$OPTARG"
		shift
		;;
	--subselect-atoms-global)
		SUBSELECT_ATOMS_FOR_GLOBAL="$OPTARG"
		shift
		;;
	--reference)
		REFERENCE_STRUCTURE_FILE_FOR_COMPARISON="$OPTARG"
		shift
		;;
	--openmm-forcefield)
		OPENMM_FORCEFIELD="$OPTARG"
		shift
		;;
	--ftdmp-root)
		FTDMP_ROOT="$OPTARG"
		shift
		;;
	--conda-path)
		CONDA_PATH="$OPTARG"
		shift
		;;
	--conda-early)
		CONDA_EARLY="$OPTARG"
		shift
		;;
	--conda-env)
		CONDA_ENV="$OPTARG"
		CONDA_ENV_FOR_GNN="$OPTARG"
		shift
		;;
	--processors)
		SCORING_PROCESSORS="$OPTARG"
		shift
		;;
	--sbatch)
		SCORING_SBATCH="$OPTARG"
		shift
		;;
	--score-symmetry)
		SCORE_SYMMETRY="$OPTARG"
		shift
		;;
	--external-scores)
		EXTERNAL_SCORES_FILE="$OPTARG"
		shift
		;;
	--remap-cadscore)
		REMAP_CADSCORE="$OPTARG"
		shift
		;;
	--crude-cadscore)
		CRUDE_CADSCORE="$OPTARG"
		shift
		;;
	--keep-top-fast)
		KEEP_TOP_FAST="$OPTARG"
		shift
		;;
	--keep-top-slow)
		KEEP_TOP_SLOW="$OPTARG"
		shift
		;;
	--limit-voromqa-light)
		LIMIT_VOROMQA_LIGHT="$OPTARG"
		shift
		;;
	--rank-names)
		SCORING_RANKS="$OPTARG"
		shift
		;;
	--ranks-top)
		SCORING_RANKS_TOP="$OPTARG"
		shift
		;;
	--jury-slices)
		SCORING_RANKS_JURY_SLICES="$OPTARG"
		shift
		;;
	--jury-cluster)
		SCORING_RANKS_JURY_CLUSTER="$OPTARG"
		shift
		;;
	--jury-maxs)
		SCORING_RANKS_JURY_MAX_VALUES="$OPTARG"
		shift
		;;
	--output-redundancy-threshold)
		REDUNDANCY_THRESHOLD="$OPTARG"
		shift
		;;
	--plot-jury-scores)
		PLOT_JURY_SCORES="$OPTARG"
		shift
		;;
	--plot-jury-diagnostics)
		PLOT_JURY_DIAGNOSTICS="$OPTARG"
		shift
		;;
	--write-pdb-file)
		WRITE_SCORES_PDB_FILE="$OPTARG"
		shift
		;;
	--write-pdb-mode)
		WRITE_SCORES_PDB_MODE="$OPTARG"
		shift
		;;
	--write-pdb-num)
		WRITE_SCORES_PDB_NUM="$OPTARG"
		shift
		;;
	--write-full-table)
		WRITE_FULL_TABLE="$OPTARG"
		shift
		;;
	-h|--help)
		HELP_MODE="true"
		;;
	*)
		echo >&2 "Error: invalid command line option '$OPTION'"
		exit 1
		;;
	esac
done

if [ "$HELP_MODE" == "true" ]
then
	print_help_and_exit
fi

[ -n "$WORKDIR" ] || { echo >&2 "Error: working directory path not provided"; exit 1;}

[ -n "$INTERFACE_SUBSELECTION" ] || { echo >&2 "Error: contacts subselection not defined"; exit 1; }

[ -n "$SCORING_RANKS" ] || { echo >&2 "Error: scoring rank names not provided"; exit 1;}

[ -n "$SCORING_RANKS_JURY_SLICES" ] || { echo >&2 "Error: scoring ranks jury slices not defined"; exit 1;}

if [ -n "$WRITE_SCORES_PDB_FILE" ] && [ "$WRITE_SCORES_PDB_NUM" -gt "0" ] && [ "$WRITE_SCORES_PDB_MODE" != "voromqa_dark" ] && [ "$WRITE_SCORES_PDB_MODE" != "voromqa_dark_and_gnn" ]
then
	echo >&2 "Error: invalid mode '$WRITE_SCORES_PDB_MODE' for writing scores in PDB files"
	exit 1
fi

################################################################################

readonly TMPLDIR=$(mktemp -d)
trap "rm -r $TMPLDIR" EXIT

cat \
| tr ' ' '\n' \
| egrep '.' \
| sort \
| uniq \
> "$TMPLDIR/input_list"

if [ ! -s "$TMPLDIR/input_list" ]
then
	echo >&2 "Error: no input in stdin"
	exit 1
fi

NUMBER_OF_INPUT_FILES="$(wc -l $TMPLDIR/input_list | awk '{print $1}')"

INPUT_FILE_PATH_PREFIX="$(dirname $(head -n 1 $TMPLDIR/input_list))/"

while read -r SINGLEINFILE
do
	if [ ! -s "$SINGLEINFILE" ]
	then
		echo >&2 "Error: no input file '$SINGLEINFILE'"
		exit 1
	fi
	SINGLEINFILE_PATH_PREFIX="$(dirname ${SINGLEINFILE})/"
	if [ "$SINGLEINFILE_PATH_PREFIX" != "$INPUT_FILE_PATH_PREFIX" ]
	then
		echo >&2 "Error: input files do not have the same path prefix"
		exit 1
	fi
done < "$TMPLDIR/input_list"

################################################################################

if [ "$SCORING_RANKS" == "generalized_voromqa" ]
then
	SCORING_RANKS="
raw_FIVb_iface_energy_rank
raw_FIVb_energy_clash_tour_rank
"
fi

if [ "$SCORING_RANKS" == "protein_protein_simplest_voromqa" ]
then
	SCORING_RANKS="
raw_FIV_iface_energy_rank
raw_FIV_energy_clash_tour_rank
"
fi

if [ "$SCORING_RANKS" == "protein_protein_voromqa_no_sr" ]
then
	SCORING_RANKS="
raw_FIV_iface_energy_rank
raw_FIV_energy_clash_tour_rank
raw_FIVb_iface_energy_rank
raw_FIVb_energy_clash_tour_rank
"
fi

if [ "$SCORING_RANKS" == "protein_protein_voromqa_with_sr" ]
then
	SCORING_RANKS="
raw_FIV_iface_energy_rank
raw_FIV_energy_clash_tour_rank
raw_FIVb_iface_energy_rank
raw_FIVb_energy_clash_tour_rank
raw_FIV_sr_iface_energy_rank
raw_FIV_sr_energy_clash_tour_rank
raw_FIVb_sr_iface_energy_rank
raw_FIVb_sr_energy_clash_tour_rank
"
fi

if [ "$SCORING_RANKS" == "protein_protein_voromqa_and_gnn_no_sr" ]
then
	SCORING_RANKS="
raw_FIV_iface_energy_rank
raw_FIV_energy_clash_tour_rank
raw_FIVb_iface_energy_rank
raw_FIVb_energy_clash_tour_rank
raw_FIGNN_sum_of_gnn_scores_rank
raw_FIGNN_average_gnn_score_rank
raw_FIGNN_gnn_scores_tour_rank
raw_FIGNN_average_pcadscore_rank
raw_FIGNN_weighted_average_pcadscore_rank
raw_FIGNN_sum_of_gnn_scores_clash_tour_rank
raw_FIGNN_sum_of_gnn_scores_FIV_energy_tour_rank
raw_FIGNN_avg_gnn_score_FIV_energy_tour_rank
raw_FIGNN_wavg_pcadscore_FIV_energy_tour_rank
raw_FIGNN_sum_of_gnn_scores_FIVb_energy_tour_rank
raw_FIGNN_avg_gnn_score_FIVb_energy_tour_rank
raw_FIGNN_wavg_pcadscore_FIVb_energy_tour_rank
"
fi

if [ "$SCORING_RANKS" == "protein_protein_voromqa_and_gnn_with_sr" ]
then
	SCORING_RANKS="
raw_FIV_iface_energy_rank
raw_FIV_energy_clash_tour_rank
raw_FIVb_iface_energy_rank
raw_FIVb_energy_clash_tour_rank
raw_FIGNN_sum_of_gnn_scores_rank
raw_FIGNN_average_gnn_score_rank
raw_FIGNN_gnn_scores_tour_rank
raw_FIGNN_average_pcadscore_rank
raw_FIGNN_weighted_average_pcadscore_rank
raw_FIGNN_sum_of_gnn_scores_clash_tour_rank
raw_FIGNN_sum_of_gnn_scores_FIV_energy_tour_rank
raw_FIGNN_avg_gnn_score_FIV_energy_tour_rank
raw_FIGNN_wavg_pcadscore_FIV_energy_tour_rank
raw_FIGNN_sum_of_gnn_scores_FIVb_energy_tour_rank
raw_FIGNN_avg_gnn_score_FIVb_energy_tour_rank
raw_FIGNN_wavg_pcadscore_FIVb_energy_tour_rank
raw_FIV_sr_iface_energy_rank
raw_FIV_sr_energy_clash_tour_rank
raw_FIVb_sr_iface_energy_rank
raw_FIVb_sr_energy_clash_tour_rank
raw_FIGNN_sr_sum_of_gnn_scores_rank
raw_FIGNN_sr_average_gnn_score_rank
raw_FIGNN_sr_gnn_scores_tour_rank
raw_FIGNN_sr_average_pcadscore_rank
raw_FIGNN_sr_weighted_average_pcadscore_rank
raw_FIGNN_sr_sum_of_gnn_scores_clash_tour_rank
raw_FIGNN_sr_sum_of_gnn_scores_FIV_sr_energy_tour_rank
raw_FIGNN_sr_avg_gnn_score_FIV_sr_energy_tour_rank
raw_FIGNN_sr_wavg_pcadscore_FIV_sr_energy_tour_rank
raw_FIGNN_sr_sum_of_gnn_scores_FIVb_sr_energy_tour_rank
raw_FIGNN_sr_avg_gnn_score_FIVb_sr_energy_tour_rank
raw_FIGNN_sr_wavg_pcadscore_FIVb_sr_energy_tour_rank
"
fi

if [ "$SCORING_RANKS" == "protein_protein_voromqa_and_global_and_gnn_no_sr" ]
then
	SCORING_RANKS="
raw_FIV_iface_energy_rank
raw_FIV_energy_clash_tour_rank
raw_FIVb_iface_energy_rank
raw_FIVb_energy_clash_tour_rank
raw_FIGNN_sum_of_gnn_scores_rank
raw_FIGNN_average_gnn_score_rank
raw_FIGNN_gnn_scores_tour_rank
raw_FIGNN_average_pcadscore_rank
raw_FIGNN_weighted_average_pcadscore_rank
raw_FIGNN_sum_of_gnn_scores_clash_tour_rank
raw_FIGNN_sum_of_gnn_scores_FIV_energy_tour_rank
raw_FIGNN_avg_gnn_score_FIV_energy_tour_rank
raw_FIGNN_wavg_pcadscore_FIV_energy_tour_rank
raw_FIGNN_sum_of_gnn_scores_FIVb_energy_tour_rank
raw_FIGNN_avg_gnn_score_FIVb_energy_tour_rank
raw_FIGNN_wavg_pcadscore_FIVb_energy_tour_rank
raw_FIV_and_FGV_light_tour_rank
raw_FIV_and_FGV_dark_tour_rank
raw_FIGNN_and_FGV_light_tour_rank
raw_FIGNN_and_FGV_dark_tour_rank
"
fi

if [ "$SCORING_RANKS" == "protein_protein_voromqa_and_global_and_gnn_with_sr" ]
then
	SCORING_RANKS="
raw_FIV_iface_energy_rank
raw_FIV_energy_clash_tour_rank
raw_FIVb_iface_energy_rank
raw_FIVb_energy_clash_tour_rank
raw_FIGNN_sum_of_gnn_scores_rank
raw_FIGNN_average_gnn_score_rank
raw_FIGNN_gnn_scores_tour_rank
raw_FIGNN_average_pcadscore_rank
raw_FIGNN_weighted_average_pcadscore_rank
raw_FIGNN_sum_of_gnn_scores_clash_tour_rank
raw_FIGNN_sum_of_gnn_scores_FIV_energy_tour_rank
raw_FIGNN_avg_gnn_score_FIV_energy_tour_rank
raw_FIGNN_wavg_pcadscore_FIV_energy_tour_rank
raw_FIGNN_sum_of_gnn_scores_FIVb_energy_tour_rank
raw_FIGNN_avg_gnn_score_FIVb_energy_tour_rank
raw_FIGNN_wavg_pcadscore_FIVb_energy_tour_rank
raw_FIV_and_FGV_light_tour_rank
raw_FIV_and_FGV_dark_tour_rank
raw_FIGNN_and_FGV_light_tour_rank
raw_FIGNN_and_FGV_dark_tour_rank
raw_FIV_sr_iface_energy_rank
raw_FIV_sr_energy_clash_tour_rank
raw_FIVb_sr_iface_energy_rank
raw_FIVb_sr_energy_clash_tour_rank
raw_FIGNN_sr_sum_of_gnn_scores_rank
raw_FIGNN_sr_average_gnn_score_rank
raw_FIGNN_sr_gnn_scores_tour_rank
raw_FIGNN_sr_average_pcadscore_rank
raw_FIGNN_sr_weighted_average_pcadscore_rank
raw_FIGNN_sr_sum_of_gnn_scores_clash_tour_rank
raw_FIGNN_sr_sum_of_gnn_scores_FIV_sr_energy_tour_rank
raw_FIGNN_sr_avg_gnn_score_FIV_sr_energy_tour_rank
raw_FIGNN_sr_wavg_pcadscore_FIV_sr_energy_tour_rank
raw_FIGNN_sr_sum_of_gnn_scores_FIVb_sr_energy_tour_rank
raw_FIGNN_sr_avg_gnn_score_FIVb_sr_energy_tour_rank
raw_FIGNN_sr_wavg_pcadscore_FIVb_sr_energy_tour_rank
"
fi

if [[ $SCORING_RANKS == "file:"* ]]
then
	SCORING_RANKS="$(cat "$(echo ${SCORING_RANKS} | sed 's|^file:||')" | tr '\n' ' ')"
fi

SCORING_RANKS="$(echo ${SCORING_RANKS} | tr '\n' ' ' | sed 's/\s\+/ /g' | sed 's/^\s\+//' | sed 's/\s\+$//')"

SCORING_MODE_PROTEIN="false"
SCORING_MODE_PROTEIN_SIDECHAIN_REBUILT="false"
SCORING_MODE_GENERIC="false"
SCORING_MODE_GENERIC_SIDECHAIN_REBUILT="false"
SCORING_MODE_PROTEIN_GNN="false"
SCORING_MODE_PROTEIN_GNN_SIDECHAIN_REBUILT="false"
SCORING_MODE_PROTEIN_GLOBAL="false"

if [[ "$SCORING_RANKS" == *"_FIV_"* ]]
then
	SCORING_MODE_PROTEIN="true"
fi

if [[ "$SCORING_RANKS" == *"_FIV_sr_"* ]]
then
	SCORING_MODE_PROTEIN_SIDECHAIN_REBUILT="true"
fi

if [[ "$SCORING_RANKS" == *"_FIVb_"* ]]
then
	SCORING_MODE_GENERIC="true"
fi

if [[ "$SCORING_RANKS" == *"_FIVb_sr_"* ]]
then
	SCORING_MODE_GENERIC_SIDECHAIN_REBUILT="true"
fi

if [[ "$SCORING_RANKS" == *"_FIGNN_"* ]]
then
	SCORING_MODE_PROTEIN_GNN="true"
fi

if [[ "$SCORING_RANKS" == *"_FIGNN_sr_"* ]]
then
	SCORING_MODE_PROTEIN_GNN_SIDECHAIN_REBUILT="true"
fi

if [[ "$SCORING_RANKS" == *"_FGV_"* ]]
then
	SCORING_MODE_PROTEIN_GLOBAL="true"
fi

if [ -z "$REFERENCE_STRUCTURE_FILE_FOR_COMPARISON" ] && [ "$SCORING_MODE_PROTEIN" != "true" ] && [ "$SCORING_MODE_PROTEIN_SIDECHAIN_REBUILT" != "true" ] && [ "$SCORING_MODE_GENERIC" != "true" ] && [ "$SCORING_MODE_GENERIC_SIDECHAIN_REBUILT" != "true" ] && [ "$SCORING_MODE_PROTEIN_GNN" != "true" ] && [ "$SCORING_MODE_PROTEIN_GNN_SIDECHAIN_REBUILT" != "true" ] && [ "$SCORING_MODE_PROTEIN_GLOBAL" != "true" ]
then
	echo >&2 "Error: no scoring action enabled"
	exit 1
fi

if [ -n "$WRITE_SCORES_PDB_FILE" ] && [ "$WRITE_SCORES_PDB_NUM" -gt "0" ]
then
	if [ "$WRITE_SCORES_PDB_MODE" == "voromqa_dark" ]
	then
		if [ "$SCORING_MODE_PROTEIN_GNN" != "true" ]
		then
			echo >&2 "Error: mode '$WRITE_SCORES_PDB_MODE' for writing scores in PDB files is incompatible with requested scoring ranks"
			exit 1
		fi
	fi
	
	if [ "$WRITE_SCORES_PDB_MODE" == "voromqa_dark_and_gnn" ]
	then
		if [ "$SCORING_MODE_GENERIC" != "true" ] || [ "$SCORING_MODE_PROTEIN_GNN" != "true" ]
		then
			echo >&2 "Error: mode '$WRITE_SCORES_PDB_MODE' for writing scores in PDB files is incompatible with requested scoring ranks"
			exit 1
		fi
	fi
fi

################################################################################

if [ "$CONDA_EARLY" == "true" ]
then
	if [ -z "$CONDA_DEFAULT_ENV" ]
	then
		if [ -z "$CONDA_PATH" ]
		then
			echo >&2 "Error: not in conda environment, and the conda path is not provided"
			exit 1
		fi
		if [ ! -s "${CONDA_PATH}/bin/activate" ]
		then
			echo >&2 "Error: no conda activation script '${CONDA_PATH}/bin/activate'"
			exit 1
		fi
		source "${CONDA_PATH}/bin/activate"
	fi
	
	if [ -n "$CONDA_ENV" ]
	then
		if [ "$CONDA_DEFAULT_ENV" != "$CONDA_ENV" ]
		then
			conda activate "$CONDA_ENV"
		fi
		
		if [ "$CONDA_DEFAULT_ENV" != "$CONDA_ENV" ]
		then
			echo >&2 "Error: no '$CONDA_ENV' environment"
			exit 1
		fi
	fi
	
	CONDA_PATH=""
fi

################################################################################

if [ -z "$FTDMPDIR" ] || [ -n "$FTDMP_ROOT" ]
then
	if [ -z "$FTDMP_ROOT" ]
	then
		export FTDMPDIR="$( cd -- "$( dirname -- "${BASH_SOURCE[0]}" )" &> /dev/null && pwd )"
	else
		export FTDMPDIR="$FTDMP_ROOT"
	fi
	export PATH="${FTDMPDIR}/core/voronota/expansion_js:${FTDMPDIR}/core/3D_Dock/progs:${FTDMPDIR}:${PATH}"
fi

################################################################################

JOB_NAME="$(cat ${TMPLDIR}/input_list | xargs cat | md5sum | awk '{print $1}')"

OUTPUT_PATH="${WORKDIR}/jobs/${JOB_NAME}"
mkdir -p "${OUTPUT_PATH}"

{
	echo "ID"
	cat "${TMPLDIR}/input_list" | xargs -L 1 basename
} \
> "${OUTPUT_PATH}/ids_to_process.txt"

CACHE_DIR="${WORKDIR}/cache"

PREV_SCORING_RESULTS_FILE="${OUTPUT_PATH}/ids_to_process.txt"

################################################################################

LOCAL_COLUMNS="false"
if [ -n "$WRITE_FULL_TABLE" ] && [ "$SCORING_MODE_PROTEIN_GNN" == "true" ]
then
	LOCAL_COLUMNS="true"
fi

ftdmp-all-score \
  --input-from-table        "$PREV_SCORING_RESULTS_FILE" \
  --structure-monomer1      "" \
  --structure-monomer2      "" \
  --structure-prefix        "$INPUT_FILE_PATH_PREFIX" \
  --structure-suffix        "" \
  --subselect-contacts      "$INTERFACE_SUBSELECTION" \
  --constraints-required    "$CONSTRAINTS_REQUIRED" \
  --constraints-banned      "$CONSTRAINTS_BANNED" \
  --constraint-clashes      "$CONSTRAINT_CLASHES" \
  --subselect-atoms-global  "$SUBSELECT_ATOMS_FOR_GLOBAL" \
  --reference               "$REFERENCE_STRUCTURE_FILE_FOR_COMPARISON" \
  --remap-cadscore          "$REMAP_CADSCORE" \
  --mode-protein            "$SCORING_MODE_PROTEIN" \
  --mode-protein-sr         "$SCORING_MODE_PROTEIN_SIDECHAIN_REBUILT" \
  --mode-generic            "$SCORING_MODE_GENERIC" \
  --mode-generic-sr         "$SCORING_MODE_GENERIC_SIDECHAIN_REBUILT" \
  --mode-protein-gnn        "$SCORING_MODE_PROTEIN_GNN" \
  --mode-protein-gnn-sr     "$SCORING_MODE_PROTEIN_GNN_SIDECHAIN_REBUILT" \
  --mode-protein-global     "$SCORING_MODE_PROTEIN_GLOBAL" \
  --external-scores         "$EXTERNAL_SCORES_FILE" \
  --conda-path              "$CONDA_PATH" \
  --conda-env               "$CONDA_ENV" \
  --score-symmetry          "$SCORE_SYMMETRY" \
  --local-columns           "$LOCAL_COLUMNS" \
  --keep-top                "$KEEP_TOP_FAST" \
  --keep-top-for-slow       "$KEEP_TOP_SLOW" \
  --clash-tour-tolerance    "0.05" \
  --parallel                "$SCORING_PROCESSORS" \
  --cache-dir               "$CACHE_DIR" \
  --sbatch                  "$SCORING_SBATCH" \
  --output-names-prefix     "raw_" \
  --output-dir              "$OUTPUT_PATH"

PREV_SCORING_RESULTS_FILE="${OUTPUT_PATH}/raw_scoring_results_with_ranks.txt"

if [ ! -s "$PREV_SCORING_RESULTS_FILE" ]
then
	echo >&2 "Error: no raw scoring results"
	exit 1
fi

################################################################################

if [ -n "$LIMIT_VOROMQA_LIGHT" ] && [ "$SCORING_MODE_PROTEIN_GLOBAL" == "true" ]
then
	cat "$PREV_SCORING_RESULTS_FILE" \
	| ftdmp-filter-table \
	  ">=${LIMIT_VOROMQA_LIGHT}" "raw_FGV_full_light_score" \
	> "${OUTPUT_PATH}/raw_scoring_results_with_ranks_limited_by_voromqa_light.txt"
	
	PREV_SCORING_RESULTS_FILE="${OUTPUT_PATH}/raw_scoring_results_with_ranks_limited_by_voromqa_light.txt"
	
	if [ ! -s "$PREV_SCORING_RESULTS_FILE" ]
	then
		echo >&2 "Error: no raw scoring results after limiting by VoroMQA-light score"
		exit 1
	fi
fi

################################################################################

if [ "$NUMBER_OF_INPUT_FILES" -gt "1" ]
then
	ftdmp-all-jury \
	  --input-scoring-table     "$PREV_SCORING_RESULTS_FILE" \
	  --structure-monomer1      "" \
	  --structure-monomer2      "" \
	  --structure-prefix        "$INPUT_FILE_PATH_PREFIX" \
	  --structure-suffix        "" \
	  --subselect-contacts      "$INTERFACE_SUBSELECTION" \
	  --rank-names              "$SCORING_RANKS" \
	  --ranks-top               "$SCORING_RANKS_TOP" \
	  --remap-cadscore          "$REMAP_CADSCORE" \
	  --crude-cadscore          "$CRUDE_CADSCORE" \
	  --jury-slices             "$SCORING_RANKS_JURY_SLICES" \
	  --jury-cluster            "$SCORING_RANKS_JURY_CLUSTER" \
	  --jury-max-values         "$SCORING_RANKS_JURY_MAX_VALUES" \
	  --redundancy-threshold    "$REDUNDANCY_THRESHOLD" \
	  --parallel                "$SCORING_PROCESSORS" \
	  --sbatch                  "$SCORING_SBATCH" \
	  --output-names-prefix     "raw_"
	
	PREV_SCORING_RESULTS_FILE="${OUTPUT_PATH}/raw_top_scoring_results_RJS.txt"
	
	if [ -n "$WRITE_FULL_TABLE" ]
	then
		mkdir -p "$(dirname ${WRITE_FULL_TABLE})"	
		cat "$PREV_SCORING_RESULTS_FILE" \
		> "$WRITE_FULL_TABLE"
	fi
	
	if [ ! -s "$PREV_SCORING_RESULTS_FILE" ]
	then
		echo >&2 "Error: no raw jury results"
		exit 1
	fi
	
	if [ -n "$PLOT_JURY_SCORES" ]
	then
		ftdmp-plot-jury-scores \
		  --input-scoring-table "$PREV_SCORING_RESULTS_FILE" \
		  --prefix-for-jury-scores 'raw_RJS' \
		  --output "${OUTPUT_PATH}/raw_RJS_plot.png"
		
		mkdir -p "$(dirname ${PLOT_JURY_SCORES})"
		cp "${OUTPUT_PATH}/raw_RJS_plot.png" "$PLOT_JURY_SCORES"
	fi
	
	if [ "$PLOT_JURY_DIAGNOSTICS" == "true" ]
	then
		ftdmp-plot-jury-diagnostics \
		  --input-prefix "${OUTPUT_PATH}/raw_top_scoring_results_RJS_diagnostic_" \
		  --output-prefix "${OUTPUT_PATH}/raw_RJS_diagnostic_plot_"
	fi
	
	PREV_SCORING_RESULTS_FILE="${OUTPUT_PATH}/raw_top_scoring_results_RJS_nonredundant.txt"
	
	if [ ! -s "$PREV_SCORING_RESULTS_FILE" ]
	then
		echo >&2 "Error: no raw non-redundant jury results"
		exit 1
	fi
fi

################################################################################

if [ -n "$WRITE_SCORES_PDB_FILE" ] && [ "$WRITE_SCORES_PDB_NUM" -gt "0" ]
then
	cat "$PREV_SCORING_RESULTS_FILE" \
	| ftdmp-extract-table-columns ID \
	| tail -n +2 \
	| head -n "$WRITE_SCORES_PDB_NUM" \
	| awk '{print NR " " $1}' \
	| while read -r RESULT_RANK ORIGINAL_FILE_BASENAME
	do
		PRETTY_RESULT_RANK="$(echo 00000000${RESULT_RANK} | sed 's/.*\(....\)$/\1/')"
		ORIGINAL_INPUT_FILE="${INPUT_FILE_PATH_PREFIX}${ORIGINAL_FILE_BASENAME}"
		OUTPUT_PDB_FILE="$(echo "${WRITE_SCORES_PDB_FILE}" | sed "s|-RANK-|${PRETTY_RESULT_RANK}|" | sed "s|-BASENAME-|${ORIGINAL_FILE_BASENAME}|")"
		echo "${ORIGINAL_INPUT_FILE} ${OUTPUT_PDB_FILE}"
	done \
	> "$TMPLDIR/pdb_files_for_writing_cores"
	
	if [ "$WRITE_SCORES_PDB_MODE" == "voromqa_dark" ] || [ "$WRITE_SCORES_PDB_MODE" == "voromqa_dark_and_gnn" ]
	then
		cat "$TMPLDIR/pdb_files_for_writing_cores" \
		| awk '{print " --input " $1 " --output-dark-pdb " $2}' \
		| xargs -L 1 -P "$SCORING_PROCESSORS" voronota-js-only-global-voromqa \
		  --cache-dir "$CACHE_DIR" \
		  --processors 1 \
		> /dev/null
	fi
	
	if [ "$WRITE_SCORES_PDB_MODE" == "voromqa_dark_and_gnn" ]
	then
		cat "$TMPLDIR/pdb_files_for_writing_cores" \
		| awk '{print " --input " $2 " --output-pdb-file " $2}' \
		| xargs -L 1 -P "$SCORING_PROCESSORS" voronota-js-voroif-gnn \
		  --conda-path "$CONDA_PATH" \
		  --conda-env "$CONDA_ENV" \
		  --gnn "${FTDMPDIR}/core/voronota/expansion_js/voroif/gnn_packages/v1" \
		  --cache-dir "$CACHE_DIR" \
		  --processors 1 \
		  --output-pdb-mode "combine" \
		> /dev/null
	fi
fi

################################################################################

{
cat << 'EOF'
ID
raw_RJS_max
raw_FICS_iface_cadscore
raw_FICS_iface_site_based_cadscore
raw_FIGNN_sum_of_gnn_scores
raw_FIGNN_average_pcadscore
raw_FGV_full_dark_score
raw_FGV_full_light_score
raw_FIV_iface_energy
raw_FIVb_iface_energy
raw_FIVb_iface_clash_score
raw_FIVb_iface_area
raw_FIGNN_sr_sum_of_gnn_scores
raw_FIGNN_sr_average_pcadscore
raw_FIV_sr_iface_energy
raw_FIVb_sr_iface_energy
raw_FIVb_sr_iface_clash_score
raw_FIVb_sr_iface_area
raw_FIV_and_FGV_light_tour_rank
raw_FIV_and_FGV_dark_tour_rank
raw_RDDS_ordered_score
EOF
} \
| egrep -f <(head -n 1 "$PREV_SCORING_RESULTS_FILE" | tr ' ' '\n' | egrep . | awk -v dollar='$' '{print "^" $1 dollar}') \
> "$TMPLDIR/columns_to_output"

cat "$PREV_SCORING_RESULTS_FILE" \
| ftdmp-extract-table-columns $(cat $TMPLDIR/columns_to_output) \
> "$TMPLDIR/restricted_table"

{
head -n 1 "$TMPLDIR/restricted_table" \
| tr ' ' '\n' \
| sed 's/^raw_RJS_max$/jury_score/' \
| sed 's/^raw_FICS_iface_cadscore$/cadscore/' \
| sed 's/^raw_FICS_iface_site_based_cadscore$/site_cadscore/' \
| sed 's/^raw_FIGNN_sum_of_gnn_scores$/GNN_sum_score/' \
| sed 's/^raw_FIGNN_average_pcadscore$/GNN_pcadscore/' \
| sed 's/^raw_FGV_full_dark_score$/voromqa_dark/' \
| sed 's/^raw_FGV_full_light_score$/voromqa_light/' \
| sed 's/^raw_FIV_iface_energy$/voromqa_energy/' \
| sed 's/^raw_FIVb_iface_energy$/gen_voromqa_energy/' \
| sed 's/^raw_FIVb_iface_area$/area/' \
| sed 's/^raw_FIVb_iface_clash_score$/clash_score/' \
| sed 's/^raw_FIGNN_sr_sum_of_gnn_scores$/sr_GNN_sum_score/' \
| sed 's/^raw_FIGNN_sr_average_pcadscore$/sr_GNN_pcadscore/' \
| sed 's/^raw_FIV_sr_iface_energy$/sr_voromqa_energy/' \
| sed 's/^raw_FIVb_sr_iface_energy$/sr_gen_voromqa_energy/' \
| sed 's/^raw_FIVb_sr_iface_area$/sr_area/' \
| sed 's/^raw_FIVb_sr_iface_clash_score$/sr_clash_score/' \
| sed 's/^raw_FIV_and_FGV_light_tour_rank$/rank_2018/' \
| sed 's/^raw_FIV_and_FGV_dark_tour_rank$/rank_2020/' \
| sed 's/^raw_RDDS_ordered_score$/redundancy/' \
| tr '\n' ' ' \
| sed 's/\s*$/\n/'

tail -n +2 "$TMPLDIR/restricted_table" \
| awk '{printf("%s", $1); for(i=2;i<=NF;i++){printf(" %.3f", $i);} print("");}' \
| sed 's/\.0\+\s/ /g' \
| sed 's/\.0\+$//g'
} \
> "${OUTPUT_PATH}/results_for_humans.txt"

cat "${OUTPUT_PATH}/results_for_humans.txt"