orca_run.sh

#!/bin/bash

#by Gabriel Libânio Silva Rodrigues

# Script to run Orca quantum chemistry program, moving the temporary files to a scratch directory given by ${orca_scr}

# Nprocs is the number of processors to be used
# Maxcore is the ORCA maximum memory per core.
# Use -mail to send an email to you in the end of the calculation if this option is configured in your system
# -a is to copy any additional files to the run directory, as example, .gbw files to read molecular orbitals
# or .xyz files for multiple structures runs. If you need more than one -a file, the multiple files must be 
# specified between double quotes like $ orca_run -i input.inp -a "file1.gbw file2.xyz"

# orca_run variables
if [ -z "${orca_dir}" ] || [ -z "${orca_src}" ] || [ -z "${orca_scr}" ]; then
	echo "
	You have to set the orca_run enviroment variables (bash profile).
	
	- Dir variable = \$orca_dir
	- Source variable = \$orca_src (source modules, variables and etc.)
	- Scratch variable = \$orca_scr
	
	Example:
	export orca_dir=path/to/orca 
	export orca_src=path/to/source 
	export orca_scr=path/to/scratch
	
	"
	exit 0
fi
source ${orca_src}
ORCAPATH=${orca_dir}
ORCASCR=${orca_scr}

# Calculation directory
CALCDIR="${PWD}"
# Variables
DATE=`date "+%d-%m-%Y"`
TIME=`date "+%H:%M:%S"`
input=""
output=""
nprocs=""
email=""
afile=""
maxcore=""
usage="SCRIPT USAGE:

orca_run -i input.inp -o output.out -p nprocs -m maxcore -a \"file1.gbw file2.xyz\" -e email@addres.com hostsender

Only input is obligatory, default output = input-basename.out
For multiple auxiliary files, use the -a option separating the file names with space and INSIDE double quotes:

-a \"file1.gbw file2.xyz\""

# Input options, passing arguments
while getopts i:o:p:a:e:m:h option
do
	case "${option}" in
		i)	input=${OPTARG}
			# New input to be placed in the runfiles folder to see modifications
			inputNEW="${input%.*}.new.inp"
			RUNDIR="${ORCASCR}/${input%.*}-$$"
			;;
		o)	output=${OPTARG};;
		p)	nprocs=${OPTARG};;
		a)	afile=${OPTARG};;
		e)	email=${OPTARG}
			eval "hostsender=\${$OPTIND}"
			shift 2
			;;
		m)	maxcore=${OPTARG};;
		h | *) # Display help.
			echo "${usage}"
			exit 0
		;;
	esac
done
shift $((OPTIND-1))

### START and show what will be done ###
echo "Orca will run on node $HOSTNAME in $DATE at $TIME with the following options" > ${input%.*}.nodes
echo "input = $input" >> ${input%.*}.nodes
echo "output = $output" >> ${input%.*}.nodes
echo "number of processors = $nprocs" >> ${input%.*}.nodes
echo "maxcore memory = $maxcore" >> ${input%.*}.nodes
echo "extrafile = $afile" >> ${input%.*}.nodes
echo "mail = $email send by host $hostsender" >> ${input%.*}.nodes
echo "scrDIR = $RUNDIR" >> ${input%.*}.nodes

### Argument Cases ###
# Alert if email will be sent
if [ -n "$email" ]; then
	echo "An email will be send to $email via host $hostsender when job is finished."
fi
# Case there is extra files to copy
if [ -n "$afile" ]; then
	for file in ${afile[@]}
		do cp "${CALCDIR}"/${file} "${RUNDIR}"
	done
fi
# Case no output is specified, output will be same input name with .out extension
if [ -z "$output" ]; then
	output=${input%.*}.out
fi
# Case nprocs was specified
if [ -n "$nprocs" ]; then
	## Erase possible %pal configuration in input
	sed -i -e "/^%pal.*$/Id" $input
	## Write the number of processors to inp file with requested %pal configuration
	echo "%Pal nprocs $nprocs end" | cat - $input > inputOR.temp && mv inputOR.temp $input
fi
if [ -n "$maxcore" ]; then
# Case maxcore was specified
	## Erase possible %maxcore configuration in input
	sed -i -e "/^%maxcore.*$/Id" $input
	## Write the number of processors to inp file with requested %maxcore configuration
	echo "%maxcore $maxcore" | cat - $input > inputOR.temp && mv inputOR.temp $input
fi

### Run the program itself and alert the user ####
echo "
If you need help try the option: orca_run.sh -h

Run directory = ${RUNDIR}
Orca will run with command:

nohup ${ORCAPATH}/orca ${RUNDIR}/${input} > "${CALCDIR}"/$output
"

#### Starting job script ####
mkdir -p "${RUNDIR}"
cd "${RUNDIR}"
cp "${CALCDIR}/$input" "${RUNDIR}" # Copy input to run dir

###### RUN ORCA #######
nohup ${ORCAPATH}/orca "${RUNDIR}/${input}" > "${CALCDIR}/${output}"

# Do operation on final ORCA files
for file in "${PWD}"; do
        if [ -d "${file}" ]; then
                ## Remove temporary files if they still exist
                if [ "${file: -4}" == ".tmp" ]; then
                        rm ${file}
                fi
                ## Move old input file to a .new.inp file so user knows what changed
                mv ${input} ${inputNEW}
                ## Finally, move the files from the calculation to their folder
                mkdir -p "${CALCDIR}/${input%.*}-runfiles"
                mv ${input%.*}* "${CALCDIR}/${input%.*}-runfiles"
        else
                echo "There are no ORCA files to move, check your calculation!"
        fi
done

# If email was requested send it
if [ -n "$email" ]; then
	echo "echo	"ssh $USER@$hostsender "echo "The job $input has ended on $(hostname)." | mail -s "JOB_DONE!" $email" 
fi