4dConcatDataTable

#!/usr/bin/env bash
set -euo pipefail
trap 'e=$?; [ $e -ne 0 ] && echo "$0 exited in error"' EXIT

# wrap around 3dbucket and 3drefit
# extract a subbrick from a list of nifti 
# label with luna ids
# 20200430 - like 4dConcatSubBriks but for a datatable file
# 20231016 - add stdin option
#
usage(){
   cat <<HEREDOC
use 3dMVM/3dLME datatable to create 4d concat image
Useful for quality checking many structurals, subject masks, or individual contrasts.
Wraps (1) 3dbucket InputFile column (2) 3drefit w/ Subj column

 USAGE:
  $(basename "$0") output.nii.gz datatable.tsv [extraidcol]

 OPTIONS:
  output.nii.gz   the file to create
                  will not overwrite (must rm to rerun)

  datatable.tsv   input file as used by AFNI's 3dMVM or 3dLME
                  tab separated.
                  must have header w/ 'Subj' and 'InputFile' columns
                  can be '-' to read from stdin

  extraidcol      optional extra column(s) to add to 3drefit ids (regexp)
                  '|' to seprate multiples: 'Sess|Sex' (id is combined Subj,Sess,Sex)

 EXAMPLES:
  4dConcatDataTable /tmp/all.nii.gz r2p.tsv

  4dConcatDataTable /tmp/all.nii.gz r2p.tsv 'session|shapes'

  mlr --tsv filter '\$InputFile =~ "ses-01"' r2p.tsv| 4dConcatDataTable /tmp/ses1.nii.gz -

HEREDOC
}

[ $# -lt 2 -o $# -gt 3 ] && usage >&2 && exit 1

prefix="$1"; shift
datatable="$1"; shift

! [[ $prefix =~ .nii$|.nii.gz$|.HEAD$ ]] &&
   echo "prefix '$prefix' must end in .nii, .nii.gz, or .HEAD" && exit 1

[ -r "$prefix" ] &&
   echo "rm $prefix # to regenerate; skipping 3dbucket and 3drefit" &&
   exit 0

if [ "$datatable" = "-" ]; then
   tty >/dev/null  &&
      echo -e "ERROR: using '$datatable' as datatable requires pipe;\n\t grep thing | $(basename "$0") $prefix $datatable $*" &&
      exit 1
   datatable=$(mktemp "${TMPDIR:-/tmp}/4dConcatTable_XXXX.tsv")
   cat > "$datatable"
   # TODO: this disables original exit in error trap
   trap 'echo "# removing $datatable" && rm "$datatable"; exit' EXIT
fi

[ ! -r "$datatable" ] &&
   echo "$datatable file doesn't exist!" && exit 1
#read ididx fileidx <<< $(sed 1q $datatable | sed 's/[ \t]\+/\n/g' | perl -lne 'print $. if m/(Subj|InputFile)/i')

# if we want to use an extra id
extraid=""; extraidpatt=""; awkextra="" 
[ $# -eq 1 ] && extraid="$1" && extraidpatt="|$extraid" 

mapfile -t idxs < <(sed 1q "$datatable" | sed 's/[ \t]\+/\n/g' | perl -lne "print $. if m/(Subj$extraidpatt|InputFile)/i")
# index of the file index is last item
# but 0-based. so minus 1
fii=${#idxs[@]}
! let fii-- && echo "ERROR: no files in $datatable!? Does input have a header?"  && exit 1

# TODO: test only 2 if not $3, 3 otherwise?

ididx=${idxs[0]}
fileidx=${idxs[$fii]}
[ ${#idxs[@]} -eq 3 ] && awkextra=' "_"'"\$${idxs[1]}"

[ -z "$fileidx" ] && echo "no InputFile in '$datatable' header?! (idxs '$ididx' '$fileidx')" && exit 1
mapfile -t ids < <(awk "(NR>1) { print \$$ididx$awkextra}" "$datatable")
mapfile -t files < <(awk "(NR>1){ print \$$fileidx}" "$datatable")
[ ${#ids[@]} -ne "${#files[@]}" ] &&
   echo "error parsing '$datatable'. have different number of ids and files!" &&
   exit 1


echo "3dbucket: combine"
dryrun 3dbucket -prefix "$prefix" "${files[@]}"
echo "3drefit: relabel"
dryrun 3drefit -relabel_all_str "${ids[*]}" "$prefix"