#!/usr/bin/perl
# cleanup of data files
$version="1.4"; $date="18 Feb 2020"; #  by Nick Manini

sub print_help {
    print STDERR <<EOF;
prunedata: cleunup pointlessly long data files

SYNTAX:  prunedata [-t threshold] datafile
-t      : change the rejection threshold from the default = $SMALL
	  [a higher threshold results in less points being kept]
-h      : print this help
EOF
return;
}


$BIG=1.e30;
$SMALL=1.e-5;
$TINY=1.e-16;

$thresh=$SMALL;


while($p=shift(@ARGV)){
# [ file ] check existence of 'file'
    if($p=~/^[^\-]/){
        if(!(-e  $p)){
            die "file $p cannot be accessed\n";
        }
        push (@FILELIST,$p);
        next;
    }
    if($p eq "-t"){ # change threshold from default
        $thresh=shift(@ARGV)+0;
	$thresh>0 or die "there is no point setting a threshold <=0\n";
        next;
    }
    if($p eq "-h"){ # print help
        print_help;
        exit;
    }
    print STDERR "Error: invalid option $p\n ";
    print_help;
    exit;
}


if ($#FILELIST==-1)
{
    push(@FILELIST,"-");    #       adds stdio for lack of arguments
}


while ($filename=shift(@FILELIST)){	# loop on filenames

    $xmax=-$BIG;
    $ymax=-$BIG;
    $xmin=$BIG;
    $ymin=$BIG;
    $xkeep=$BIG;
    $ykeep=$BIG;

    open(INF, $filename);       #       Open for input
    while (<INF>){		# compute the max & min for x and y:
	s/^\s+//;
	@data=split(/[()\s]+/);

	$x=$data[0]+0.;
	$y=$data[1]+0.;

	if($x<$xmin){$xmin=$x;}
	if($y<$ymin){$ymin=$y;}
	if($x>$xmax){$xmax=$x;}
	if($y>$ymax){$ymax=$y;}
    }
    close(INF);

    $xrange=$xmax-$xmin;
    $yrange=$ymax-$ymin;
    $xcheck=$xrange+$TINY;
    $ycheck=$yrange+$TINY;
    
    print "# pruned with ",$xcheck," ",$ycheck,"\n";
    
    open(INF, $filename);       #       re-open for input

    while (<INF>){		# loop on lines in the file
	s/^\s+//;
	@data=split(/[()\s]+/);

	$xnext=$data[0]+0.;
	$ynext=$data[1]+0.;
	$lnext=$_;

	$xratio=abs($x-$xkeep)/$xcheck;
	$yratio=abs($y-$ykeep)/$ycheck;

	if(  # sufficient displacement:
	    (0.99*$xratio+0.01*$yratio > $thresh &&
	     0.01*$xratio+0.99*$yratio > $thresh)
	     || # a local max or minimum:
	     ($y-$ykeep)*($y-$ynext)>0
	    ){
#	    print $x," ",$y,"\n";
	    print $l;
	    $xkeep=$x;
	    $ykeep=$y;
	}
	$x=$xnext;
	$y=$ynext;
	$l=$lnext;
    }
    close(INF);
    if($x!=$xkeep){
#       print $x," ",$y,"\n";
	print $l;
    } # print last point anyway
}