Ensure dir-to-pdf saves to dirname.pdf not dirname/.pdf, and handle all different naming conventions

1 files changed, 19 insertions, 4 deletions

diff --git a/dir-to-pdf.sh b/dir-to-pdf.sh
index 345d908..a0dc95f 100755
--- a/dir-to-pdf.sh
+++ b/dir-to-pdf.sh
@@ -44,7 +44,7 @@ if ! test -d "$1"; then
 fi
 
 if test $# -eq 2 ; then
-	outfile="$2"
+	outfile=`echo "$2" | sed 's/\/$//'`
 else
 	outfile="$1.pdf"
 fi
@@ -58,12 +58,27 @@ fi
 mkdir -p "$tmpdir" || exit 1
 
 echo "Copying hocrs and converting pngs from $1 to $tmpdir"
-find "$1" -maxdepth 1 -type f -name '*.unpapered.png' | while read i; do
-	b=`basename "$i" .unpapered.png`
+n=`find "$1" -maxdepth 1 -type f -name '*.unpapered.png' | wc -l`
+if test $n -gt 0 ; then
+	imgsuffix=".unpapered.png"
+	hocrsuffix="_unpapered_"
+else
+	n=`find "$1" -maxdepth 1 -type f -name '*.binarized.png' | wc -l`
+	if test $n -gt 0 ; then
+		imgsuffix=".binarized.png"
+		hocrsuffix="_"
+	else
+		echo "Error: no pages found"
+		exit 1
+	fi
+fi
+
+find "$1" -maxdepth 1 -type f -name '*'"$imgsuffix" | while read i; do
+	b=`basename "$i" "$imgsuffix"`
 
 	hocr=""
 	for t in $TRAININGS; do
-		n=`echo "$i" | sed "s/.unpapered.png/_unpapered_$t.hocr/"`
+		n=`echo "$i" | sed "s/${imgsuffix}/${hocrsuffix}$t.hocr/"`
 		test -f "$n" && hocr="$n"
 	done
 	if test -z "$hocr"; then