Make fullocrdir.sh only do things that haven't been done before

1 files changed, 6 insertions, 1 deletions

diff --git a/fullocrdir.sh b/fullocrdir.sh
index bac8121..7be36c5 100755
--- a/fullocrdir.sh
+++ b/fullocrdir.sh
@@ -31,6 +31,8 @@ find "$1" -maxdepth 1 -type f -name '*.jpg' | while read f; do
 	b=`basename "$f" .jpg`
 	d=`dirname "$f"`
 
+	test -f "$d/${b}_bin0.2.png" && echo "Skipping preprocessing for $b; .${b}_bin0.2.png already exists" && continue
+
 	preprocmulti "$f" "$d/$b" || exit 1
 done
 
@@ -39,7 +41,8 @@ find "$1" -maxdepth 1 -type f -name '*_bin?.?.png' | while read f; do
 	b=`basename "$f" .png`
 	d=`dirname "$f"`
 
-	# TODO: ensure to run the correct command here
+	test -f "$d/$b.hocr" && echo "Skipping tesseract for $b; .hocr already exists" && continue
+
 	tesseract -l $TRAINING "$f" "$d/$b" hocr || exit 1
 done
 
@@ -48,6 +51,8 @@ find "$1" -maxdepth 1 -type f -name '*.hocr' | while read f; do
 	b=`basename "$f" .hocr`
 	d=`dirname "$f"`
 
+	test -f "$d/${b}.conf" && echo "Skipping pgconf for $b; ${b}.conf already exists" && continue
+
 	pgconf "$f" > "$d/$b.conf" 2>/dev/null || rm -f "$d/$b.conf"
 done