Add ocr pdf script

3 files changed, 48 insertions, 0 deletions

diff --git a/extras/mkocrpdf-cuneform.sh b/extras/mkocrpdf-cuneform.sh
new file mode 100644
index 0000000..94e10eb
--- /dev/null
+++ b/extras/mkocrpdf-cuneform.sh
@@ -0,0 +1,21 @@
+#!/bin/sh
+#
+# Makes a pdf with embedded text as extracted by cuneiform
+#
+# Requires imagemagick, pdftk, hocr2pdf and cuneiform
+
+for i in `ls *png`
+do
+	a=`basename $i .png`
+	echo processing $a
+
+	convert $i $a.bmp
+	cuneiform -f hocr -o $a.html $a.bmp
+	rm -f $a.bmp
+
+	# hocr2pdf has a habit of segfaulting, so fall back to convert
+	hocr2pdf -i $i -o $a.pdf < $a.html || convert $i $a.pdf
+	rm -f $a.html
+done
+
+pdftk *pdf cat output book.pdf
diff --git a/extras/mkocrpdf.sh b/extras/mkocrpdf.sh
new file mode 100644
index 0000000..165059a
--- /dev/null
+++ b/extras/mkocrpdf.sh
@@ -0,0 +1,22 @@
+#!/bin/sh
+#
+# Makes a PDF with embedded text extracted by tesseract
+#
+# Requires imagemagick, pdftk, hocr2pdf and tesseract
+#
+# Also requires this tesseract configuration:
+# echo 'tessedit_create_hocr 1' > /usr/local/share/tessdata/configs/hocr
+
+for i in `ls *png`
+do
+	a=`basename $i .png`
+	echo processing $a
+
+	tesseract $i $a hocr 2>/dev/null
+
+	# hocr2pdf has a habit of segfaulting, so fall back to convert
+	hocr2pdf -i $i -o $a.pdf < $a.html || convert $i $a.pdf
+	rm -f $a.html
+done
+
+pdftk *pdf cat output book.pdf
diff --git a/extras/mkpdf.sh b/extras/mkpdf.sh
new file mode 100755
index 0000000..80e776f
--- /dev/null
+++ b/extras/mkpdf.sh
@@ -0,0 +1,5 @@
+#!/bin/sh
+# See COPYING file for copyright and license details.
+# Requires imagemagick
+
+convert *.png book.pdf