From cd0c0a821361f5ee7c52ee60fb0ed5b758e53620 Mon Sep 17 00:00:00 2001
From: Nick White <git@njw.me.uk>
Date: Sun, 30 Oct 2011 12:29:45 +0000
Subject: Add ocr pdf script

---
 extras/mkocrpdf.sh | 22 ++++++++++++++++++++++
 1 file changed, 22 insertions(+)
 create mode 100644 extras/mkocrpdf.sh

(limited to 'extras/mkocrpdf.sh')

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
-- 
cgit v1.2.3