update20140518

pdt2txt.sh

@@ -1,12 +1,12 @@
 #!/bin/bash
 for FILE in "$@" 
 do
-	pdfseparate "$FILE" pdf2txtTemp-%d.pdf
+	pdftoppm $FILE pdf2txtTemp
-	for i in pdf2txtTemp-*.pdf 
+	for i in pdf2txtTemp-*.ppm 
 	do 
-		convert -monochrome -normalize -density 300 $i $i.png
+		ppm2tiff  $i $i.tiff
-		tesseract $i.png  $i -l eng+deu 
+		tesseract $i.tiff $i -l eng+deu 
 	done
-	cat pdf2txtTemp-*.pdf.txt > "$FILE".txt
+	cat pdf2txtTemp-*.ppm.txt > "$FILE".txt
 	rm  pdf2txtTemp-*.*
 done

reverssh.sh

@@ -0,0 +1 @@
+ ssh -fN -R 5182:localhost:22  ssh@109.226.173.73 -p 5187 -v

streamcloud2strm.sh

@@ -0,0 +1,10 @@
+#!/bin/bash
+for FILE in "$@" 
+do
+	grep "file:" "$FILE" > result 
+	sed -i "s/file: \"//g" result
+	sed -i "s/\t//g" result
+	sed -i "s/\",//g" result
+	mv -f result "${FILE%.html}.strm"
+	rm "$FILE"
+done