3
0
Fork 0
forked from mirrors/nixpkgs

Adding training results files for some languages to tesseract to be able to do OCR directly.

svn path=/nixpkgs/trunk/; revision=26956
This commit is contained in:
Lluís Batlle i Rossell 2011-04-24 20:01:19 +00:00
parent 47c44f9315
commit 9a0a0c92c7

View file

@ -1,5 +1,22 @@
{ stdenv, fetchurl, libtiff }:
let
f = lang : sha256 : let
src = fetchurl {
url = "http://tesseract-ocr.googlecode.com/files/${lang}.traineddata.gz";
inherit sha256;
};
in
"gunzip -c ${src} > $out/share/tessdata/${lang}.traineddata";
extraLanguages = ''
${f "cat" "1qndk8qygw9bq7nzn7kzgxkm3jhlq7jgvdqpj5id4rrcaavjvifw"}
${f "rus" "0yjzks189bgcmi2vr4v0l0fla11qdrw3cb1nvpxl9mdis8qr9vcc"}
${f "spa" "1q1hw3qi95q5ww3l02fbhjqacxm34cp65fkbx10wjdcg0s5p9q2x"}
${f "nld" "0cbqfhl2rwb1mg4y1140nw2vhhcilc0nk7bfbnxw6bzj1y5n49i8"}
'';
in
stdenv.mkDerivation {
name = "tesseract-3.0.0";
@ -10,6 +27,8 @@ stdenv.mkDerivation {
buildInputs = [ libtiff ];
postInstall = extraLanguages;
meta = {
description = "OCR engine";
homepage = http://code.google.com/p/tesseract-ocr/;