Currently, I am starting the local server this way:
java -cp languagetool-server.jar org.languagetool.server.HTTPServer --port 8081 “*” &
The post is like this:
function postit($s) {
$ch = curl_init();
curl_setopt($ch, CURLOPT_URL,"http://localhost:8081/v2/check");
curl_setopt($ch, CURLOPT_POST, 1);
curl_setopt($ch, CURLOPT_POSTFIELDS,"language=nl&level=picky&text=".urlencode($s));
curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
$server_output = curl_exec($ch);
curl_close($ch);
return $server_output;
}
But this does not test all rules for Dutch. 'ALL; meaning:
- all disabled rules for Dutch
- all categories for Dutch
- full spell checking for Dutch
- all picky rules
- all Java rules
- preferably also remote rules when uses in the Open Source version
(but of course: not the Premium ones, since that is not Open Source)
Is this possible?
I am continuously finding mistakes in the current test output, mistakes I am sure there is a (picky, diabled or otherwise) rule for. Even some spelling errors seem to slip through the net (though these might actually be mistakes in the spelling files).
I need this to split the corpus (23 GB) as good as possible in a ‘good’ and a ‘bad’ part.