RSS AI Zone
Hey all,
has someone managed to use an external pos-tagger on a system that is NOT windows? Right now i try to implement a german version of treetagger. On a windows-system everything works perfectly fine with the German bot.
But i have to work with Linux/Mac so i have to change the simpletopic ~xPOSTAGSENTENCE, because the “tag-german” script is not existing on a Linux/Installation of treetagger.
I tried changing it to “tree-tagger-german” as it does basically the same as “tag-german”. But then i get the following output:
LAURA: Hey!
plpl: > :prepare hilft
TokenControl: DO_NUMBER_MERGE DO_PROPERNAME_MERGE DO_DATE_MERGE
Original User Input: hilft
Tokenized into: hilft
Actual used input: hilft
reading parameters ...
tagging ...
finished.
entering null word to dictionary in sentence
Concepts:
1: hilft (raw):
+~
+hilft
//
1: hilft (canonical): //
Sequences:
The canonical in the example should be “helfen”. Again, this works on windows and it works on Linux/Mac too, if i call “tree-tagger-german chat.txt” in a terminal window. There seems to be a missing link between ChatScript and Treetagger.
So i don’t know what the “entering null word to dictionary in sentence” error means. I found the error in the dictionarySystem.cpp file, but can’t really figure it out.
WORDP StoreWord(char* word, uint64 properties)
{
if (!*word) // servers dont want long lists of bugs from strange inputs // we require something
{
if (!server) ReportBug((char*)"entering null word to dictionary in sentence")
return StoreWord((char*)"emptystring"); // we require something
