rsennrich / smorlemma Goto Github PK

Hi,

I am trying to compile a new SMOR analyzer with the current Wiktionary. However, despite following the instructions (i.e. getting all the dependencies, generating XML output from Zmorge etc) I am getting this error:

ulimit -d 50000000; sfst/fst-compiler-utf8-epsilonfilter morph-lemma.fst morph-lemma.a
morph-lemma.fst: 8
symbols.fst: 80
morph-lemma.fst: 10
reading transducer from morph2.a...finished
morph-lemma.fst: 47
Allocation of memory failed in Mem::add_buffer!
Makefile:47: recipe for target 'morph-lemma.a' failed
make: *** [morph-lemma.a] Error 1

when i adjust ulimit to less than 50GB i am simply hitting the memory problem earlier and i don't have access to instances wit more memory than that. What am i doing wrong? I am using Ubuntu 18.04 and I installed all the dependencies using apt-get.

The wiki-lexicon.xml that i get out of current Wiktionary is about 18MB (in case the problem lies further upstream - e.g. perhaps something compiled incorrectly?). Any help would be very appreciated.

Hi, is there a tagset mapping from SMOR to the "Interset" tagset?
https://metacpan.org/source/ZEMAN/Lingua-Interset-2.025/lib/Lingua/Interset/Tagset/DE
http://universaldependencies.github.io/docs/