smefin inprovement

In order to make a smesmn bidix we need a better smefin.

Many t nodes contain parantheses. They shall be moved into elements.

   <e>
      <lg>
         <l pos="N">beassanvuohta</l>
      </lg>
      <mg>
         <tg xml:lang="fin">
            <t pos="N">pääsy (rel.)</t>
         </tg>
      </mg>
   </e>

The Finnish words should be proofread. The following command gives all words, the next one only the ones we do not recognize.

cat n_smefin.xml|grep '<l '|tr '<' '>'|cut -d">" -f3|see
cat n_smefin.xml|grep '<l '|tr '<' '>'|cut -d">" -f3|ufin|grep '?'|cut -f1|see

There are some Saami words not recognized:

cat *_smefin.xml|grep '<l '|tr '<' '>'|cut -d">" -f3|usme|grep '?'|wc -l

2921 out of 13131 translations contain a space. This number will get smaller as the parentheses are removed, but some will remain. Todo:

1. First remove parentheses to (above)
2. Then take out entries where all translations contain spaces,
3. Then look at they separately, and try to add one-word translations if possible