Output formats
Some changes to textual output formats:
json:
- indent only by 2 (we have many levels of nesting)
- sort alphabetically, so that the order of attributes in the output is constant across multiple runs
lu_list:
- is now tab-separated and contains: lu_id, lexeme_id, list of lu lemmas with aspect info
txt:
- some changes of whitespace and treatment of comments, so that txt output is idempotent in the following sense:
./vallex-cli --output-format txt -i ../aktualni_data/data-txt/v-vallex.txt -o vallex.txt grep ''
./vallex-cli --output-format txt -i vallex.txt -o vallex2.txt grep ''
then vallex.txt = vallex2.txt
;
currently, txt output of the attribute recipr
is not correct, so we have to additionally run
sed -i 's/%all% //' vallex2.txt
for the above equality to hold