Useful commandoes for linguists

Words in an analysed text

How many words, not numerals

cat <analysedtext> | grep '"<' | grep '[a-zA-Z]'  | wc -l

How many uniq lemmas, not numerals

cat <analysedtext> | grep -v '"<' | cut -d '"' -f2 | grep '[a-zA-Z]' | sort -u | wc -l

Compounds in an analysed text

The syntactic analysis is important for getting the correct lemma through disambiguation. Many compounds are lexicalised in our analyser, and therefore we have to analyse the lemmas once more to find the compounds.
For all commands: Add ‘sort -u’ instead of ‘uniq’ to get numbers for uniqs

How many compound nouns

cat <analysedtext> | grep -v '"<' | cut -d '"' -f2 | grep '[a-záčžA-ZÁČŽ]' | usme | egrep 'Cmp.*N\+' |cut -f1 | uniq

How many compounds with noun in the first part

We don’t want Der/NomAct as N:

cat <analysedtext> | grep -v '"<' | cut -d '"' -f2 | grep '[a-záčžA-ZÁČŽ]' | usme | egrep 'N\+.*Cmp' | grep -v 'NomAct.*Cmp' | cut -f1 | uniq

How many compounds with verb in the first part

We don’t want Der/NomAg as V.

cat <analysedtext>| grep -v '"<' | cut -d '"' -f2 | grep '[a-záčžA-ZÁČŽ]' | usme | egrep 'V\+.*Cmp' | grep -v 'NomAg.*Cmp' | cut -f1 | uniq

How many compounds with adjective the in first part

We don’t want +N as A:

cat <analysedtext> | grep -v '"<' | cut -d '"' -f2 | grep '[a-záčžA-ZÁČŽ]' | usme | sed 's/^$/¢/' | tr "\n" "€" | tr "¢" "\n" | egrep 'A\+[A-Za-z\+]*Cmp' ]( egrep -v 'N\+[A-Za-z\+)*Cmp' |cut -f1 |tr -d "€"  | uniq

How many compounds with adverb in the first part

cat <analysedtext> | grep -v '"<' | cut -d '"' -f2 | grep '[a-záčžA-ZÁČŽ]' | usme | egrep 'Adv\+.*Cmp' | grep -v 'NomAct.*Cmp' | cut -f1 | uniq