Skip to content

Commit

Permalink
add missing column for conllu
Browse files Browse the repository at this point in the history
  • Loading branch information
@arysin
arysin committed Oct 24, 2024
1 parent bc50c92 commit b008835
Show file tree
Hide file tree
Showing 2 changed files with 32 additions and 32 deletions.
2 changes: 1 addition & 1 deletion src/main/groovy/ua/net/nlp/tools/tag/UdModule.groovy
View file
Original file line number Diff line number Diff line change
Expand Up @@ -148,7 +148,7 @@ class UdModule {

def miscStr = misc ? misc.join("|") : "_"

sb.append("${tkn.value}\t${tkn.lemma}\t${udPos}\t${tkn.tags}\t${udTagsStr}\t_\t_\t${miscStr}")
sb.append("${tkn.value}\t${tkn.lemma}\t${udPos}\t${tkn.tags}\t${udTagsStr}\t_\t_\t_\t${miscStr}")

sb.append('\n')
}
Expand Down
62 changes: 31 additions & 31 deletions src/test/groovy/ua/net/nlp/tools/tag/TagTextVerticalOutputTest.groovy
View file
Original file line number Diff line number Diff line change
Expand Up @@ -98,19 +98,19 @@ class TagTextVerticalOutputTest {
def expected =
"""# sent_id = 1
# text = $text
1 А а CCONJ conj:coord _ _ _ _
2 треба треба ADV noninfl:&predic _ _ _ Uninflect=Yes
3 далі далі ADV adv:compc:&predic Degree=Cmp _ _ SemTags=1:dist:2:time
4 воно воно PRON noun:unanim:n:v_naz:&pron:pers:3 Animacy=Anim,Inan|Case=Nom|Gender=Neut|Number=Sing|Person=3|PronType=Prs _ _ SemTags=1:conc:deictic
5 - - PUNCT punct _ _ _ _
6 озеро озеро NOUN noun:inanim:n:v_naz Animacy=Inan|Case=Nom|Gender=Neut|Number=Sing _ _ _
7 Світязь Світязь PROPN noun:inanim:m:v_naz:prop:geo:xp1 Animacy=Inan|Case=Nom|Gender=Masc|NameType=Geo|Number=Sing _ _ SemTags=1:conc:loc
8 де де ADV adv:&pron:int:rel PronType=Int|PronType=Rel _ _ _
9 я я PRON noun:anim:s:v_naz:&pron:pers:1 Animacy=Anim|Case=Nom|Number=Sing|Person=1|PronType=Prs _ _ SemTags=1:conc:hum:deictic
10 затримався затриматися VERB verb:rev:perf:past:m Aspect=Perf|Gender=Masc|Mood=Ind|Number=Sing|Reflex=Yes|Tense=Past|VerbForm=Fin _ _ SpaceAfter=No
11 , , PUNCT punct _ _ _ _
12 тисяча тисяча NOUN noun:inanim:f:v_naz:&numr Animacy=Inan|Case=Nom|Gender=Fem|Number=Sing|NumType=Card _ _ SemTags=1:abst:quantity:absol:2:abst:quantity&max:3:conc:hum:group:part|SpaceAfter=No
13 ... ... PUNCT punct _ _ _ _
1 А а CCONJ conj:coord _ _ _ _ _
2 треба треба ADV noninfl:&predic _ _ _ _ Uninflect=Yes
3 далі далі ADV adv:compc:&predic Degree=Cmp _ _ _ SemTags=1:dist:2:time
4 воно воно PRON noun:unanim:n:v_naz:&pron:pers:3 Animacy=Anim,Inan|Case=Nom|Gender=Neut|Number=Sing|Person=3|PronType=Prs _ _ _ SemTags=1:conc:deictic
5 - - PUNCT punct _ _ _ _ _
6 озеро озеро NOUN noun:inanim:n:v_naz Animacy=Inan|Case=Nom|Gender=Neut|Number=Sing _ _ _ _
7 Світязь Світязь PROPN noun:inanim:m:v_naz:prop:geo:xp1 Animacy=Inan|Case=Nom|Gender=Masc|NameType=Geo|Number=Sing _ _ _ SemTags=1:conc:loc
8 де де ADV adv:&pron:int:rel PronType=Int|PronType=Rel _ _ _ _
9 я я PRON noun:anim:s:v_naz:&pron:pers:1 Animacy=Anim|Case=Nom|Number=Sing|Person=1|PronType=Prs _ _ _ SemTags=1:conc:hum:deictic
10 затримався затриматися VERB verb:rev:perf:past:m Aspect=Perf|Gender=Masc|Mood=Ind|Number=Sing|Reflex=Yes|Tense=Past|VerbForm=Fin _ _ _ SpaceAfter=No
11 , , PUNCT punct _ _ _ _ _
12 тисяча тисяча NOUN noun:inanim:f:v_naz:&numr Animacy=Inan|Case=Nom|Gender=Fem|Number=Sing|NumType=Card _ _ _ SemTags=1:abst:quantity:absol:2:abst:quantity&max:3:conc:hum:group:part|SpaceAfter=No
13 ... ... PUNCT punct _ _ _ _ _
""".toString()

assertEquals expected, adjustResult(tagged.tagged)
Expand All @@ -126,27 +126,27 @@ class TagTextVerticalOutputTest {
expected =
"""# sent_id = 1
# text = $sent1
1 Шановні шановний ADJ adj:p:v_naz:compb Case=Nom|Degree=Pos|Number=Plur _ _ _
2 колеги колега NOUN noun:anim:p:v_naz Animacy=Anim|Case=Nom|Gender=Fem|Number=Plur _ _ SpaceAfter=No
3 , , PUNCT punct _ _ _ _
4 прошу просити VERB verb:imperf:pres:s:1 Aspect=Imp|Mood=Ind|Number=Sing|Person=1|Tense=Pres|VerbForm=Fin _ _ _
5 вставте вставити VERB verb:perf:impr:p:2 Aspect=Perf|Mood=Imp|Number=Plur|Person=2|VerbForm=Fin _ _ _
6 картки картка NOUN noun:inanim:p:v_zna Animacy=Inan|Case=Acc|Gender=Fem|Number=Plur _ _ SpaceAfter=No
7 , , PUNCT punct _ _ _ _
8 зараз зараз ADV adv:&pron:dem PronType=Dem _ _ _
9 проведемо провести VERB verb:perf:futr:p:1 Aspect=Perf|Mood=Ind|Number=Plur|Person=1|Tense=Fut|VerbForm=Fin _ _ _
10 реєстрацію реєстрація NOUN noun:inanim:f:v_zna Animacy=Inan|Case=Acc|Gender=Fem|Number=Sing _ _ _
11 натисканням натискання NOUN noun:inanim:p:v_dav Animacy=Inan|Case=Dat|Gender=Neut|Number=Plur _ _ _
12 зеленої зелений ADJ adj:f:v_rod:compb Case=Gen|Degree=Pos|Gender=Fem|Number=Sing _ _ _
13 кнопки кнопка NOUN noun:inanim:f:v_rod Animacy=Inan|Case=Gen|Gender=Fem|Number=Sing _ _ _
14 приладів прилад NOUN noun:inanim:p:v_rod Animacy=Inan|Case=Gen|Gender=Masc|Number=Plur _ _ SpaceAfter=No
15 . . PUNCT punct _ _ _ _
1 Шановні шановний ADJ adj:p:v_naz:compb Case=Nom|Degree=Pos|Number=Plur _ _ _ _
2 колеги колега NOUN noun:anim:p:v_naz Animacy=Anim|Case=Nom|Gender=Fem|Number=Plur _ _ _ SpaceAfter=No
3 , , PUNCT punct _ _ _ _ _
4 прошу просити VERB verb:imperf:pres:s:1 Aspect=Imp|Mood=Ind|Number=Sing|Person=1|Tense=Pres|VerbForm=Fin _ _ _ _
5 вставте вставити VERB verb:perf:impr:p:2 Aspect=Perf|Mood=Imp|Number=Plur|Person=2|VerbForm=Fin _ _ _ _
6 картки картка NOUN noun:inanim:p:v_zna Animacy=Inan|Case=Acc|Gender=Fem|Number=Plur _ _ _ SpaceAfter=No
7 , , PUNCT punct _ _ _ _ _
8 зараз зараз ADV adv:&pron:dem PronType=Dem _ _ _ _
9 проведемо провести VERB verb:perf:futr:p:1 Aspect=Perf|Mood=Ind|Number=Plur|Person=1|Tense=Fut|VerbForm=Fin _ _ _ _
10 реєстрацію реєстрація NOUN noun:inanim:f:v_zna Animacy=Inan|Case=Acc|Gender=Fem|Number=Sing _ _ _ _
11 натисканням натискання NOUN noun:inanim:p:v_dav Animacy=Inan|Case=Dat|Gender=Neut|Number=Plur _ _ _ _
12 зеленої зелений ADJ adj:f:v_rod:compb Case=Gen|Degree=Pos|Gender=Fem|Number=Sing _ _ _ _
13 кнопки кнопка NOUN noun:inanim:f:v_rod Animacy=Inan|Case=Gen|Gender=Fem|Number=Sing _ _ _ _
14 приладів прилад NOUN noun:inanim:p:v_rod Animacy=Inan|Case=Gen|Gender=Masc|Number=Plur _ _ _ SpaceAfter=No
15 . . PUNCT punct _ _ _ _ _
# sent_id = 2
# text = $sent2
1 Рада рада NOUN noun:inanim:f:v_naz Animacy=Inan|Case=Nom|Gender=Fem|Number=Sing _ _ _
2 України Україна PROPN noun:inanim:f:v_rod:prop:geo Animacy=Inan|Case=Gen|Gender=Fem|NameType=Geo|Number=Sing _ _ SpaceAfter=No
3 . . PUNCT punct _ _ _ _
1 Рада рада NOUN noun:inanim:f:v_naz Animacy=Inan|Case=Nom|Gender=Fem|Number=Sing _ _ _ _
2 України Україна PROPN noun:inanim:f:v_rod:prop:geo Animacy=Inan|Case=Gen|Gender=Fem|NameType=Geo|Number=Sing _ _ _ SpaceAfter=No
3 . . PUNCT punct _ _ _ _ _
""".toString()

// println tagged.tagged.replace(" ", '\ ')
Expand Down

0 comments on commit b008835

Please sign in to comment.