Add french language and upgrade stanford-nlp to 4.0.0

This commit is contained in:
Eike Kettner
2020-04-21 23:33:15 +02:00
parent 9945b43266
commit fdb46da26d
13 changed files with 208 additions and 54 deletions

View File

@ -31,7 +31,7 @@ object Dependencies {
val PostgresVersion = "42.2.16"
val PureConfigVersion = "0.13.0"
val Slf4jVersion = "1.7.30"
val StanfordNlpVersion = "3.9.2"
val StanfordNlpVersion = "4.0.0"
val TikaVersion = "1.24.1"
val YamuscaVersion = "0.6.2"
val SwaggerUIVersion = "3.32.3"
@ -135,11 +135,16 @@ object Dependencies {
)
val stanfordNlpModels = Seq(
("edu.stanford.nlp" % "stanford-corenlp" % StanfordNlpVersion)
.classifier("models"),
("edu.stanford.nlp" % "stanford-corenlp" % StanfordNlpVersion)
.classifier("models-german"),
("edu.stanford.nlp" % "stanford-corenlp" % StanfordNlpVersion).classifier(
"models-english"
)
("edu.stanford.nlp" % "stanford-corenlp" % StanfordNlpVersion)
.classifier("models-french"),
("edu.stanford.nlp" % "stanford-corenlp" % StanfordNlpVersion)
.classifier(
"models-english"
)
)
val tika = Seq(

View File

@ -68,7 +68,18 @@ object NerModelsPlugin extends AutoPlugin {
}
private val nerModels = List(
"german.conll.germeval2014.hgc_175m_600.crf.ser.gz",
"english.all.3class.distsim.crf.ser.gz"
"german.distsim.crf.ser.gz",
"english.conll.4class.distsim.crf.ser.gz",
"french-wikiner-4class.crf.ser.gz",
"french-mwt-statistical.tsv",
"french-mwt.tagger",
"french-mwt.tsv",
"german-mwt.tsv",
"german-ud.tagger",
"german-ud.tagger.props",
"french-ud.tagger",
"french-ud.tagger.props",
"english-left3words-distsim.tagger",
"english-left3words-distsim.tagger.props"
)
}