Add latvian language

This commit is contained in:
Eike Kettner
2021-03-09 00:19:33 +01:00
parent e4ef299582
commit 9991ad5fcc
6 changed files with 74 additions and 1 deletions

View File

@ -17,6 +17,7 @@ object DateFind {
def findDates(text: String, lang: Language): Stream[Pure, NerDateLabel] =
TextSplitter
.splitToken(text, " \t.,\n\r/".toSet)
.filter(w => lang != Language.Latvian || w.value != "gada")
.sliding(3)
.filter(_.length == 3)
.flatMap(q =>
@ -55,6 +56,10 @@ object DateFind {
case ((m, d), y) =>
List(SimpleDate(y, m, d))
}
def lavLong =
(readYear >> readDay >> readMonth(Language.Latvian)).map { case ((y, d), m) =>
List(SimpleDate(y, m, d))
}
// ymd ✔, ydm, dmy ✔, dym, myd, mdy ✔
def fromParts(parts: List[Word], lang: Language): List[SimpleDate] = {
@ -77,6 +82,7 @@ object DateFind {
case Language.Russian => dmy.or(ymd).or(mdy)
case Language.Swedish => ymd.or(dmy).or(mdy)
case Language.Dutch => dmy.or(ymd).or(mdy)
case Language.Latvian => dmy.or(lavLong).or(ymd)
}
p.read(parts) match {
case Result.Success(sds, _) =>

View File

@ -42,6 +42,8 @@ object MonthName {
finnish
case Language.Russian =>
russian
case Language.Latvian =>
latvian
}
private val numbers = List(
@ -267,4 +269,19 @@ object MonthName {
List("nov", "november"),
List("dec", "december")
)
private val latvian = List(
List("janvāris", "janv."),
List("februāris", "febr."),
List("marts"),
List("aprīlis", "apr."),
List("maijs"),
List("jūnijs", "jūn."),
List("jūlijs", "jūl."),
List("augusts", "aug."),
List("septembris", "sept."),
List("oktobris", "okt."),
List("novembris", "nov."),
List("decembris", "dec.")
)
}

View File

@ -103,4 +103,36 @@ object DateFindSpec extends SimpleTestSuite {
)
}
test("find latvian dates") {
assertEquals(
DateFind.findDates("on 2020. gada 30. jūlijs there", Language.Latvian).toVector,
Vector(
NerDateLabel(
LocalDate.of(2020, 7, 30),
NerLabel("2020. gada 30. jūlijs", NerTag.Date, 3, 24)
)
)
)
assertEquals(
DateFind.findDates("Lai gan 30.07.2020", Language.Latvian).toVector,
Vector(
NerDateLabel(
LocalDate.of(2020, 7, 30),
NerLabel("30.07.2020", NerTag.Date, 8, 18)
)
)
)
assertEquals(
DateFind
.findDates("Es nevaru šodien 2020.gada 30.oktobris iet uz", Language.Latvian)
.toVector,
Vector(
NerDateLabel(
LocalDate.of(2020, 10, 30),
NerLabel("2020.gada 30.oktobris", NerTag.Date, 17, 38)
)
)
)
}
}