TheAnachronism/docspell
1
0
Fork 0
mirror of https://github.com/TheAnachronism/docspell.git synced 2025-06-22 02:18:26 +00:00
Code Issues Packages Projects Releases Wiki Activity

Autoformat

Browse Source
This commit is contained in:
Eike Kettner
2020-09-09 00:29:32 +02:00
parent 570b7de43f
commit c658677032
34 changed files with 279 additions and 319 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
modules/extract/src/main/scala/docspell/extract/poi/PoiExtract.scala
View File

@ -44,14 +44,14 @@ object PoiExtract {
getDocx(data)
case PoiType.msoffice =>
EitherT(getDoc[F](data))
.recoverWith({
case _ => EitherT(getXls[F](data))
.recoverWith({ case _ =>
EitherT(getXls[F](data))
})
.value
case PoiType.ooxml =>
EitherT(getDocx[F](data))
.recoverWith({
case _ => EitherT(getXlsx[F](data))
.recoverWith({ case _ =>
EitherT(getXlsx[F](data))
})
.value
case mt =>

15
modules/extract/src/test/scala/docspell/extract/odf/OdfExtractTest.scala
View File

@ -14,15 +14,14 @@ object OdfExtractTest extends SimpleTestSuite {
)
test("test extract from odt") {
files.foreach {
case (file, len) =>
val is = file.toJavaUrl.map(_.openStream()).fold(sys.error, identity)
val str1 = OdfExtract.get(is).fold(throw _, identity)
assertEquals(str1.length, len)
files.foreach { case (file, len) =>
val is = file.toJavaUrl.map(_.openStream()).fold(sys.error, identity)
val str1 = OdfExtract.get(is).fold(throw _, identity)
assertEquals(str1.length, len)
val data = file.readURL[IO](8192, blocker)
val str2 = OdfExtract.get[IO](data).unsafeRunSync().fold(throw _, identity)
assertEquals(str2, str1)
val data = file.readURL[IO](8192, blocker)
val str2 = OdfExtract.get[IO](data).unsafeRunSync().fold(throw _, identity)
assertEquals(str2, str1)
}
}

26
modules/extract/src/test/scala/docspell/extract/pdfbox/PdfboxExtractTest.scala
View File

@ -14,24 +14,22 @@ object PdfboxExtractTest extends SimpleTestSuite {
)
test("extract text from text PDFs by inputstream") {
textPDFs.foreach {
case (file, txt) =>
val url = file.toJavaUrl.fold(sys.error, identity)
val str = PdfboxExtract.getText(url.openStream()).fold(throw _, identity)
val received = removeFormatting(str.value)
val expect = removeFormatting(txt)
assertEquals(received, expect)
textPDFs.foreach { case (file, txt) =>
val url = file.toJavaUrl.fold(sys.error, identity)
val str = PdfboxExtract.getText(url.openStream()).fold(throw _, identity)
val received = removeFormatting(str.value)
val expect = removeFormatting(txt)
assertEquals(received, expect)
}
}
test("extract text from text PDFs via Stream") {
textPDFs.foreach {
case (file, txt) =>
val data = file.readURL[IO](8192, blocker)
val str = PdfboxExtract.getText(data).unsafeRunSync().fold(throw _, identity)
val received = removeFormatting(str.value)
val expect = removeFormatting(txt)
assertEquals(received, expect)
textPDFs.foreach { case (file, txt) =>
val data = file.readURL[IO](8192, blocker)
val str = PdfboxExtract.getText(data).unsafeRunSync().fold(throw _, identity)
val received = removeFormatting(str.value)
val expect = removeFormatting(txt)
assertEquals(received, expect)
}
}

29
modules/extract/src/test/scala/docspell/extract/poi/PoiExtractTest.scala
View File

@ -17,23 +17,22 @@ object PoiExtractTest extends SimpleTestSuite {
)
test("extract text from ms office files") {
officeFiles.foreach {
case (file, len) =>
val str1 = PoiExtract
.get[IO](file.readURL[IO](8192, blocker), MimeTypeHint.none)
.unsafeRunSync()
.fold(throw _, identity)
officeFiles.foreach { case (file, len) =>
val str1 = PoiExtract
.get[IO](file.readURL[IO](8192, blocker), MimeTypeHint.none)
.unsafeRunSync()
.fold(throw _, identity)
val str2 = PoiExtract
.get[IO](
file.readURL[IO](8192, blocker),
MimeTypeHint(Some(file.path.segments.last), None)
)
.unsafeRunSync()
.fold(throw _, identity)
val str2 = PoiExtract
.get[IO](
file.readURL[IO](8192, blocker),
MimeTypeHint(Some(file.path.segments.last), None)
)
.unsafeRunSync()
.fold(throw _, identity)
assertEquals(str1, str2)
assertEquals(str1.length, len)
assertEquals(str1, str2)
assertEquals(str1.length, len)
}
}
}