Make glob matching case-insensitive by default

modules/common/src/main/scala/docspell/common/Glob.scala

@@ -8,7 +8,7 @@ import io.circe.{Decoder, Encoder}
 trait Glob {
 
   /** Matches the input string against this glob. */
-  def matches(in: String): Boolean
+  def matches(caseSensitive: Boolean)(in: String): Boolean
 
   /** If this glob consists of multiple segments, it is the same as
     * `matches`. If it is only a single segment, it is matched against
@@ -25,42 +25,6 @@ trait Glob {
 }
 
 object Glob {
-  private val separator = '/'
-  private val anyChar   = '|'
-
-  val all = new Glob {
-    def matches(in: String)             = true
-    def matchFilenameOrPath(in: String) = true
-    val asString                        = "*"
-  }
-
-  def pattern(pattern: Pattern): Glob =
-    PatternGlob(pattern)
-
-  /** A simple glob supporting `*` and `?`. */
-  final private case class PatternGlob(pattern: Pattern) extends Glob {
-    def matches(in: String): Boolean =
-      pattern.parts
-        .zipWith(Glob.split(in, Glob.separator))(_.matches(_))
-        .forall(identity)
-
-    def matchFilenameOrPath(in: String): Boolean =
-      if (pattern.parts.tail.isEmpty) matches(split(in, separator).last)
-      else matches(in)
-
-    def asString: String =
-      pattern.asString
-  }
-
-  final private case class AnyGlob(globs: NonEmptyList[Glob]) extends Glob {
-    def matches(in: String) =
-      globs.exists(_.matches(in))
-    def matchFilenameOrPath(in: String) =
-      globs.exists(_.matchFilenameOrPath(in))
-    def asString =
-      globs.toList.map(_.asString).mkString(anyChar.toString)
-  }
-
   def apply(in: String): Glob = {
     def single(str: String) =
       PatternGlob(Pattern(split(str, separator).map(makeSegment)))
@@ -75,6 +39,42 @@ object Glob {
       }
   }
 
+  private val separator = '/'
+  private val anyChar   = '|'
+
+  val all = new Glob {
+    def matches(caseSensitive: Boolean)(in: String) = true
+    def matchFilenameOrPath(in: String)             = true
+    val asString                                    = "*"
+  }
+
+  def pattern(pattern: Pattern): Glob =
+    PatternGlob(pattern)
+
+  /** A simple glob supporting `*` and `?`. */
+  final private case class PatternGlob(pattern: Pattern) extends Glob {
+    def matches(caseSensitive: Boolean)(in: String): Boolean =
+      pattern.parts
+        .zipWith(Glob.split(in, Glob.separator))(_.matches(caseSensitive)(_))
+        .forall(identity)
+
+    def matchFilenameOrPath(in: String): Boolean =
+      if (pattern.parts.tail.isEmpty) matches(true)(split(in, separator).last)
+      else matches(true)(in)
+
+    def asString: String =
+      pattern.asString
+  }
+
+  final private case class AnyGlob(globs: NonEmptyList[Glob]) extends Glob {
+    def matches(caseSensitive: Boolean)(in: String) =
+      globs.exists(_.matches(caseSensitive)(in))
+    def matchFilenameOrPath(in: String) =
+      globs.exists(_.matchFilenameOrPath(in))
+    def asString =
+      globs.toList.map(_.asString).mkString(anyChar.toString)
+  }
+
   case class Pattern(parts: NonEmptyList[Segment]) {
     def asString =
       parts.map(_.asString).toList.mkString(separator.toString)
@@ -86,12 +86,12 @@ object Glob {
   }
 
   case class Segment(tokens: NonEmptyList[Token]) {
-    def matches(in: String): Boolean =
-      consume(in).exists(_.isEmpty)
+    def matches(caseSensitive: Boolean)(in: String): Boolean =
+      consume(in, caseSensitive).exists(_.isEmpty)
 
-    def consume(in: String): Option[String] =
+    def consume(in: String, caseSensitive: Boolean): Option[String] =
       tokens.foldLeft(in.some) { (rem, token) =>
-        rem.flatMap(token.consume)
+        rem.flatMap(token.consume(caseSensitive))
       }
 
     def asString: String =
@@ -103,34 +103,47 @@ object Glob {
   }
 
   sealed trait Token {
-    def consume(str: String): Option[String]
+    def consume(caseSensitive: Boolean)(str: String): Option[String]
 
     def asString: String
   }
   object Token {
     case class Literal(asString: String) extends Token {
-      def consume(str: String): Option[String] =
-        if (str.startsWith(asString)) str.drop(asString.length).some
+      def consume(caseSensitive: Boolean)(str: String): Option[String] =
+        if (str.startsWith(asString, caseSensitive)) str.drop(asString.length).some
         else None
     }
     case class Until(value: String) extends Token {
-      def consume(str: String): Option[String] =
+      def consume(caseSensitive: Boolean)(str: String): Option[String] =
         if (value.isEmpty) Some("")
         else
-          str.indexOf(value) match {
-            case -1 => None
-            case n  => str.substring(n + value.length).some
-          }
+          str
+            .findFirst(value, caseSensitive)
+            .map(n => str.substring(n + value.length))
       val asString =
         s"*$value"
     }
     case object Single extends Token {
-      def consume(str: String): Option[String] =
-        if (str.isEmpty()) None
+      def consume(caseSensitive: Boolean)(str: String): Option[String] =
+        if (str.isEmpty) None
         else Some(str.drop(1))
 
       val asString = "?"
     }
+
+    implicit final class StringHelper(val str: String) extends AnyVal {
+      def findFirst(sub: String, caseSensitive: Boolean): Option[Int] = {
+        val vstr = if (caseSensitive) str else str.toLowerCase
+        val vsub = if (caseSensitive) sub else sub.toLowerCase
+        Option(vstr.indexOf(vsub)).filter(_ >= 0)
+      }
+
+      def startsWith(prefix: String, caseSensitive: Boolean): Boolean = {
+        val vstr    = if (caseSensitive) str else str.toLowerCase
+        val vprefix = if (caseSensitive) prefix else prefix.toLowerCase
+        vstr.startsWith(vprefix)
+      }
+    }
   }
 
   private def split(str: String, sep: Char): NonEmptyList[String] =
@@ -139,6 +152,7 @@ object Glob {
       .getOrElse(NonEmptyList.of(str))
 
   private def makeSegment(str: String): Segment = {
+    @annotation.tailrec
     def loop(rem: String, res: List[Token]): List[Token] =
       if (rem.isEmpty) res
       else

modules/common/src/test/scala/docspell/common/GlobTest.scala

@@ -6,8 +6,10 @@ import Glob._
 object GlobTest extends SimpleTestSuite {
 
   test("literals") {
-    assert(Glob.pattern(Pattern(Segment(Token.Literal("hello")))).matches("hello"))
-    assert(!Glob.pattern(Pattern(Segment(Token.Literal("hello")))).matches("hello1"))
+    assert(Glob.pattern(Pattern(Segment(Token.Literal("hello")))).matches(true)("hello"))
+    assert(
+      !Glob.pattern(Pattern(Segment(Token.Literal("hello")))).matches(true)("hello1")
+    )
   }
 
   test("single wildcards 1") {
@@ -16,19 +18,19 @@ object GlobTest extends SimpleTestSuite {
         Pattern(Segment(Token.Literal("s"), Token.Until("p"), Token.Until("t")))
       )
 
-    assert(glob.matches("snapshot"))
-    assert(!glob.matches("snapshots"))
+    assert(glob.matches(true)("snapshot"))
+    assert(!glob.matches(true)("snapshots"))
   }
 
   test("single wildcards 2") {
     val glob =
       Glob.pattern(Pattern(Segment(Token.Literal("test."), Token.Until(""))))
 
-    assert(glob.matches("test.txt"))
-    assert(glob.matches("test.pdf"))
-    assert(glob.matches("test.converted.pdf"))
-    assert(!glob.matches("test1.txt"))
-    assert(!glob.matches("atest.txt"))
+    assert(glob.matches(true)("test.txt"))
+    assert(glob.matches(true)("test.pdf"))
+    assert(glob.matches(true)("test.converted.pdf"))
+    assert(!glob.matches(true)("test1.txt"))
+    assert(!glob.matches(true)("atest.txt"))
   }
 
   test("single parsing") {
@@ -60,12 +62,12 @@ object GlobTest extends SimpleTestSuite {
   }
 
   test("with splitting") {
-    assert(Glob("a/b/*").matches("a/b/hello"))
-    assert(!Glob("a/b/*").matches("/a/b/hello"))
-    assert(Glob("/a/b/*").matches("/a/b/hello"))
-    assert(!Glob("/a/b/*").matches("a/b/hello"))
-    assert(!Glob("*/a/b/*").matches("a/b/hello"))
-    assert(Glob("*/a/b/*").matches("test/a/b/hello"))
+    assert(Glob("a/b/*").matches(true)("a/b/hello"))
+    assert(!Glob("a/b/*").matches(true)("/a/b/hello"))
+    assert(Glob("/a/b/*").matches(true)("/a/b/hello"))
+    assert(!Glob("/a/b/*").matches(true)("a/b/hello"))
+    assert(!Glob("*/a/b/*").matches(true)("a/b/hello"))
+    assert(Glob("*/a/b/*").matches(true)("test/a/b/hello"))
   }
 
   test("asString") {
@@ -79,9 +81,9 @@ object GlobTest extends SimpleTestSuite {
   }
 
   test("simple matches") {
-    assert(Glob("/test.*").matches("/test.pdf"))
-    assert(!Glob("/test.*").matches("test.pdf"))
-    assert(!Glob("test.*").matches("/test.pdf"))
+    assert(Glob("/test.*").matches(true)("/test.pdf"))
+    assert(!Glob("/test.*").matches(true)("test.pdf"))
+    assert(!Glob("test.*").matches(true)("/test.pdf"))
   }
 
   test("matchFilenameOrPath") {
@@ -100,12 +102,24 @@ object GlobTest extends SimpleTestSuite {
   }
 
   test("anyglob") {
-    assert(Glob("*.pdf|*.txt").matches("test.pdf"))
-    assert(Glob("*.pdf|*.txt").matches("test.txt"))
-    assert(!Glob("*.pdf|*.txt").matches("test.xls"))
-    assert(Glob("*.pdf | *.txt").matches("test.pdf"))
-    assert(Glob("*.pdf | mail.html").matches("test.pdf"))
-    assert(Glob("*.pdf | mail.html").matches("mail.html"))
-    assert(!Glob("*.pdf | mail.html").matches("test.docx"))
+    assert(Glob("*.pdf|*.txt").matches(true)("test.pdf"))
+    assert(Glob("*.pdf|*.txt").matches(true)("test.txt"))
+    assert(!Glob("*.pdf|*.txt").matches(true)("test.xls"))
+    assert(Glob("*.pdf | *.txt").matches(true)("test.pdf"))
+    assert(Glob("*.pdf | mail.html").matches(true)("test.pdf"))
+    assert(Glob("*.pdf | mail.html").matches(true)("mail.html"))
+    assert(!Glob("*.pdf | mail.html").matches(true)("test.docx"))
+  }
+
+  test("case insensitive") {
+    assert(Glob("*hello*").matches(false)("hello world"))
+    assert(Glob("*hello*").matches(false)("world hello"))
+    assert(Glob("*hello*").matches(false)("Hello world"))
+    assert(Glob("*hello*").matches(false)("world Hello"))
+    assert(Glob("*hello*").matches(false)("World Hello"))
+    assert(Glob("*hello*").matches(false)("Hello World"))
+    assert(Glob("*Hello*").matches(false)("world hello"))
+    assert(Glob("*heLLo*").matches(false)("Hello world"))
+    assert(Glob("*hellO*").matches(false)("world Hello"))
   }
 }

modules/joex/src/main/scala/docspell/joex/mail/ReadMail.scala

@@ -51,11 +51,11 @@ object ReadMail {
       (Stream
         .eval(bodyEntry)
         .flatMap(e => Stream.emits(e.toSeq))
-        .filter(a => glob.matches(a.name)) ++
+        .filter(a => glob.matches(caseSensitive = false)(a.name)) ++
         Stream
           .eval(TnefExtract.replace(mail))
           .flatMap(m => Stream.emits(m.attachments.all))
-          .filter(a => a.filename.exists(glob.matches))
+          .filter(a => a.filename.exists(glob.matches(caseSensitive = false)))
           .map(a =>
             Binary(a.filename.getOrElse("noname"), a.mimeType.toLocal, a.content)
           ))

modules/joex/src/main/scala/docspell/joex/process/ExtractArchive.scala

@@ -243,7 +243,9 @@ object ExtractArchive {
       )
 
     def filterNames(filter: Glob): Extracted =
-      copy(files = files.filter(ra => filter.matches(ra.name.getOrElse(""))))
+      copy(files =
+        files.filter(ra => filter.matches(caseSensitive = false)(ra.name.getOrElse("")))
+      )
 
     def setMeta(m: MetaProposal): Extracted =
       setMeta(MetaProposalList.of(m))

modules/joex/src/main/scala/docspell/joex/scanmailbox/ScanMailboxTask.scala

@@ -182,7 +182,7 @@ object ScanMailboxTask {
       ctx.args.subjectFilter match {
         case Some(sf) =>
           def check(mh: MailHeader): F[Option[MailHeader]] =
-            if (sf.matches(mh.subject))
+            if (sf.matches(caseSensitive = false)(mh.subject))
               ctx.logger.debug(
                 s"Including mail '${mh.subject}', it matches the filter."
               ) *> Option(mh).pure[F]