Merge pull request #8001 from kaushalmodi/fix-isupper-islower-try2

Make isUpper (and variants) work for strings with non-alpha chars

changelog.md

@@ -42,6 +42,12 @@
 - ``math.`mod` `` for floats now behaves the same as ``mod`` for integers
   (previously it used floor division like Python). Use ``math.floorMod`` for the old behavior.
 
+- For string inputs, ``unicode.isUpper`` and ``unicode.isLower`` now require a
+  second mandatory parameter ``skipNonAlpha``.
+
+- For string inputs, ``strutils.isUpperAscii`` and ``strutils.isLowerAscii`` now
+  require a second mandatory parameter ``skipNonAlpha``.
+
 #### Breaking changes in the compiler
 
 - The undocumented ``#? braces`` parsing mode was removed.

lib/pure/strutils.nim

@@ -150,23 +150,52 @@ proc isSpaceAscii*(s: string): bool {.noSideEffect, procvar,
   ## characters and there is at least one character in `s`.
   isImpl isSpaceAscii
 
-proc isLowerAscii*(s: string): bool {.noSideEffect, procvar,
-  rtl, extern: "nsuIsLowerAsciiStr".} =
-  ## Checks whether or not `s` contains all lower case characters.
-  ##
-  ## This checks ASCII characters only.
-  ## Returns true if all characters in `s` are lower case
-  ## and there is at least one character  in `s`.
-  isImpl isLowerAscii
+template isCaseImpl(s, charProc, skipNonAlpha) =
+  var hasAtleastOneAlphaChar = false
+  if s.len == 0: return false
+  for c in s:
+    if skipNonAlpha:
+      var charIsAlpha = c.isAlphaAscii()
+      if not hasAtleastOneAlphaChar:
+        hasAtleastOneAlphaChar = charIsAlpha
+      if charIsAlpha and (not charProc(c)):
+        return false
+    else:
+      if not charProc(c):
+        return false
+  return if skipNonAlpha: hasAtleastOneAlphaChar else: true
 
-proc isUpperAscii*(s: string): bool {.noSideEffect, procvar,
-  rtl, extern: "nsuIsUpperAsciiStr".} =
-  ## Checks whether or not `s` contains all upper case characters.
+proc isLowerAscii*(s: string, skipNonAlpha: bool): bool =
+  ## Checks whether ``s`` is lower case.
   ##
   ## This checks ASCII characters only.
-  ## Returns true if all characters in `s` are upper case
-  ## and there is at least one character in `s`.
-  isImpl isUpperAscii
+  ##
+  ## If ``skipNonAlpha`` is true, returns true if all alphabetical
+  ## characters in ``s`` are lower case.  Returns false if none of the
+  ## characters in ``s`` are alphabetical.
+  ##
+  ## If ``skipNonAlpha`` is false, returns true only if all characters
+  ## in ``s`` are alphabetical and lower case.
+  ##
+  ## For either value of ``skipNonAlpha``, returns false if ``s`` is
+  ## an empty string.
+  isCaseImpl(s, isLowerAscii, skipNonAlpha)
+
+proc isUpperAscii*(s: string, skipNonAlpha: bool): bool =
+  ## Checks whether ``s`` is upper case.
+  ##
+  ## This checks ASCII characters only.
+  ##
+  ## If ``skipNonAlpha`` is true, returns true if all alphabetical
+  ## characters in ``s`` are upper case.  Returns false if none of the
+  ## characters in ``s`` are alphabetical.
+  ##
+  ## If ``skipNonAlpha`` is false, returns true only if all characters
+  ## in ``s`` are alphabetical and upper case.
+  ##
+  ## For either value of ``skipNonAlpha``, returns false if ``s`` is
+  ## an empty string.
+  isCaseImpl(s, isUpperAscii, skipNonAlpha)
 
 proc toLowerAscii*(c: char): char {.noSideEffect, procvar,
   rtl, extern: "nsuToLowerAsciiChar".} =
@@ -2516,19 +2545,34 @@ when isMainModule:
     doAssert(not isLowerAscii('A'))
     doAssert(not isLowerAscii('5'))
     doAssert(not isLowerAscii('&'))
+    doAssert(not isLowerAscii(' '))
 
-    doAssert isLowerAscii("abcd")
-    doAssert(not isLowerAscii("abCD"))
-    doAssert(not isLowerAscii("33aa"))
+    doAssert isLowerAscii("abcd", false)
+    doAssert(not isLowerAscii("33aa", false))
+    doAssert(not isLowerAscii("a b", false))
+
+    doAssert(not isLowerAscii("abCD", true))
+    doAssert isLowerAscii("33aa", true)
+    doAssert isLowerAscii("a b", true)
+    doAssert isLowerAscii("1, 2, 3 go!", true)
+    doAssert(not isLowerAscii(" ", true))
+    doAssert(not isLowerAscii("(*&#@(^#$ ", true)) # None of the string chars are alphabets
 
     doAssert isUpperAscii('A')
     doAssert(not isUpperAscii('b'))
     doAssert(not isUpperAscii('5'))
     doAssert(not isUpperAscii('%'))
 
-    doAssert isUpperAscii("ABC")
-    doAssert(not isUpperAscii("AAcc"))
-    doAssert(not isUpperAscii("A#$"))
+    doAssert isUpperAscii("ABC", false)
+    doAssert(not isUpperAscii("A#$", false))
+    doAssert(not isUpperAscii("A B", false))
+
+    doAssert(not isUpperAscii("AAcc", true))
+    doAssert isUpperAscii("A#$", true)
+    doAssert isUpperAscii("A B", true)
+    doAssert isUpperAscii("1, 2, 3 GO!", true)
+    doAssert(not isUpperAscii(" ", true))
+    doAssert(not isUpperAscii("(*&#@(^#$ ", true)) # None of the string chars are alphabets
 
     doAssert rsplit("foo bar", seps=Whitespace) == @["foo", "bar"]
     doAssert rsplit(" foo bar", seps=Whitespace, maxsplit=1) == @[" foo", "bar"]
@@ -2601,4 +2645,3 @@ bar
   nonStaticTests()
   staticTests()
   static: staticTests()
-

lib/pure/unicode.nim

@@ -1392,7 +1392,7 @@ proc isCombining*(c: Rune): bool {.rtl, extern: "nuc$1", procvar.} =
     (c >= 0xfe20 and c <= 0xfe2f))
 
 template runeCheck(s, runeProc) =
-  ## Common code for rune.isLower, rune.isUpper, etc
+  ## Common code for isAlpha and isSpace.
   result = if len(s) == 0: false else: true
 
   var
@@ -1403,16 +1403,6 @@ template runeCheck(s, runeProc) =
     fastRuneAt(s, i, rune, doInc=true)
     result = runeProc(rune) and result
 
-proc isUpper*(s: string): bool {.noSideEffect, procvar,
-  rtl, extern: "nuc$1Str".} =
-  ## Returns true iff `s` contains all upper case unicode characters.
-  runeCheck(s, isUpper)
-
-proc isLower*(s: string): bool {.noSideEffect, procvar,
-  rtl, extern: "nuc$1Str".} =
-  ## Returns true iff `s` contains all lower case unicode characters.
-  runeCheck(s, isLower)
-
 proc isAlpha*(s: string): bool {.noSideEffect, procvar,
   rtl, extern: "nuc$1Str".} =
   ## Returns true iff `s` contains all alphabetic unicode characters.
@@ -1423,6 +1413,56 @@ proc isSpace*(s: string): bool {.noSideEffect, procvar,
   ## Returns true iff `s` contains all whitespace unicode characters.
   runeCheck(s, isWhiteSpace)
 
+template runeCaseCheck(s, runeProc, skipNonAlpha) =
+  ## Common code for rune.isLower and rune.isUpper.
+  if len(s) == 0: return false
+
+  var
+    i = 0
+    rune: Rune
+    hasAtleastOneAlphaRune = false
+
+  while i < len(s):
+    fastRuneAt(s, i, rune, doInc=true)
+    if skipNonAlpha:
+      var runeIsAlpha = isAlpha(rune)
+      if not hasAtleastOneAlphaRune:
+        hasAtleastOneAlphaRune = runeIsAlpha
+      if runeIsAlpha and (not runeProc(rune)):
+        return false
+    else:
+      if not runeProc(rune):
+        return false
+  return if skipNonAlpha: hasAtleastOneAlphaRune else: true
+
+proc isLower*(s: string, skipNonAlpha: bool): bool =
+  ## Checks whether ``s`` is lower case.
+  ##
+  ## If ``skipNonAlpha`` is true, returns true if all alphabetical
+  ## runes in ``s`` are lower case.  Returns false if none of the
+  ## runes in ``s`` are alphabetical.
+  ##
+  ## If ``skipNonAlpha`` is false, returns true only if all runes in
+  ## ``s`` are alphabetical and lower case.
+  ##
+  ## For either value of ``skipNonAlpha``, returns false if ``s`` is
+  ## an empty string.
+  runeCaseCheck(s, isLower, skipNonAlpha)
+
+proc isUpper*(s: string, skipNonAlpha: bool): bool =
+  ## Checks whether ``s`` is upper case.
+  ##
+  ## If ``skipNonAlpha`` is true, returns true if all alphabetical
+  ## runes in ``s`` are upper case.  Returns false if none of the
+  ## runes in ``s`` are alphabetical.
+  ##
+  ## If ``skipNonAlpha`` is false, returns true only if all runes in
+  ## ``s`` are alphabetical and upper case.
+  ##
+  ## For either value of ``skipNonAlpha``, returns false if ``s`` is
+  ## an empty string.
+  runeCaseCheck(s, isUpper, skipNonAlpha)
+
 template convertRune(s, runeProc) =
   ## Convert runes in `s` using `runeProc` as the converter.
   result = newString(len(s))
@@ -1755,25 +1795,39 @@ when isMainModule:
   doAssert(not isSpace(""))
   doAssert(not isSpace("ΑΓc   \td"))
 
-  doAssert isLower("a")
-  doAssert isLower("γ")
-  doAssert(not isLower("Γ"))
-  doAssert(not isLower("4"))
-  doAssert(not isLower(""))
+  doAssert(not isLower(' '.Rune))
 
-  doAssert isLower("abcdγ")
-  doAssert(not isLower("abCDΓ"))
-  doAssert(not isLower("33aaΓ"))
+  doAssert isLower("a", false)
+  doAssert isLower("γ", true)
+  doAssert(not isLower("Γ", false))
+  doAssert(not isLower("4", true))
+  doAssert(not isLower("", false))
+  doAssert isLower("abcdγ", false)
+  doAssert(not isLower("33aaΓ", false))
+  doAssert(not isLower("a b", false))
 
-  doAssert isUpper("Γ")
-  doAssert(not isUpper("b"))
-  doAssert(not isUpper("α"))
-  doAssert(not isUpper("✓"))
-  doAssert(not isUpper(""))
+  doAssert(not isLower("abCDΓ", true))
+  doAssert isLower("a b", true)
+  doAssert isLower("1, 2, 3 go!", true)
+  doAssert(not isLower(" ", true))
+  doAssert(not isLower("(*&#@(^#$✓ ", true)) # None of the string runes are alphabets
 
-  doAssert isUpper("ΑΒΓ")
-  doAssert(not isUpper("AAccβ"))
-  doAssert(not isUpper("A#$β"))
+  doAssert(not isUpper(' '.Rune))
+
+  doAssert isUpper("Γ", false)
+  doAssert(not isUpper("α", false))
+  doAssert(not isUpper("", false))
+  doAssert isUpper("ΑΒΓ", false)
+  doAssert(not isUpper("A#$β", false))
+  doAssert(not isUpper("A B", false))
+
+  doAssert(not isUpper("b", true))
+  doAssert(not isUpper("✓", true))
+  doAssert(not isUpper("AAccβ", true))
+  doAssert isUpper("A B", true)
+  doAssert isUpper("1, 2, 3 GO!", true)
+  doAssert(not isUpper(" ", true))
+  doAssert(not isUpper("(*&#@(^#$✓ ", true)) # None of the string runes are alphabets
 
   doAssert toUpper("Γ") == "Γ"
   doAssert toUpper("b") == "B"