add Slugify for making URL-safe slugs out of text

Slugify lowercases text and keeps only ASCII letters and digits, turning every
run of other characters into a single hyphen with none left at either end.
Comes with a doc comment and table-driven tests. (closes #9)

Model: opus-5

exists.go

@@ -1,31 +0,0 @@
-package util
-
-import "os"
-
-// FileExists reports whether there is a regular file at path. A directory, a
-// device node or a socket gives false, and so does anything the process is not
-// allowed to look at, since from the caller's point of view there is no usable
-// file there either way.
-//
-// Symbolic links are followed, so a link pointing at a regular file gives true
-// and a link pointing at nothing gives false.
-func FileExists(path string) bool {
-	info, err := os.Stat(path)
-	if err != nil {
-		return false
-	}
-
-	return info.Mode().IsRegular()
-}
-
-// DirExists reports whether there is a directory at path. As with FileExists,
-// anything that cannot be looked at gives false, and symbolic links are
-// followed.
-func DirExists(path string) bool {
-	info, err := os.Stat(path)
-	if err != nil {
-		return false
-	}
-
-	return info.IsDir()
-}

exists_test.go

@@ -1,57 +0,0 @@
-package util
-
-import (
-	"os"
-	"path/filepath"
-	"testing"
-)
-
-func TestFileExistsAndDirExists(t *testing.T) {
-	root := t.TempDir()
-
-	file := filepath.Join(root, "a-file")
-	if err := os.WriteFile(file, []byte("contents"), 0644); err != nil {
-		t.Fatalf("could not write the test file: %v", err)
-	}
-
-	directory := filepath.Join(root, "a-directory")
-	if err := os.Mkdir(directory, 0755); err != nil {
-		t.Fatalf("could not make the test directory: %v", err)
-	}
-
-	linkToFile := filepath.Join(root, "link-to-file")
-	if err := os.Symlink(file, linkToFile); err != nil {
-		t.Fatalf("could not make the test link: %v", err)
-	}
-
-	brokenLink := filepath.Join(root, "broken-link")
-	if err := os.Symlink(filepath.Join(root, "nothing-here"), brokenLink); err != nil {
-		t.Fatalf("could not make the broken test link: %v", err)
-	}
-
-	tests := []struct {
-		name          string
-		path          string
-		expectedFile  bool
-		expectedIsDir bool
-	}{
-		{"a regular file", file, true, false},
-		{"a directory", directory, false, true},
-		{"a link to a file", linkToFile, true, false},
-		{"a link to nothing", brokenLink, false, false},
-		{"a path that is not there", filepath.Join(root, "missing"), false, false},
-		{"a path whose parent is not there", filepath.Join(root, "missing", "deeper"), false, false},
-		{"the empty path", "", false, false},
-	}
-
-	for _, test := range tests {
-		t.Run(test.name, func(t *testing.T) {
-			if got := FileExists(test.path); got != test.expectedFile {
-				t.Errorf("FileExists: expected %v got %v", test.expectedFile, got)
-			}
-			if got := DirExists(test.path); got != test.expectedIsDir {
-				t.Errorf("DirExists: expected %v got %v", test.expectedIsDir, got)
-			}
-		})
-	}
-}

slugify.go

@@ -0,0 +1,36 @@
+package util
+
+import "strings"
+
+// Slugify returns input as lowercase ASCII letters, digits and hyphens, which
+// is safe to put in a URL path, a filename or a fragment identifier. Every run
+// of other characters becomes a single hyphen, and the result never starts or
+// ends with one.
+//
+// Characters outside ASCII are treated as separators rather than being folded
+// to their nearest ASCII relative, because doing that properly needs a Unicode
+// table that this package does not carry. Text with no ASCII letters or digits
+// in it therefore slugifies to an empty string.
+func Slugify(input string) string {
+	var out strings.Builder
+	pendingHyphen := false
+
+	for _, character := range strings.ToLower(input) {
+		if (character >= 'a' && character <= 'z') || (character >= '0' && character <= '9') {
+			if pendingHyphen {
+				out.WriteByte('-')
+				pendingHyphen = false
+			}
+			out.WriteRune(character)
+			continue
+		}
+
+		// Remember that a separator was seen, but only write a hyphen once
+		// another usable character turns up, so nothing trails off the end.
+		if out.Len() > 0 {
+			pendingHyphen = true
+		}
+	}
+
+	return out.String()
+}

slugify_test.go

@@ -0,0 +1,34 @@
+package util
+
+import "testing"
+
+func TestSlugify(t *testing.T) {
+	tests := []struct {
+		name     string
+		input    string
+		expected string
+	}{
+		{"already a slug", "hello-world", "hello-world"},
+		{"mixed case", "Hello World", "hello-world"},
+		{"punctuation", "What's new?", "what-s-new"},
+		{"run of separators", "one   ---   two", "one-two"},
+		{"leading separators", "///one", "one"},
+		{"trailing separators", "one///", "one"},
+		{"separators at both ends", "  --one two--  ", "one-two"},
+		{"digits are kept", "Go 1.14 release", "go-1-14-release"},
+		{"underscores are separators", "some_name_here", "some-name-here"},
+		{"nothing usable", "!!! ???", ""},
+		{"empty string", "", ""},
+		{"outside ascii", "Grüße, Welt", "gr-e-welt"},
+		{"only characters outside ascii", "日本語", ""},
+	}
+
+	for _, test := range tests {
+		t.Run(test.name, func(t *testing.T) {
+			got := Slugify(test.input)
+			if got != test.expected {
+				t.Errorf("expected %q got %q", test.expected, got)
+			}
+		})
+	}
+}