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

humanbytes.go

@@ -1,38 +0,0 @@
-package util
-
-import (
-	"fmt"
-	"math"
-)
-
-const bytesPerUnit = 1024
-
-var byteUnits = []string{"KiB", "MiB", "GiB", "TiB", "PiB", "EiB"}
-
-// HumanBytes writes a byte count the way a person would read it, using powers
-// of 1024: "0 B", "1023 B", "1.0 KiB", "1.5 MiB", "16.0 EiB". Counts below 1024
-// are given as whole bytes, and anything larger gets one digit after the
-// decimal point.
-func HumanBytes(bytes uint64) string {
-	if bytes < bytesPerUnit {
-		return fmt.Sprintf("%d B", bytes)
-	}
-
-	value := float64(bytes)
-	unit := -1
-
-	for value >= bytesPerUnit && unit < len(byteUnits)-1 {
-		value /= bytesPerUnit
-		unit++
-	}
-
-	// Rounding happens after the unit has been chosen, so a count just short
-	// of the next threshold would otherwise come out as "1024.0 KiB" rather
-	// than "1.0 MiB".
-	if unit < len(byteUnits)-1 && math.Round(value*10)/10 >= bytesPerUnit {
-		value /= bytesPerUnit
-		unit++
-	}
-
-	return fmt.Sprintf("%.1f %s", value, byteUnits[unit])
-}

humanbytes_test.go

@@ -1,33 +0,0 @@
-package util
-
-import "testing"
-
-func TestHumanBytes(t *testing.T) {
-	tests := []struct {
-		name     string
-		input    uint64
-		expected string
-	}{
-		{"zero", 0, "0 B"},
-		{"one byte", 1, "1 B"},
-		{"just below a kibibyte", 1023, "1023 B"},
-		{"exactly a kibibyte", 1024, "1.0 KiB"},
-		{"half a kibibyte more", 1536, "1.5 KiB"},
-		{"rounds up into the next unit", 1048575, "1.0 MiB"},
-		{"exactly a mebibyte", 1 << 20, "1.0 MiB"},
-		{"a gibibyte", 1 << 30, "1.0 GiB"},
-		{"a tebibyte", 1 << 40, "1.0 TiB"},
-		{"a pebibyte", 1 << 50, "1.0 PiB"},
-		{"an exbibyte", 1 << 60, "1.0 EiB"},
-		{"the largest count there is", ^uint64(0), "16.0 EiB"},
-	}
-
-	for _, test := range tests {
-		t.Run(test.name, func(t *testing.T) {
-			got := HumanBytes(test.input)
-			if got != test.expected {
-				t.Errorf("expected %q got %q", test.expected, 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)
+			}
+		})
+	}
+}