refactor(natural): replace maruel/natural with custom natural sort implementation

This commit is contained in:
Deluan
2026-03-14 09:59:51 -04:00
parent 55331b5fd9
commit d79b812467
3 changed files with 215 additions and 1 deletions
+1 -1
View File
@@ -13,13 +13,13 @@ import (
"time" "time"
"github.com/Masterminds/squirrel" "github.com/Masterminds/squirrel"
"github.com/maruel/natural"
"github.com/navidrome/navidrome/conf" "github.com/navidrome/navidrome/conf"
"github.com/navidrome/navidrome/core" "github.com/navidrome/navidrome/core"
"github.com/navidrome/navidrome/core/external" "github.com/navidrome/navidrome/core/external"
"github.com/navidrome/navidrome/core/ffmpeg" "github.com/navidrome/navidrome/core/ffmpeg"
"github.com/navidrome/navidrome/log" "github.com/navidrome/navidrome/log"
"github.com/navidrome/navidrome/model" "github.com/navidrome/navidrome/model"
"github.com/navidrome/navidrome/utils/natural"
) )
type albumArtworkReader struct { type albumArtworkReader struct {
+98
View File
@@ -0,0 +1,98 @@
// Package natural provides natural (alphanumeric) string comparison.
// When both strings have digit sequences at the same position, they are
// compared numerically (so "file2" < "file10"); otherwise bytes are
// compared one-by-one. No allocations are made.
package natural
import "strings"
// Compare returns a negative value if a < b, zero if a == b,
// or a positive value if a > b using natural sort ordering.
//
// When two numeric segments are numerically equal (e.g. "01" vs "1"),
// comparison continues with the remaining suffixes. If one or both
// strings end at the digit boundary, the raw strings are compared
// lexically, which makes leading zeros significant as a tie-breaker
// (e.g. "a01" < "a1", "a0" < "a00").
func Compare(a, b string) int {
ia, ib := 0, 0
for ia < len(a) && ib < len(b) {
ca, cb := a[ia], b[ib]
da, db := isDigit(ca), isDigit(cb)
switch {
case da && db:
// Both are in digit sequences — compare numerically.
endA := ia
for endA < len(a) && isDigit(a[endA]) {
endA++
}
endB := ib
for endB < len(b) && isDigit(b[endB]) {
endB++
}
if c := compareNumbers(a[ia:endA], b[ib:endB]); c != 0 {
return c
}
// Numerically equal. If both sides have trailing data, continue
// comparing after the digit runs. Otherwise fall through to
// lexical comparison of the full remaining strings (which makes
// leading-zero differences significant as a tie-breaker).
if endA < len(a) && endB < len(b) {
ia = endA
ib = endB
continue
}
return strings.Compare(a[ia:], b[ib:])
case da != db:
return int(ca) - int(cb)
default:
if ca != cb {
return int(ca) - int(cb)
}
ia++
ib++
}
}
return (len(a) - ia) - (len(b) - ib)
}
// compareNumbers compares two digit strings numerically.
// Leading zeros are stripped before comparison.
func compareNumbers(a, b string) int {
// Strip leading zeros.
sa := stripZeros(a)
sb := stripZeros(b)
// Different lengths after stripping means different magnitude.
if len(sa) != len(sb) {
return len(sa) - len(sb)
}
// Same length — compare digit by digit.
for i := range len(sa) {
if sa[i] != sb[i] {
return int(sa[i]) - int(sb[i])
}
}
return 0
}
// stripZeros returns s with leading '0' bytes removed.
// If s is all zeros, returns the last byte (a single "0").
func stripZeros(s string) string {
i := 0
for i < len(s) && s[i] == '0' {
i++
}
if i == len(s) && len(s) > 0 {
return s[len(s)-1:]
}
return s[i:]
}
func isDigit(c byte) bool {
return c >= '0' && c <= '9'
}
+116
View File
@@ -0,0 +1,116 @@
package natural_test
import (
"testing"
"github.com/navidrome/navidrome/utils/natural"
. "github.com/onsi/ginkgo/v2"
. "github.com/onsi/gomega"
)
func TestNatural(t *testing.T) {
RegisterFailHandler(Fail)
RunSpecs(t, "Natural Suite")
}
var _ = Describe("Compare", func() {
DescribeTable("returns correct ordering",
func(a, b string, expected int) {
result := natural.Compare(a, b)
if expected < 0 {
Expect(result).To(BeNumerically("<", 0), "expected %q < %q", a, b)
} else if expected > 0 {
Expect(result).To(BeNumerically(">", 0), "expected %q > %q", a, b)
} else {
Expect(result).To(Equal(0), "expected %q == %q", a, b)
}
},
// Basic string ordering
Entry("a < b", "a", "b", -1),
Entry("b > a", "b", "a", 1),
Entry("a < aa (prefix)", "a", "aa", -1),
Entry("aa > a", "aa", "a", 1),
// Equal strings
Entry("equal strings return 0", "abc", "abc", 0),
Entry("both empty", "", "", 0),
Entry("a01 == a01", "a01", "a01", 0),
Entry("a1 == a1", "a1", "a1", 0),
// Empty string edge cases
Entry("empty < non-empty", "", "a", -1),
Entry("non-empty > empty", "a", "", 1),
// Numeric comparison
Entry("2 < 10 numerically", "2", "10", -1),
Entry("10 > 2 numerically", "10", "2", 1),
Entry("equal numbers", "42", "42", 0),
Entry("9 < 10", "9", "10", -1),
Entry("99 < 100", "99", "100", -1),
// Simple numeric segments (from original library)
Entry("a0 < a1", "a0", "a1", -1),
Entry("a0 < a00", "a0", "a00", -1),
Entry("a00 < a01", "a00", "a01", -1),
Entry("a01 < a1", "a01", "a1", -1),
Entry("a01 < a2", "a01", "a2", -1),
Entry("a01x < a2x", "a01x", "a2x", -1),
Entry("a01 > a00", "a01", "a00", 1),
Entry("a2 > a01", "a2", "a01", 1),
Entry("a2x > a01x", "a2x", "a01x", 1),
// Multiple numeric groups (from original library)
Entry("a0b00 < a00b1", "a0b00", "a00b1", -1),
Entry("a0b00 < a00b01", "a0b00", "a00b01", -1),
Entry("a00b0 < a0b00", "a00b0", "a0b00", -1),
Entry("a00b00 < a0b01", "a00b00", "a0b01", -1),
Entry("a00b00 < a0b1", "a00b00", "a0b1", -1),
Entry("a00b00 > a0b0", "a00b00", "a0b0", 1),
Entry("a00b01 > a0b00", "a00b01", "a0b00", 1),
Entry("a00b00 == a0b00", "a00b00", "a0b00", 0),
// Leading zeros at end of string — lexical tie-break
Entry("file01 < file1", "file01", "file1", -1),
// Prefix comparison
Entry("abc < abcd", "abc", "abcd", -1),
Entry("abcd > abc", "abcd", "abc", 1),
// Navidrome use cases: cover art sorting
Entry("cover < cover.1", "cover", "cover.1", -1),
Entry("cover.1 < cover.2", "cover.1", "cover.2", -1),
Entry("cover.2 < cover.10", "cover.2", "cover.10", -1),
// Navidrome use cases: disc sorting
Entry("disc1 < disc2", "disc1", "disc2", -1),
Entry("disc2 < disc10", "disc2", "disc10", -1),
Entry("disc1 < disc10", "disc1", "disc10", -1),
// Multiple numeric segments
Entry("a1b2 < a1b10", "a1b2", "a1b10", -1),
Entry("a2b1 > a1b2", "a2b1", "a1b2", 1),
// Numbers at the start
Entry("2abc < 10abc", "2abc", "10abc", -1),
// Numbers larger than uint64 max (from original library)
Entry("large: fewer digits < more digits",
"a99999999999999999999", "a100000000000000000000", -1),
Entry("large: digit-by-digit comparison",
"a123456789012345678901234567890", "a123456789012345678901234567891", -1),
Entry("large: more digits > fewer digits",
"a999999999999999999999", "a1000000000000000000000", -1),
Entry("large: 20 digits < 100 digits by length",
"a20000000000000000000", "a100000000000000000000", -1),
Entry("large: 100 digits > 20 digits",
"a100000000000000000000", "a20000000000000000000", 1),
Entry("large: reverse of above",
"a1000000000000000000000", "a999999999999999999999", 1),
Entry("large: equal",
"a100000000000000000000", "a100000000000000000000", 0),
Entry("large: leading zeros with trailing data",
"a00000000000000000000001x", "a1x", 0),
Entry("large: leading zeros with trailing data (2)",
"a099999999999999999999x", "a99999999999999999999x", 0),
)
})