fix: address code review findings
This commit is contained in:
@@ -18,6 +18,10 @@ var (
|
||||
parenRe = regexp.MustCompile(`\([^)]*\)`)
|
||||
yearRe = regexp.MustCompile(`\b(1[0-9]{3}|2[0-9]{3})\b`)
|
||||
spaceRe = regexp.MustCompile(`\s+`)
|
||||
// wordRe matches any alphabetic character. Used to decide whether a title
|
||||
// that collapses entirely to a year actually had other words worth keeping
|
||||
// (e.g. "1989 (Deluxe)") versus being a bare year title (e.g. "1989").
|
||||
wordRe = regexp.MustCompile(`[a-z]`)
|
||||
)
|
||||
|
||||
// NormalizeString normalizes a string for fuzzy matching by:
|
||||
@@ -28,6 +32,10 @@ var (
|
||||
// - Collapsing multiple spaces into one
|
||||
// - Trimming leading/trailing whitespace
|
||||
func NormalizeString(s string) string {
|
||||
// Capture the original input; used after stripping to tell a bare year
|
||||
// title apart from a title that merely collapses to a year.
|
||||
original := s
|
||||
|
||||
// Convert to lowercase
|
||||
s = strings.ToLower(s)
|
||||
|
||||
@@ -37,11 +45,23 @@ func NormalizeString(s string) string {
|
||||
// Remove parenthesized content (e.g., (Deluxe), (Remastered))
|
||||
s = parenRe.ReplaceAllString(s, "")
|
||||
|
||||
// Remove years (4-digit numbers between 1000-2999). If stripping the
|
||||
// year would empty the entire string (e.g. an album literally titled
|
||||
// "1989" or "2112"), keep the original form so the title can still match.
|
||||
// Remove years (4-digit numbers between 1000-2999). If stripping the year
|
||||
// would empty the entire string, we must decide what to keep:
|
||||
// - A bare year title (e.g. "1989", "2112") has no other words, so keep
|
||||
// the year so it can still match itself (the user owns that album).
|
||||
// - A title that had OTHER words alongside the year (e.g. "1989 (Deluxe)")
|
||||
// collapses to empty on purpose: it is a distinct release group that
|
||||
// must NOT be considered already-present just because the user owns the
|
||||
// standard "1989". Collapsing to empty makes it score 0.0 against a
|
||||
// plain "1989", correctly reporting the reissue as missing.
|
||||
stripped := yearRe.ReplaceAllString(s, "")
|
||||
if strings.TrimSpace(stripped) != "" {
|
||||
if strings.TrimSpace(stripped) == "" {
|
||||
if wordRe.MatchString(strings.ToLower(original)) {
|
||||
s = ""
|
||||
} else {
|
||||
s = strings.TrimSpace(s)
|
||||
}
|
||||
} else {
|
||||
s = stripped
|
||||
}
|
||||
|
||||
|
||||
Reference in New Issue
Block a user