package normalize

import "testing"

func TestNormalizeString_Basic(t *testing.T) {
	tests := []struct {
		input    string
		expected string
	}{
		// Lowercase conversion
		{"DARK SIDE OF THE MOON", "dark side of the moon"},
		// Special character removal
		{"Dark Side of the Moon!", "dark side of the moon"},
		{"Dark-Side-of-the-Moon", "dark side of the moon"},
		{"Dark_Side_of_the_Moon", "dark side of the moon"},
		// Bracket removal
		{"Dark Side of the Moon [Deluxe Edition]", "dark side of the moon"},
		{"Dark Side of the Moon [Remastered 2020]", "dark side of the moon"},
		{"Album [2023 Remix]", "album"},
		// Parenthesis removal
		{"Dark Side of the Moon (Deluxe)", "dark side of the moon"},
		{"Album (Remastered)", "album"},
		// Year removal
		{"Dark Side of the Moon 1973", "dark side of the moon"},
		// Standalone reissue keywords (no brackets/parens) are stripped
		{"Album 2020 Remastered", "album"},
		{"The Wall 2011 Remaster", "the wall"},
		{"X Deluxe", "x"},
		{"Y Expanded Edition", "y"},
		{"Z Remix", "z"},
		// Space collapsing
		{"Dark   Side   of   the   Moon", "dark side of the moon"},
		// Trim
		{"  Dark Side of the Moon  ", "dark side of the moon"},
		// Combined
		{"The Dark Side of the Moon [2011 Remaster] (Deluxe Edition)", "the dark side of the moon"},
		// Empty
		{"", ""},
		// Only special chars
		{"!@#$%^&*()", ""},
		// Digits that are not years should stay
		{"30 Seconds to Mars", "30 seconds to mars"},
		{"1941 - The Greatest Hits", "the greatest hits"},
		// Year-only title is preserved (not collapsed to empty) so it can still match
		{"1989", "1989"},
		{"2112", "2112"},
		// A year-plus-suffix title collapses to empty: it is a distinct release
		// group (e.g. "1989 (Deluxe)") and must NOT match a bare "1989".
		{"1989 (Deluxe)", ""},
		{"1989 [Deluxe Edition]", ""},
		{"2112 (Remastered)", ""},
		// Regression: a year with a bracketed/suffixed year must NOT collapse to
		// the bare year (it falsely matched "1989" before). It collapses to empty.
		{"1989 [2020]", ""},
		{"1989 2020", ""},
		// Both tokens are years → both stripped → empty (no album words remain).
		{"3000 2000", ""},
		{"1989 RMX", "rmx"},
		// Regression: year regex must cover ALL 4-digit years, not just 1000-2999.
		// A reissue of a year-titled album outside that range must still collapse
		// to empty so it is correctly reported as missing and does NOT falsely
		// match a bare year-titled local album.
		{"3000", "3000"},
		{"3000 (Remastered)", ""},
		{"3010 [Deluxe Edition]", ""},
		{"4000 (Remastered)", ""},
	}

	for _, tt := range tests {
		t.Run(tt.input, func(t *testing.T) {
			got := NormalizeString(tt.input)
			if got != tt.expected {
				t.Errorf("NormalizeString(%q) = %q, want %q", tt.input, got, tt.expected)
			}
		})
	}
}

func TestNormalizeArtistName(t *testing.T) {
	tests := []struct {
		input    string
		expected string
	}{
		{"Pink Floyd", "pink floyd"},
		{"The Beatles", "beatles"},
		{"A Perfect Circle", "perfect circle"},
		{"An Orchestra", "orchestra"},
		{"  The  Who  ", "who"},
		{"THE WHO", "who"},
		// No stripping needed
		{"Radiohead", "radiohead"},
		// Already stripped
		{"Beatles", "beatles"},
	}

	for _, tt := range tests {
		t.Run(tt.input, func(t *testing.T) {
			got := NormalizeArtistName(tt.input)
			if got != tt.expected {
				t.Errorf("NormalizeArtistName(%q) = %q, want %q", tt.input, got, tt.expected)
			}
		})
	}
}