playlink/internal/engine/layout.go

package engine

import (
	"fmt"
	"os"
	"path/filepath"
	"sort"
	"strings"

	"github.com/cdnutter/playlink/internal/store"
)

// Layout warnings flag real layout problems only. Content that exists in just
// one library is normal (a singles library naturally has artists and songs
// the albums library lacks), so it is never counted against the layout.

var accentFold = strings.NewReplacer(
	"á", "a", "à", "a", "â", "a", "ä", "a", "ã", "a", "å", "a", "é", "e", "è", "e", "ê", "e", "ë", "e",
	"í", "i", "ì", "i", "î", "i", "ï", "i", "ó", "o", "ò", "o", "ô", "o", "ö", "o", "õ", "o", "ø", "o",
	"ú", "u", "ù", "u", "û", "u", "ü", "u", "ñ", "n", "ç", "c", "ý", "y", "ÿ", "y", "œ", "oe", "æ", "ae", "ß", "ss",
)

// folderKey identifies a folder name regardless of case, punctuation, accents,
// "&" vs "and" and a leading/trailing "The": "AC/DC", "AC-DC" and "ACDC", or
// "The Beatles" and "Beatles, The", get the same key.
func folderKey(name string) string {
	s := accentFold.Replace(strings.ToLower(strings.TrimSpace(name)))
	s = strings.ReplaceAll(s, "&", " and ")
	s = strings.TrimPrefix(s, "the ")
	s = strings.TrimSuffix(s, ", the")
	var b strings.Builder
	for _, w := range words(s) {
		b.WriteString(w)
	}
	return b.String()
}

// pathKey applies folderKey to every component of a relative folder path.
func pathKey(dir string) string {
	parts := strings.Split(dir, string(filepath.Separator))
	for i, p := range parts {
		parts[i] = folderKey(p)
	}
	return strings.Join(parts, "/")
}

func subdirs(dir string) []string {
	ents, _ := os.ReadDir(dir)
	var out []string
	for _, en := range ents {
		if en.IsDir() && !strings.HasPrefix(en.Name(), ".") {
			out = append(out, en.Name())
		}
	}
	return out
}

// renamedFolders finds slave folders that are the same folder as a master
// folder under a different name. It checks top-level folders and, for
// top-level folders both libraries share, the folders inside them.
func renamedFolders(l *libs) [][2]string {
	var out [][2]string
	compare := func(prefix string, master, slave []string) {
		exact := map[string]bool{}
		byKey := map[string]string{}
		for _, m := range master {
			exact[m] = true
			if k := folderKey(m); k != "" {
				byKey[k] = m
			}
		}
		for _, s := range slave {
			if exact[s] {
				continue
			}
			if m, ok := byKey[folderKey(s)]; ok {
				out = append(out, [2]string{filepath.Join(prefix, s), filepath.Join(prefix, m)})
			}
		}
	}
	mTop, sTop := subdirs(l.master.Path), subdirs(l.slave.Path)
	compare("", mTop, sTop)
	shared := map[string]bool{}
	for _, m := range mTop {
		shared[m] = true
	}
	for _, s := range sTop {
		if shared[s] {
			compare(s, subdirs(filepath.Join(l.master.Path, s)), subdirs(filepath.Join(l.slave.Path, s)))
		}
	}
	sort.Slice(out, func(i, j int) bool { return out[i][0] < out[j][0] })
	return out
}

// maxWarningItems caps how many rows a warning carries.
const maxWarningItems = 200

func renamedWarning(pairs [][2]string) *store.Warning {
	if len(pairs) == 0 {
		return nil
	}
	w := &store.Warning{Total: len(pairs),
		Detail: "Songs inside these folders can only match by path once the names agree. Rename one side so they match."}
	if len(pairs) == 1 {
		w.Title = "1 folder is named differently in the two libraries"
	} else {
		w.Title = fmt.Sprintf("%d folders are named differently in the two libraries", len(pairs))
	}
	for _, p := range pairs[:min(maxWarningItems, len(pairs))] {
		w.Items = append(w.Items, store.WarningItem{Slave: p[0], Master: p[1]})
	}
	return w
}

// layoutStats watches, during a scan, where songs found in both libraries
// live. Songs only in the slave library are ignored.
type layoutStats struct {
	shared int               // slave files with a master counterpart
	moved  int               // ...that live in a different folder in the master
	pairs  map[[2]string]int // (slave album folder, master album folder) -> songs
}

func (ls *layoutStats) add(relSlave, relMaster, kind, reason string) {
	if kind == store.KindNone || relMaster == "" {
		return
	}
	ls.shared++
	if kind != store.KindElsewhere && kind != store.KindApprox {
		return
	}
	if albumKey(relSlave) == albumKey(relMaster) {
		return // only a disc folder differs
	}
	if pathKey(AlbumFolder(relSlave)) == pathKey(AlbumFolder(relMaster)) {
		return // same folder under another name: the renamed-folders warning covers it
	}
	// A single, EP or compilation matched to its album version is the
	// intended behaviour, not a layout problem.
	if st := parseTitle(relSlave); st.single || st.compilation || strings.Contains(reason, "album version preferred") {
		return
	}
	ls.moved++
	if ls.pairs == nil {
		ls.pairs = map[[2]string]int{}
	}
	ls.pairs[[2]string{AlbumFolder(relSlave), AlbumFolder(relMaster)}]++
}

func (ls *layoutStats) warning() *store.Warning {
	if ls.shared < 20 || ls.moved*4 <= ls.shared {
		return nil
	}
	items := make([]store.WarningItem, 0, len(ls.pairs))
	for p, n := range ls.pairs {
		items = append(items, store.WarningItem{Slave: p[0], Master: p[1], Count: n})
	}
	sort.Slice(items, func(i, j int) bool {
		if items[i].Count != items[j].Count {
			return items[i].Count > items[j].Count
		}
		return items[i].Slave < items[j].Slave
	})
	w := &store.Warning{
		Title: fmt.Sprintf("%d of %d songs found in both libraries are in a different folder in the master library", ls.moved, ls.shared),
		Detail: "The libraries seem to use different folder layouts here, so these songs only match through the fallbacks. " +
			"Folders are listed with the most affected songs first.",
		Total: len(items),
		Items: items[:min(maxWarningItems, len(items))],
	}
	return w
}