navidrome/scanner/tag_scanner.go

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

400 lines
11 KiB
Go
Raw Normal View History

2020-01-16 22:53:48 +01:00
package scanner
import (
"context"
"os"
"path/filepath"
"sort"
2020-01-16 22:53:48 +01:00
"strings"
"time"
2020-08-03 05:17:13 +02:00
"github.com/deluan/navidrome/conf"
2020-01-24 01:44:08 +01:00
"github.com/deluan/navidrome/log"
"github.com/deluan/navidrome/model"
2020-07-27 17:12:39 +02:00
"github.com/deluan/navidrome/model/request"
"github.com/deluan/navidrome/scanner/metadata"
"github.com/deluan/navidrome/utils"
2020-01-16 22:53:48 +01:00
)
type TagScanner struct {
rootFolder string
ds model.DataStore
mapper *mediaFileMapper
2020-07-27 17:12:39 +02:00
plsSync *playlistSync
cnt *counters
2020-01-16 22:53:48 +01:00
}
func NewTagScanner(rootFolder string, ds model.DataStore) *TagScanner {
2020-01-16 22:53:48 +01:00
return &TagScanner{
rootFolder: rootFolder,
mapper: newMediaFileMapper(rootFolder),
2020-07-27 17:12:39 +02:00
plsSync: newPlaylistSync(ds),
ds: ds,
2020-01-16 22:53:48 +01:00
}
}
2020-06-11 23:36:09 +02:00
type (
2020-07-12 18:35:23 +02:00
counters struct {
added int64
updated int64
deleted int64
}
2020-06-11 23:36:09 +02:00
)
const (
2020-07-27 17:12:39 +02:00
// filesBatchSize used for batching file metadata extraction
2020-06-11 23:36:09 +02:00
filesBatchSize = 100
)
2020-07-27 17:12:39 +02:00
// Scanner algorithm overview:
// Load all directories under the music folder, with their ModTime (self or any non-dir children, whichever is newer)
// Load all directories from the DB
// Compare both collections to find changed folders (based on lastModifiedSince) and deleted folders
// For each deleted folder: delete all files from DB whose path starts with the delete folder path (non-recursively)
// For each changed folder: get all files from DB whose path starts with the changed folder (non-recursively), check each file:
// if file in folder is newer, update the one in DB
2020-07-27 17:12:39 +02:00
// if file in folder does not exists in DB, add it
// for each file in the DB that is not found in the folder, delete it from DB
// Create new albums/artists, update counters:
// collect all albumIDs and artistIDs from previous steps
// refresh the collected albums and artists with the metadata from the mediafiles
2020-07-27 17:12:39 +02:00
// For each changed folder, process playlists:
// If the playlist is not in the DB, import it, setting sync = true
// If the playlist is in the DB and sync == true, import it, or else skip it
// Delete all empty albums, delete all empty artists, clean-up playlists
2020-01-16 22:53:48 +01:00
func (s *TagScanner) Scan(ctx context.Context, lastModifiedSince time.Time) error {
2020-07-27 17:12:39 +02:00
ctx = s.withAdminUser(ctx)
start := time.Now()
2020-07-27 17:12:39 +02:00
allFSDirs, err := s.getDirTree(ctx)
if err != nil {
return err
}
2020-07-12 19:30:03 +02:00
2020-07-27 17:12:39 +02:00
allDBDirs, err := s.getDBDirTree(ctx)
2020-01-16 22:53:48 +01:00
if err != nil {
return err
}
2020-07-27 17:12:39 +02:00
changedDirs := s.getChangedDirs(ctx, allFSDirs, allDBDirs, lastModifiedSince)
deletedDirs := s.getDeletedDirs(ctx, allFSDirs, allDBDirs)
if len(changedDirs)+len(deletedDirs) == 0 {
log.Debug(ctx, "No changes found in Music Folder", "folder", s.rootFolder)
2020-01-16 22:53:48 +01:00
return nil
}
if log.CurrentLevel() >= log.LevelTrace {
2020-07-27 17:12:39 +02:00
log.Info(ctx, "Folder changes detected", "changedFolders", len(changedDirs), "deletedFolders", len(deletedDirs),
"changed", strings.Join(changedDirs, ";"), "deleted", strings.Join(deletedDirs, ";"))
} else {
2020-07-27 17:12:39 +02:00
log.Info(ctx, "Folder changes detected", "changedFolders", len(changedDirs), "deletedFolders", len(deletedDirs))
}
2020-01-16 22:53:48 +01:00
2020-07-27 17:12:39 +02:00
s.cnt = &counters{}
2020-01-16 22:53:48 +01:00
2020-07-27 17:12:39 +02:00
for _, dir := range deletedDirs {
err := s.processDeletedDir(ctx, dir)
2020-01-16 22:53:48 +01:00
if err != nil {
2020-07-27 17:12:39 +02:00
log.Error("Error removing deleted folder from DB", "path", dir, err)
2020-01-16 22:53:48 +01:00
}
}
2020-07-27 17:12:39 +02:00
for _, dir := range changedDirs {
err := s.processChangedDir(ctx, dir)
2020-01-16 22:53:48 +01:00
if err != nil {
2020-07-27 17:12:39 +02:00
log.Error("Error updating folder in the DB", "path", dir, err)
2020-01-16 22:53:48 +01:00
}
}
2020-07-27 17:12:39 +02:00
plsCount := 0
2020-08-03 05:17:13 +02:00
if conf.Server.AutoImportPlaylists {
// Now that all mediafiles are imported/updated, search for and import playlists
u, _ := request.UserFrom(ctx)
for _, dir := range changedDirs {
info := allFSDirs[dir]
if info.hasPlaylist {
if !u.IsAdmin {
log.Warn("Playlists will not be imported, as there are no admin users yet, "+
"Please create an admin user first, and then update the playlists for them to be imported", "dir", dir)
} else {
plsCount = s.plsSync.processPlaylists(ctx, dir)
}
2020-07-27 17:12:39 +02:00
}
}
2020-08-03 05:17:13 +02:00
} else {
log.Debug("Playlist auto-import is disabled")
2020-01-16 22:53:48 +01:00
}
2020-07-27 17:12:39 +02:00
err = s.ds.GC(log.NewContext(ctx))
log.Info("Finished processing Music Folder", "folder", s.rootFolder, "elapsed", time.Since(start),
"added", s.cnt.added, "updated", s.cnt.updated, "deleted", s.cnt.deleted, "playlistsImported", plsCount)
return err
}
func (s *TagScanner) getDirTree(ctx context.Context) (dirMap, error) {
start := time.Now()
log.Trace(ctx, "Loading directory tree from music folder", "folder", s.rootFolder)
dirs, err := loadDirTree(ctx, s.rootFolder)
2020-01-16 22:53:48 +01:00
if err != nil {
2020-07-27 17:12:39 +02:00
return nil, err
2020-01-16 22:53:48 +01:00
}
2020-07-27 17:12:39 +02:00
log.Debug("Directory tree loaded from music folder", "total", len(dirs), "elapsed", time.Since(start))
return dirs, nil
}
2020-01-16 22:53:48 +01:00
2020-07-27 17:12:39 +02:00
func (s *TagScanner) getDBDirTree(ctx context.Context) (map[string]struct{}, error) {
start := time.Now()
log.Trace(ctx, "Loading directory tree from database", "folder", s.rootFolder)
2020-07-27 17:12:39 +02:00
repo := s.ds.MediaFile(ctx)
dirs, err := repo.FindPathsRecursively(s.rootFolder)
if err != nil {
return nil, err
}
resp := map[string]struct{}{}
for _, d := range dirs {
resp[filepath.Clean(d)] = struct{}{}
}
2020-01-18 05:28:11 +01:00
2020-07-27 17:12:39 +02:00
log.Debug("Directory tree loaded from DB", "total", len(resp), "elapsed", time.Since(start))
return resp, nil
2020-01-16 22:53:48 +01:00
}
2020-07-27 17:12:39 +02:00
func (s *TagScanner) getChangedDirs(ctx context.Context, fsDirs dirMap, dbDirs map[string]struct{}, lastModified time.Time) []string {
start := time.Now()
log.Trace(ctx, "Checking for changed folders")
var changed []string
for d, info := range fsDirs {
_, inDB := dbDirs[d]
if (!inDB && (info.hasAudioFiles)) || info.modTime.After(lastModified) {
changed = append(changed, d)
}
}
2020-07-27 17:12:39 +02:00
sort.Strings(changed)
log.Debug(ctx, "Finished changed folders check", "total", len(changed), "elapsed", time.Since(start))
return changed
}
func (s *TagScanner) getDeletedDirs(ctx context.Context, fsDirs dirMap, dbDirs map[string]struct{}) []string {
start := time.Now()
log.Trace(ctx, "Checking for deleted folders")
var deleted []string
for d := range dbDirs {
if _, ok := fsDirs[d]; !ok {
deleted = append(deleted, d)
}
2020-01-16 22:53:48 +01:00
}
2020-07-27 17:12:39 +02:00
sort.Strings(deleted)
log.Debug(ctx, "Finished deleted folders check", "total", len(deleted), "elapsed", time.Since(start))
return deleted
2020-01-16 22:53:48 +01:00
}
2020-07-27 17:12:39 +02:00
func (s *TagScanner) processDeletedDir(ctx context.Context, dir string) error {
start := time.Now()
buffer := newRefreshBuffer(ctx, s.ds)
mfs, err := s.ds.MediaFile(ctx).FindAllByPath(dir)
if err != nil {
return err
}
2020-07-27 17:12:39 +02:00
c, err := s.ds.MediaFile(ctx).DeleteByPath(dir)
if err != nil {
return err
2020-01-16 22:53:48 +01:00
}
2020-07-27 17:12:39 +02:00
s.cnt.deleted += c
for _, t := range mfs {
buffer.accumulate(t)
}
err = buffer.flush()
log.Info(ctx, "Finished processing deleted folder", "path", dir, "purged", len(mfs), "elapsed", time.Since(start))
return err
2020-01-16 22:53:48 +01:00
}
2020-07-27 17:12:39 +02:00
func (s *TagScanner) processChangedDir(ctx context.Context, dir string) error {
2020-01-16 22:53:48 +01:00
start := time.Now()
2020-07-27 17:12:39 +02:00
buffer := newRefreshBuffer(ctx, s.ds)
2020-01-16 22:53:48 +01:00
// Load folder's current tracks from DB into a map
currentTracks := map[string]model.MediaFile{}
ct, err := s.ds.MediaFile(ctx).FindAllByPath(dir)
2020-01-16 22:53:48 +01:00
if err != nil {
return err
}
for _, t := range ct {
currentTracks[t.Path] = t
2020-01-16 22:53:48 +01:00
}
// Load tracks FileInfo from the folder
2020-07-27 17:12:39 +02:00
files, err := loadAllAudioFiles(dir)
2020-01-16 22:53:48 +01:00
if err != nil {
return err
}
// If no files to process, return
if len(files)+len(currentTracks) == 0 {
return nil
}
2020-07-27 17:12:39 +02:00
orphanTracks := map[string]model.MediaFile{}
for k, v := range currentTracks {
orphanTracks[k] = v
}
// If track from folder is newer than the one in DB, select for update/insert in DB
log.Trace(ctx, "Processing changed folder", "dir", dir, "tracksInDB", len(currentTracks), "tracksInFolder", len(files))
var filesToUpdate []string
for filePath, info := range files {
c, ok := currentTracks[filePath]
2020-07-12 18:35:23 +02:00
if !ok {
filesToUpdate = append(filesToUpdate, filePath)
2020-07-27 17:12:39 +02:00
s.cnt.added++
2020-07-12 18:35:23 +02:00
}
if ok && info.ModTime().After(c.UpdatedAt) {
filesToUpdate = append(filesToUpdate, filePath)
2020-07-27 17:12:39 +02:00
s.cnt.updated++
}
2020-07-27 17:12:39 +02:00
// Force a refresh of the album and artist, to cater for cover art files
buffer.accumulate(c)
// Only leaves in orphanTracks the ones not found in the folder. After this loop any remaining orphanTracks
// are considered gone from the music folder and will be deleted from DB
delete(orphanTracks, filePath)
}
2020-01-16 22:53:48 +01:00
numUpdatedTracks := 0
numPurgedTracks := 0
if len(filesToUpdate) > 0 {
2020-07-27 17:12:39 +02:00
numUpdatedTracks, err = s.addOrUpdateTracksInDB(ctx, dir, currentTracks, filesToUpdate, buffer)
if err != nil {
return err
}
}
2020-07-27 17:12:39 +02:00
if len(orphanTracks) > 0 {
numPurgedTracks, err = s.deleteOrphanSongs(ctx, dir, orphanTracks, buffer)
if err != nil {
return err
}
}
2020-07-27 17:12:39 +02:00
err = buffer.flush()
log.Info(ctx, "Finished processing changed folder", "dir", dir, "updated", numUpdatedTracks,
"purged", numPurgedTracks, "elapsed", time.Since(start))
return err
}
2020-07-27 17:12:39 +02:00
func (s *TagScanner) deleteOrphanSongs(ctx context.Context, dir string, tracksToDelete map[string]model.MediaFile, buffer *refreshBuffer) (int, error) {
numPurgedTracks := 0
2020-01-16 22:53:48 +01:00
2020-07-27 17:12:39 +02:00
log.Debug(ctx, "Deleting orphan tracks from DB", "dir", dir, "numTracks", len(tracksToDelete))
// Remaining tracks from DB that are not in the folder are deleted
for _, ct := range tracksToDelete {
numPurgedTracks++
buffer.accumulate(ct)
if err := s.ds.MediaFile(ctx).Delete(ct.ID); err != nil {
return 0, err
}
2020-07-27 17:12:39 +02:00
s.cnt.deleted++
2020-01-16 22:53:48 +01:00
}
2020-07-27 17:12:39 +02:00
return numPurgedTracks, nil
2020-01-16 22:53:48 +01:00
}
2020-07-27 17:12:39 +02:00
func (s *TagScanner) addOrUpdateTracksInDB(ctx context.Context, dir string, currentTracks map[string]model.MediaFile, filesToUpdate []string, buffer *refreshBuffer) (int, error) {
numUpdatedTracks := 0
log.Trace(ctx, "Updating mediaFiles in DB", "dir", dir, "numFiles", len(filesToUpdate))
// Break the file list in chunks to avoid calling ffmpeg with too many parameters
chunks := utils.BreakUpStringSlice(filesToUpdate, filesBatchSize)
for _, chunk := range chunks {
// Load tracks Metadata from the folder
newTracks, err := s.loadTracks(chunk)
if err != nil {
2020-07-27 17:12:39 +02:00
return 0, err
}
2020-07-27 17:12:39 +02:00
// If track from folder is newer than the one in DB, update/insert in DB
log.Trace(ctx, "Updating mediaFiles in DB", "dir", dir, "files", chunk, "numFiles", len(chunk))
for i := range newTracks {
n := newTracks[i]
// Keep current annotations if the track is in the DB
if t, ok := currentTracks[n.Path]; ok {
n.Annotations = t.Annotations
}
2020-07-27 17:12:39 +02:00
err := s.ds.MediaFile(ctx).Put(&n)
if err != nil {
return 0, err
}
buffer.accumulate(n)
numUpdatedTracks++
}
}
2020-07-27 17:12:39 +02:00
return numUpdatedTracks, nil
}
2020-01-16 22:53:48 +01:00
func (s *TagScanner) loadTracks(filePaths []string) (model.MediaFiles, error) {
mds, err := metadata.Extract(filePaths...)
if err != nil {
return nil, err
}
var mfs model.MediaFiles
for _, md := range mds {
mf := s.mapper.toMediaFile(md)
mfs = append(mfs, mf)
2020-01-16 22:53:48 +01:00
}
return mfs, nil
2020-01-16 22:53:48 +01:00
}
2020-07-27 17:12:39 +02:00
func (s *TagScanner) withAdminUser(ctx context.Context) context.Context {
u, err := s.ds.User(ctx).FindFirstAdmin()
if err != nil {
log.Warn(ctx, "No admin user found!", err)
u = &model.User{}
}
ctx = request.WithUsername(ctx, u.UserName)
return request.WithUser(ctx, *u)
}
func loadAllAudioFiles(dirPath string) (map[string]os.FileInfo, error) {
dir, err := os.Open(dirPath)
if err != nil {
return nil, err
}
files, err := dir.Readdir(-1)
if err != nil {
return nil, err
}
audioFiles := make(map[string]os.FileInfo)
for _, f := range files {
if f.IsDir() {
continue
}
if strings.HasPrefix(f.Name(), ".") {
continue
}
filePath := filepath.Join(dirPath, f.Name())
if !utils.IsAudioFile(filePath) {
continue
}
fi, err := os.Stat(filePath)
if err != nil {
log.Error("Could not stat file", "filePath", filePath, err)
} else {
audioFiles[filePath] = fi
}
}
return audioFiles, nil
}