navidrome/scanner/controller.go
Deluan Quintão 758e64c999
feat(scanner): per-library PID configuration (#6252)
* feat(model): add per-library PID config columns

* refactor(metadata): pass PID config to ToMediaFile and add spec validation

* feat(scanner): rescan only libraries whose PID config changed

* feat(server): validate library PID config and rescan on change

* feat(ui): edit per-library PID config

* fix(ui): label the PID mode selects

* fix: tighten per-library PID rescan edge cases

An interrupted PID rescan no longer upgrades every library to a full scan, a save that loses the race for the scanner logs at debug, the confirm dialog only shows when the effective PID spec changes, and it now gets translation keys.

* refactor(metadata): pass the library to ToMediaFile

ToMediaFile and core.Inspect took the library ID and its PID config as
separate arguments, so a caller could mix values from two libraries. They
now take the model.Library and resolve the effective PID config from it.

* chore: tidy per-library PID comments, PropTypes and migration

Trim comments that restated the code, add PropTypes to the new UI
components, and recreate the migration with make migration-sql.

* fix(ui): show the PID spec help under its input

* feat(cmd): make inspect use the file's library PID config

inspect always used the global PID config, so it showed different IDs than
the scanner for files in a library with an override. It now finds the
file's library in the DB and uses its effective config, falling back to
the global config when there is no DB or the file is outside every
library. It never creates a DB. The library path matcher moves from
core/playlists to model so both can use it.

* refactor: simplify per-library PID code

Share the DB-file check between CLI commands, move ErrAlreadyScanning to
model so core no longer imports scanner, read the libraries once for
insights, and let ValidatePIDSpec accept an empty spec and look tags up
directly. In the scanner, use FullScanInProgress instead of a second
flag, and skip recomputing album IDs when the album spec did not change.
In the UI, share the PID inputs between Create and Edit, and use docsUrl.

* feat(ui): add section titles to Library Create and pre-fill Custom PID specs

Custom now starts from the global spec, so admins edit a working spec
instead of typing one from scratch.

* fix(inspect): map files with the library-relative path the scanner uses

Inspect gave metadata the file's directory as typed, so folder-based PIDs
never matched the DB. It now uses the path relative to the library root,
through the scanner's helper, which moves to model.

* fix(scanner): say when a PID rescan only covers target folders

* fix: reject tag aliases in album PID specs and match root libraries

Tags are stored under canonical names, so an alias in a spec always reads
as empty. In an album spec that gives every album the same ID, so album
specs now require the tag name. Track specs keep accepting aliases, since
the default one uses them. LibraryMatcher now matches paths under a
library at the filesystem root.

* refactor(model): move the tag alias lookup to tag_mappings.go

* test: run the library matcher and inspect tests on Windows

Build test paths with filepath instead of Unix literals, so they use the
OS separator like filepath.Abs output, and drop the Windows skips.

* feat(ui): add pt-BR translations for per-library PID settings
2026-10-02 05:05:32 -04:00

399 lines
12 KiB
Go

package scanner
import (
"context"
"errors"
"fmt"
"slices"
"sync"
"sync/atomic"
"time"
"github.com/navidrome/navidrome/conf"
"github.com/navidrome/navidrome/consts"
"github.com/navidrome/navidrome/core/auth"
"github.com/navidrome/navidrome/core/metrics"
"github.com/navidrome/navidrome/core/playlists"
"github.com/navidrome/navidrome/db"
"github.com/navidrome/navidrome/log"
"github.com/navidrome/navidrome/model"
"github.com/navidrome/navidrome/model/request"
"github.com/navidrome/navidrome/server/events"
"github.com/navidrome/navidrome/utils/pl"
"github.com/navidrome/navidrome/utils/singleton"
"golang.org/x/time/rate"
)
var (
ErrAlreadyScanning = model.ErrAlreadyScanning
)
func New(rootCtx context.Context, ds model.DataStore, broker events.Broker,
pls playlists.Playlists, m metrics.Metrics) model.Scanner {
c := &controller{
rootCtx: rootCtx,
ds: ds,
broker: broker,
pls: pls,
metrics: m,
devExternalScanner: conf.Server.DevExternalScanner,
}
if !c.devExternalScanner {
c.limiter = new(rate.Sometimes{Interval: conf.Server.DevActivityPanelUpdateRate})
}
return c
}
func (s *controller) getScanner() scanner {
if s.devExternalScanner {
return &scannerExternal{}
}
return &scannerImpl{ds: s.ds, pls: s.pls}
}
// CallScan starts an in-process scan of specific library/folder pairs.
// If targets is empty, it scans all libraries.
// This is meant to be called from the command line (see cmd/scan.go).
func CallScan(ctx context.Context, ds model.DataStore, pls playlists.Playlists, fullScan bool, targets []model.ScanTarget) (<-chan *ProgressInfo, error) {
release, err := lockScan(ctx)
if err != nil {
return nil, err
}
defer release()
ctx = auth.WithAdminUser(ctx, ds)
progress := make(chan *ProgressInfo, 100)
go func() {
defer close(progress)
scanner := &scannerImpl{ds: ds, pls: pls}
scanner.scanFolders(ctx, fullScan, targets, progress)
}()
return progress, nil
}
func IsScanning() bool {
return running.Load()
}
type ProgressInfo struct {
LibID int
FileCount uint32
Path string
Phase string
ChangesDetected bool
Warning string
Error string
ForceUpdate bool
}
// scanner defines the interface for different scanner implementations.
// This allows for swapping between in-process and external scanners.
type scanner interface {
// scanFolders performs the actual scanning of folders. If targets is nil, it scans all libraries.
scanFolders(ctx context.Context, fullScan bool, targets []model.ScanTarget, progress chan<- *ProgressInfo)
}
type controller struct {
rootCtx context.Context //nolint:containedctx // scanner lifecycle ctx
ds model.DataStore
broker events.Broker
metrics metrics.Metrics
pls playlists.Playlists
limiter *rate.Sometimes
devExternalScanner bool
count atomic.Uint32
folderCount atomic.Uint32
changesDetected bool
}
// getLastScanTime returns the most recent scan time across all libraries
func (s *controller) getLastScanTime(ctx context.Context) (time.Time, error) {
libs, err := s.ds.Library().GetAll(ctx, model.QueryOptions{
Sort: "last_scan_at",
Order: "desc",
Max: 1,
})
if err != nil {
return time.Time{}, fmt.Errorf("getting libraries: %w", err)
}
if len(libs) == 0 {
return time.Time{}, nil
}
return libs[0].LastScanAt, nil
}
// getScanInfo retrieves scan status from the database
func (s *controller) getScanInfo(ctx context.Context) (scanType string, elapsed time.Duration, lastErr string) {
lastErr, _ = s.ds.Property().DefaultGet(ctx, consts.LastScanErrorKey, "")
scanType, _ = s.ds.Property().DefaultGet(ctx, consts.LastScanTypeKey, "")
startTimeStr, _ := s.ds.Property().DefaultGet(ctx, consts.LastScanStartTimeKey, "")
if startTimeStr != "" {
startTime, err := time.Parse(time.RFC3339, startTimeStr)
if err == nil {
if running.Load() {
elapsed = time.Since(startTime)
} else {
// If scan is not running, calculate elapsed time using the most recent scan time
lastScanTime, err := s.getLastScanTime(ctx)
if err == nil && !lastScanTime.IsZero() {
elapsed = lastScanTime.Sub(startTime)
}
}
}
}
return scanType, elapsed, lastErr
}
func (s *controller) Status(ctx context.Context) (*model.ScannerStatus, error) {
lastScanTime, err := s.getLastScanTime(ctx)
if err != nil {
return nil, fmt.Errorf("getting last scan time: %w", err)
}
scanType, elapsed, lastErr := s.getScanInfo(ctx)
if running.Load() {
status := &model.ScannerStatus{
Scanning: true,
LastScan: lastScanTime,
Count: s.count.Load(),
FolderCount: s.folderCount.Load(),
LastError: lastErr,
ScanType: scanType,
ElapsedTime: elapsed,
}
return status, nil
}
count, folderCount, err := s.getCounters(ctx)
if err != nil {
return nil, fmt.Errorf("getting library stats: %w", err)
}
return &model.ScannerStatus{
Scanning: false,
LastScan: lastScanTime,
Count: uint32(count),
FolderCount: uint32(folderCount),
LastError: lastErr,
ScanType: scanType,
ElapsedTime: elapsed,
}, nil
}
func (s *controller) getCounters(ctx context.Context) (int64, int64, error) {
libs, err := s.ds.Library().GetAll(ctx)
if err != nil {
return 0, 0, fmt.Errorf("library count: %w", err)
}
var count, folderCount int64
for _, l := range libs {
count += int64(l.TotalSongs)
folderCount += int64(l.TotalFolders)
}
return count, folderCount, nil
}
func (s *controller) ScanAll(requestCtx context.Context, fullScan bool) ([]string, error) {
return s.ScanFolders(requestCtx, fullScan, nil)
}
func (s *controller) ScanFolders(requestCtx context.Context, fullScan bool, targets []model.ScanTarget) ([]string, error) {
release, err := lockScan(requestCtx)
if err != nil {
return nil, err
}
defer release()
// Prepare the context for the scan
ctx := request.AddValues(s.rootCtx, requestCtx)
ctx = auth.WithAdminUser(ctx, s.ds)
// A quick scan is promoted to a full one when it resumes an interrupted full scan; that happens
// inside the scanner (possibly in a subprocess), so mirror it here for the analysis gate. Must
// be read before the scan: ScanEnd clears the flag.
effectiveFullScan := EffectiveFullScan(ctx, s.ds, fullScan, targets)
if effectiveFullScan || s.includesUnscannedLibrary(ctx, targets) {
if err := db.MarkOptimizePending(ctx); err != nil {
log.Error(ctx, "Scanner: Error marking DB analysis pending", err)
}
}
// Send the initial scan status event
s.sendMessage(ctx, &events.ScanStatus{Scanning: true, Count: 0, FolderCount: 0})
progress := make(chan *ProgressInfo, 100)
go func() {
defer close(progress)
scanner := s.getScanner()
scanner.scanFolders(ctx, fullScan, targets, progress)
}()
// Wait for the scan to finish, sending progress events to all connected clients
scanWarnings, scanError := s.trackProgress(ctx, progress)
for _, w := range scanWarnings {
log.Warn(ctx, fmt.Sprintf("Scan warning: %s", w))
}
// Store scan error in database so it can be displayed in the UI
if scanError != nil {
_ = s.ds.Property().Put(ctx, consts.LastScanErrorKey, scanError.Error())
}
// Refresh the query-planner statistics after a successful full scan. This must run in the
// server process: with the external scanner, an ANALYZE in the subprocess is invisible to the
// server's pooled connections; their shared schema cache keeps the old statistics until the
// process restarts.
if effectiveFullScan && scanError == nil {
if err := db.Optimize(ctx); err != nil {
log.Error(ctx, "Scanner: Error analyzing DB", err)
}
}
// If changes were detected, send a refresh event to all clients
if s.changesDetected {
log.Debug(ctx, "Library changes imported. Sending refresh event")
s.broker.SendBroadcastMessage(ctx, &events.RefreshResource{})
}
// Send the final scan status event, with totals
if count, folderCount, err := s.getCounters(ctx); err != nil {
s.metrics.WriteAfterScanMetrics(ctx, false)
return scanWarnings, err
} else {
scanType, elapsed, lastErr := s.getScanInfo(ctx)
s.metrics.WriteAfterScanMetrics(ctx, true)
s.sendMessage(ctx, &events.ScanStatus{
Scanning: false,
Count: count,
FolderCount: folderCount,
Error: lastErr,
ScanType: scanType,
ElapsedTime: elapsed,
})
}
return scanWarnings, scanError
}
// This is a global variable that is used to prevent multiple scans from running at the same time.
// "There can be only one" - https://youtu.be/sqcLjcSloXs?si=VlsjEOjTJZ68zIyg
var (
running atomic.Bool
scanMaintenanceMux sync.Mutex
)
func lockScan(ctx context.Context) (func(), error) {
if !running.CompareAndSwap(false, true) {
log.Debug(ctx, "Scanner already running, ignoring request")
return func() {}, ErrAlreadyScanning
}
scanMaintenanceMux.Lock()
return func() {
scanMaintenanceMux.Unlock()
running.Store(false)
}, nil
}
// LockForMaintenance prevents a scan from starting while database maintenance is running.
func LockForMaintenance() (func(), bool) {
if !scanMaintenanceMux.TryLock() {
return func() {}, false
}
if running.Load() {
scanMaintenanceMux.Unlock()
return func() {}, false
}
return scanMaintenanceMux.Unlock, true
}
// EffectiveFullScan reports whether a scan was requested as full, will resume an interrupted full scan,
// or will rescan a library in full because its PID config changed, in one of the included libraries.
func EffectiveFullScan(ctx context.Context, ds model.DataStore, fullScan bool, targets []model.ScanTarget) bool {
if fullScan {
return true
}
return anyIncludedLibrary(ctx, ds, targets, func(library model.Library) bool {
return library.FullScanInProgress || library.NeedsPIDRescan()
})
}
func (s *controller) includesUnscannedLibrary(ctx context.Context, targets []model.ScanTarget) bool {
return anyIncludedLibrary(ctx, s.ds, targets, func(library model.Library) bool {
return library.LastScanAt.IsZero()
})
}
// anyIncludedLibrary reports whether any library included in the scan (all of them when targets is
// empty) matches pred.
func anyIncludedLibrary(ctx context.Context, ds model.DataStore, targets []model.ScanTarget, pred func(model.Library) bool) bool {
libraries, err := ds.Library().GetAll(ctx)
if err != nil {
return false
}
if len(targets) == 0 {
return slices.ContainsFunc(libraries, pred)
}
targeted := make(map[int]struct{}, len(targets))
for _, target := range targets {
targeted[target.LibraryID] = struct{}{}
}
return slices.ContainsFunc(libraries, func(library model.Library) bool {
_, ok := targeted[library.ID]
return ok && pred(library)
})
}
func (s *controller) trackProgress(ctx context.Context, progress <-chan *ProgressInfo) ([]string, error) {
s.count.Store(0)
s.folderCount.Store(0)
s.changesDetected = false
var warnings []string
var errs []error
for p := range pl.ReadOrDone(ctx, progress) {
if p.Error != "" {
errs = append(errs, errors.New(p.Error))
continue
}
if p.Warning != "" {
warnings = append(warnings, p.Warning)
continue
}
if p.ChangesDetected {
s.changesDetected = true
continue
}
s.count.Add(p.FileCount)
if p.FileCount > 0 {
s.folderCount.Add(1)
}
scanType, elapsed, lastErr := s.getScanInfo(ctx)
status := &events.ScanStatus{
Scanning: true,
Count: int64(s.count.Load()),
FolderCount: int64(s.folderCount.Load()),
Error: lastErr,
ScanType: scanType,
ElapsedTime: elapsed,
}
if s.limiter != nil && !p.ForceUpdate {
s.limiter.Do(func() { s.sendMessage(ctx, status) })
} else {
s.sendMessage(ctx, status)
}
}
return warnings, errors.Join(errs...)
}
func (s *controller) sendMessage(ctx context.Context, status *events.ScanStatus) {
s.broker.SendBroadcastMessage(ctx, status)
}
// GetInstance returns the scanner singleton: Status reads the progress counters of the controller
// running the scan, and scheduler, watcher and signal scans do not start from the API's injector.
func GetInstance(rootCtx context.Context, ds model.DataStore, broker events.Broker,
pls playlists.Playlists, m metrics.Metrics) model.Scanner {
return singleton.GetInstance(func() *controller {
return New(rootCtx, ds, broker, pls, m).(*controller)
})
}