Added an option to merge artists without MBIDs to those with one, fixes #642

This commit is contained in:
emeric
2025-04-04 18:30:14 +02:00
parent 7b12dc0f70
commit faabb7ec4a
35 changed files with 1080 additions and 174 deletions
@@ -32,7 +32,10 @@
#include "IFileScanOperation.hpp"
#include "ScanContext.hpp"
#include "ScannerSettings.hpp"
#include "Utils.hpp"
#include "helpers/ArtistHelpers.hpp"
#include "metadata/Types.hpp"
namespace lms::scanner
{
@@ -41,10 +44,13 @@ namespace lms::scanner
class ArtistInfoFileScanOperation : public IFileScanOperation
{
public:
ArtistInfoFileScanOperation(const FileToScan& file, db::Db& db)
ArtistInfoFileScanOperation(const FileToScan& file, const ScannerSettings& settings, db::Db& db)
: _file{ file.file }
, _mediaLibrary{ file.mediaLibrary }
, _db{ db } {}
, _settings{ settings }
, _db{ db }
{
}
~ArtistInfoFileScanOperation() override = default;
ArtistInfoFileScanOperation(const ArtistInfoFileScanOperation&) = delete;
ArtistInfoFileScanOperation& operator=(const ArtistInfoFileScanOperation&) = delete;
@@ -59,6 +65,7 @@ namespace lms::scanner
const std::filesystem::path _file;
const MediaLibraryInfo _mediaLibrary;
const ScannerSettings& _settings;
db::Db& _db;
std::optional<metadata::ArtistInfo> _parsedArtistInfo;
@@ -76,12 +83,7 @@ namespace lms::scanner
}
_parsedArtistInfo = metadata::parseArtistInfo(ifs);
if (!_parsedArtistInfo->mbid.has_value())
{
LMS_LOG(DBUPDATER, DEBUG, "Discarding artist info in file " << _file << ": no mbid set");
_parsedArtistInfo.reset();
}
else if (_parsedArtistInfo->name.empty())
if (_parsedArtistInfo->name.empty())
{
LMS_LOG(DBUPDATER, DEBUG, "Discarding artist info in file " << _file << ": no name set");
_parsedArtistInfo.reset();
@@ -125,6 +127,8 @@ namespace lms::scanner
artistInfo.modify()->setAbsoluteFilePath(_file);
}
artistInfo.modify()->setName(_parsedArtistInfo->name);
artistInfo.modify()->setSortName(_parsedArtistInfo->sortName);
artistInfo.modify()->setLastWriteTime(fileInfo->lastWriteTime);
artistInfo.modify()->setType(_parsedArtistInfo->type);
artistInfo.modify()->setGender(_parsedArtistInfo->gender);
@@ -134,12 +138,8 @@ namespace lms::scanner
db::MediaLibrary::pointer mediaLibrary{ db::MediaLibrary::find(dbSession, _mediaLibrary.id) }; // may be null if settings are updated in // => next scan will correct this
artistInfo.modify()->setDirectory(utils::getOrCreateDirectory(dbSession, _file.parent_path(), mediaLibrary));
db::Artist::pointer artist{ db::Artist::find(dbSession, *_parsedArtistInfo->mbid) };
if (!artist)
artist = dbSession.create<db::Artist>(_parsedArtistInfo->name, _parsedArtistInfo->mbid);
artist.modify()->setName(_parsedArtistInfo->name);
artist.modify()->setSortName(_parsedArtistInfo->sortName);
const metadata::Artist artistMetadata{ _parsedArtistInfo->mbid, _parsedArtistInfo->name, _parsedArtistInfo->sortName.empty() ? std::nullopt : std::make_optional<std::string>(_parsedArtistInfo->sortName) };
db::Artist::pointer artist{ helpers::getOrCreateArtist(dbSession, artistMetadata, helpers::AllowFallbackOnMBIDEntry{ _settings.allowArtistMBIDFallback }) };
artistInfo.modify()->setArtist(artist);
if (added)
@@ -155,8 +155,9 @@ namespace lms::scanner
}
} // namespace
ArtistInfoFileScanner::ArtistInfoFileScanner(db::Db& db)
: _db{ db }
ArtistInfoFileScanner::ArtistInfoFileScanner(const ScannerSettings& settings, db::Db& db)
: _settings{ settings }
, _db{ db }
{
}
@@ -202,6 +203,6 @@ namespace lms::scanner
std::unique_ptr<IFileScanOperation> ArtistInfoFileScanner::createScanOperation(const FileToScan& fileToScan) const
{
return std::make_unique<ArtistInfoFileScanOperation>(fileToScan, _db);
return std::make_unique<ArtistInfoFileScanOperation>(fileToScan, _settings, _db);
}
} // namespace lms::scanner
@@ -31,10 +31,12 @@ namespace lms
namespace lms::scanner
{
struct ScannerSettings;
class ArtistInfoFileScanner : public IFileScanner
{
public:
ArtistInfoFileScanner(db::Db& db);
ArtistInfoFileScanner(const ScannerSettings& _settings, db::Db& db);
~ArtistInfoFileScanner() override = default;
ArtistInfoFileScanner(const ArtistInfoFileScanner&) = delete;
ArtistInfoFileScanner& operator=(const ArtistInfoFileScanner&) = delete;
@@ -45,6 +47,7 @@ namespace lms::scanner
bool needsScan(ScanContext& context, const FileToScan& file) const override;
std::unique_ptr<IFileScanOperation> createScanOperation(const FileToScan& fileToScan) const override;
const ScannerSettings& _settings;
db::Db& _db;
};
} // namespace lms::scanner
@@ -46,92 +46,30 @@
#include "ScanContext.hpp"
#include "ScannerSettings.hpp"
#include "Utils.hpp"
#include "helpers/ArtistHelpers.hpp"
namespace lms::scanner
{
namespace
{
db::Artist::pointer createArtist(db::Session& session, const metadata::Artist& artistInfo)
void createTrackArtistLinks(db::Session& session, const db::Track::pointer& track, db::TrackArtistLinkType linkType, std::string_view role, std::span<const metadata::Artist> artists, helpers::AllowFallbackOnMBIDEntry allowArtistMBIDFallback)
{
db::Artist::pointer artist{ session.create<db::Artist>(artistInfo.name) };
if (artistInfo.mbid)
artist.modify()->setMBID(artistInfo.mbid);
artist.modify()->setSortName(artistInfo.sortName ? *artistInfo.sortName : artistInfo.name);
return artist;
}
std::string optionalMBIDAsString(const std::optional<core::UUID>& uuid)
{
return uuid ? std::string{ uuid->getAsString() } : "<no MBID>";
}
void updateArtistIfNeeded(db::Artist::pointer artist, const metadata::Artist& artistInfo)
{
// Name may have been updated
if (artist->getName() != artistInfo.name)
for (const metadata::Artist& artistInfo : artists)
{
LMS_LOG(DBUPDATER, DEBUG, "Artist [" << optionalMBIDAsString(artist->getMBID()) << "], updated name from '" << artist->getName() << "' to '" << artistInfo.name << "'");
artist.modify()->setName(artistInfo.name);
}
db::Artist::pointer artist{ helpers::getOrCreateArtist(session, artistInfo, allowArtistMBIDFallback) };
// Sortname may have been updated
// As the sort name is quite often not filled in, we update it only if already set (for now?)
if (artistInfo.sortName && *artistInfo.sortName != artist->getSortName())
{
LMS_LOG(DBUPDATER, DEBUG, "Artist [" << optionalMBIDAsString(artist->getMBID()) << "], updated sort name from '" << artist->getSortName() << "' to '" << *artistInfo.sortName << "'");
artist.modify()->setSortName(*artistInfo.sortName);
const bool matchedUsingMbid{ artist->getMBID() == artistInfo.mbid };
db::TrackArtistLink::pointer link{ session.create<db::TrackArtistLink>(track, artist, linkType, role, matchedUsingMbid) };
link.modify()->setArtistName(artistInfo.name);
if (artistInfo.sortName)
link.modify()->setArtistSortName(*artistInfo.sortName);
}
}
std::vector<db::Artist::pointer> getOrCreateArtists(db::Session& session, const std::vector<metadata::Artist>& artistsInfo, bool allowFallbackOnMBIDEntries)
void createTrackArtistLinks(db::Session& session, const db::Track::pointer& track, db::TrackArtistLinkType linkType, std::span<const metadata::Artist> artists, helpers::AllowFallbackOnMBIDEntry allowArtistMBIDFallback)
{
std::vector<db::Artist::pointer> artists;
for (const metadata::Artist& artistInfo : artistsInfo)
{
db::Artist::pointer artist;
// First try to get by MBID
if (artistInfo.mbid)
{
artist = db::Artist::find(session, *artistInfo.mbid);
if (!artist)
artist = createArtist(session, artistInfo);
else
updateArtistIfNeeded(artist, artistInfo);
artists.emplace_back(std::move(artist));
continue;
}
// Fall back on artist name (collisions may occur)
if (!artistInfo.name.empty())
{
for (const db::Artist::pointer& sameNamedArtist : db::Artist::find(session, artistInfo.name))
{
// Do not fallback on artist that is correctly tagged
if (!allowFallbackOnMBIDEntries && sameNamedArtist->getMBID())
continue;
artist = sameNamedArtist;
break;
}
// No Artist found with the same name and without MBID -> creating
if (!artist)
artist = createArtist(session, artistInfo);
else
updateArtistIfNeeded(artist, artistInfo);
artists.emplace_back(std::move(artist));
continue;
}
}
return artists;
constexpr std::string_view noRole{};
createTrackArtistLinks(session, track, linkType, noRole, artists, allowArtistMBIDFallback);
}
db::ReleaseType::pointer getOrCreateReleaseType(db::Session& session, std::string_view name)
@@ -609,7 +547,7 @@ namespace lms::scanner
return;
}
if (_parsedTrack->mbid && (!track || _settings.skipDuplicateMBID))
if (_parsedTrack->mbid && (!track || _settings.skipDuplicateTrackMBID))
{
std::vector<db::Track::pointer> duplicateTracks{ db::Track::findByMBID(dbSession, *_parsedTrack->mbid) };
@@ -627,7 +565,7 @@ namespace lms::scanner
}
// Skip duplicate track MBID
if (_settings.skipDuplicateMBID)
if (_settings.skipDuplicateTrackMBID)
{
for (db::Track::pointer& otherTrack : duplicateTracks)
{
@@ -739,41 +677,20 @@ namespace lms::scanner
track.modify()->setDirectory(directory);
track.modify()->clearArtistLinks();
// Do not fallback on artists with the same name but having a MBID for artist and releaseArtists, as it may be corrected by properly tagging files
for (const db::Artist::pointer& artist : getOrCreateArtists(dbSession, _parsedTrack->artists, false))
track.modify()->addArtistLink(db::TrackArtistLink::create(dbSession, track, artist, db::TrackArtistLinkType::Artist));
const helpers::AllowFallbackOnMBIDEntry allowFallback{ _settings.allowArtistMBIDFallback };
createTrackArtistLinks(dbSession, track, db::TrackArtistLinkType::Artist, _parsedTrack->artists, allowFallback);
if (_parsedTrack->medium && _parsedTrack->medium->release)
{
for (const db::Artist::pointer& releaseArtist : getOrCreateArtists(dbSession, _parsedTrack->medium->release->artists, false))
track.modify()->addArtistLink(db::TrackArtistLink::create(dbSession, track, releaseArtist, db::TrackArtistLinkType::ReleaseArtist));
}
createTrackArtistLinks(dbSession, track, db::TrackArtistLinkType::ReleaseArtist, _parsedTrack->medium->release->artists, allowFallback);
// Allow fallbacks on artists with the same name even if they have MBID, since there is no tag to indicate the MBID of these artists
// We could ask MusicBrainz to get all the information, but that would heavily slow down the import process
for (const db::Artist::pointer& conductor : getOrCreateArtists(dbSession, _parsedTrack->conductorArtists, true))
track.modify()->addArtistLink(db::TrackArtistLink::create(dbSession, track, conductor, db::TrackArtistLinkType::Conductor));
for (const db::Artist::pointer& composer : getOrCreateArtists(dbSession, _parsedTrack->composerArtists, true))
track.modify()->addArtistLink(db::TrackArtistLink::create(dbSession, track, composer, db::TrackArtistLinkType::Composer));
for (const db::Artist::pointer& lyricist : getOrCreateArtists(dbSession, _parsedTrack->lyricistArtists, true))
track.modify()->addArtistLink(db::TrackArtistLink::create(dbSession, track, lyricist, db::TrackArtistLinkType::Lyricist));
for (const db::Artist::pointer& mixer : getOrCreateArtists(dbSession, _parsedTrack->mixerArtists, true))
track.modify()->addArtistLink(db::TrackArtistLink::create(dbSession, track, mixer, db::TrackArtistLinkType::Mixer));
createTrackArtistLinks(dbSession, track, db::TrackArtistLinkType::Conductor, _parsedTrack->conductorArtists, allowFallback);
createTrackArtistLinks(dbSession, track, db::TrackArtistLinkType::Composer, _parsedTrack->composerArtists, allowFallback);
createTrackArtistLinks(dbSession, track, db::TrackArtistLinkType::Lyricist, _parsedTrack->lyricistArtists, allowFallback);
createTrackArtistLinks(dbSession, track, db::TrackArtistLinkType::Mixer, _parsedTrack->mixerArtists, allowFallback);
createTrackArtistLinks(dbSession, track, db::TrackArtistLinkType::Remixer, _parsedTrack->remixerArtists, allowFallback);
for (const auto& [role, performers] : _parsedTrack->performerArtists)
{
for (const db::Artist::pointer& performer : getOrCreateArtists(dbSession, performers, true))
track.modify()->addArtistLink(db::TrackArtistLink::create(dbSession, track, performer, db::TrackArtistLinkType::Performer, role));
}
for (const db::Artist::pointer& producer : getOrCreateArtists(dbSession, _parsedTrack->producerArtists, true))
track.modify()->addArtistLink(db::TrackArtistLink::create(dbSession, track, producer, db::TrackArtistLinkType::Producer));
for (const db::Artist::pointer& remixer : getOrCreateArtists(dbSession, _parsedTrack->remixerArtists, true))
track.modify()->addArtistLink(db::TrackArtistLink::create(dbSession, track, remixer, db::TrackArtistLinkType::Remixer));
createTrackArtistLinks(dbSession, track, db::TrackArtistLinkType::Performer, role, performers, allowFallback);
track.modify()->setScanVersion(_settings.scanVersion);
if (_parsedTrack->medium && _parsedTrack->medium->release)