Implement the OpenSubsonic transcoding extension (getTranscodeDecision/getTranscodeStream)

Adds a real capability-matching decision engine: getTranscodeDecision compares the source file's actual scanned format/channels against the client's declared directPlayProfiles first, and only when nothing matches does it search transcodingProfiles for a codec VibeSonic's ffmpeg pipeline can actually encode (mp3/opus/aac), applying any codecProfile limitations (bitrate/channel required constraints) the client listed. canTranscode is never set true for a codec the server has no encoder for, and canDirectPlay is never set true without a matching client profile. getTranscodeStream takes the opaque transcodeParams token the decision returned (never re-derived) plus a real time offset and streams actual transcoded audio via the existing seek-capable transcoding.Manager. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014r98AtcBbevpscWLsfwVvv
2026-08-23 08:23:08 -05:00
parent ec28053d36
commit 6fb46a989e
3 changed files with 502 additions and 1 deletions
@@ -55,7 +55,7 @@ func TestGetOpenSubsonicExtensions(t *testing.T) {
 	// set once Phase 3 is done.
 	for _, want := range []string{
 		"apiKeyAuthentication", "formPost", "getPodcastEpisode", "topSongsByArtistId", "transcodeOffset",
-		"indexBasedQueue", "playbackReport", "songLyrics",
+		"indexBasedQueue", "playbackReport", "songLyrics", "transcoding",
 	} {
 		if !names[want] {
 			t.Errorf("extension %q not advertised: %v", want, names)
@@ -0,0 +1,324 @@
+package subsonic
+
+import (
+	"encoding/json"
+	"fmt"
+	"net/url"
+	"strconv"
+	"strings"
+
+	"git.noroutes.net/NoRoutes/vibetunes/internal/database"
+	"git.noroutes.net/NoRoutes/vibetunes/internal/streaming"
+	"git.noroutes.net/NoRoutes/vibetunes/internal/transcoding"
+)
+
+func init() {
+	registerExtension("transcoding", 1)
+	register("getTranscodeDecision", handleGetTranscodeDecision)
+	register("getTranscodeStream", handleGetTranscodeStream)
+}
+
+// codecFormats are the transcode targets VibeSonic's ffmpeg pipeline can
+// actually produce. A client's transcodingProfile.audioCodec must name
+// one of these for VibeSonic to be able to honor it — canTranscode is
+// never set true for a codec we can't actually encode.
+var codecFormats = map[string]transcoding.Format{
+	"mp3":  transcoding.FormatMP3,
+	"opus": transcoding.FormatOpus,
+	"aac":  transcoding.FormatAAC,
+}
+
+func containsFold(list []string, want string) bool {
+	for _, s := range list {
+		if strings.EqualFold(s, want) {
+			return true
+		}
+	}
+	return false
+}
+
+// matchesDirectPlay reports whether a source file, as actually scanned
+// (its real container/format, channel count), satisfies one of the
+// client's declared direct-play profiles.
+func matchesDirectPlay(mf *database.MediaFile, profiles []DirectPlayProfile) bool {
+	for _, p := range profiles {
+		if len(p.Protocols) > 0 && !containsFold(p.Protocols, "http") {
+			continue
+		}
+		if !containsFold(p.Containers, mf.Format) {
+			continue
+		}
+		if len(p.AudioCodecs) > 0 && !containsFold(p.AudioCodecs, mf.Format) {
+			continue
+		}
+		if p.MaxAudioChannels > 0 && mf.Channels > p.MaxAudioChannels {
+			continue
+		}
+		return true
+	}
+	return false
+}
+
+// limitationOK evaluates one CodecProfile limitation against a candidate
+// value VibeSonic would actually produce for that transcode.
+func limitationOK(lim Limitation, value int) bool {
+	if len(lim.Values) == 0 {
+		return true
+	}
+	for _, raw := range lim.Values {
+		want, err := strconv.Atoi(raw)
+		if err != nil {
+			continue
+		}
+		switch lim.Comparison {
+		case "Equals":
+			if value == want {
+				return true
+			}
+		case "NotEquals":
+			if value != want {
+				return true
+			}
+		case "LessThanEqual":
+			if value <= want {
+				return true
+			}
+		case "GreaterThanEqual":
+			if value >= want {
+				return true
+			}
+		}
+	}
+	return false
+}
+
+// codecProfileFor finds the CodecProfile (if any) the client declared
+// for a given audio codec name.
+func codecProfileFor(profiles []CodecProfile, codec string) *CodecProfile {
+	for i := range profiles {
+		if strings.EqualFold(profiles[i].Type, "audio") && strings.EqualFold(profiles[i].Name, codec) {
+			return &profiles[i]
+		}
+	}
+	return nil
+}
+
+// transcodePlan is what VibeSonic decided to actually produce for a
+// chosen transcoding profile: a real ffmpeg format, bitrate, and channel
+// count that satisfy every limitation the client marked required.
+type transcodePlan struct {
+	format    transcoding.Format
+	container string
+	codec     string
+	bitrate   int
+	channels  int
+}
+
+// planTranscode picks the first client-declared transcoding profile
+// VibeSonic can actually fulfill: a codec its ffmpeg pipeline supports,
+// with a bitrate/channel count that satisfies every "required" codec
+// limitation the client listed. It never claims to satisfy a required
+// limitation it doesn't actually meet.
+func planTranscode(mf *database.MediaFile, client ClientInfo) *transcodePlan {
+	for _, tp := range client.TranscodingProfiles {
+		if tp.Protocol != "" && !strings.EqualFold(tp.Protocol, "http") {
+			continue
+		}
+		format, ok := codecFormats[strings.ToLower(tp.AudioCodec)]
+		if !ok {
+			continue
+		}
+
+		bitrate := client.MaxTranscodingAudioBitrate
+		if bitrate <= 0 {
+			bitrate = 192
+		}
+		channels := mf.Channels
+		if tp.MaxAudioChannels > 0 && channels > tp.MaxAudioChannels {
+			channels = tp.MaxAudioChannels
+		}
+
+		if cp := codecProfileFor(client.CodecProfiles, tp.AudioCodec); cp != nil {
+			ok := true
+			for _, lim := range cp.Limitations {
+				switch lim.Name {
+				case "audioBitrate":
+					if !limitationOK(lim, bitrate) {
+						if !lim.Required {
+							continue
+						}
+						ok = false
+					}
+				case "audioChannels":
+					if !limitationOK(lim, channels) {
+						if !lim.Required {
+							continue
+						}
+						ok = false
+					}
+				}
+				if !ok {
+					break
+				}
+			}
+			if !ok {
+				continue
+			}
+		}
+
+		return &transcodePlan{
+			format:    format,
+			container: tp.Container,
+			codec:     tp.AudioCodec,
+			bitrate:   bitrate,
+			channels:  channels,
+		}
+	}
+	return nil
+}
+
+func handleGetTranscodeDecision(rc *Context) *apiError {
+	if !rc.User.Allows(rc.User.Permissions.Stream) {
+		return errUnauthorized("streaming")
+	}
+	mediaID, aerr := rc.Params.Required("mediaId")
+	if aerr != nil {
+		return aerr
+	}
+	mediaType, aerr := rc.Params.Required("mediaType")
+	if aerr != nil {
+		return aerr
+	}
+
+	var client ClientInfo
+	if rc.R.Body != nil {
+		defer rc.R.Body.Close()
+		if err := json.NewDecoder(rc.R.Body).Decode(&client); err != nil {
+			return newError(ErrGeneric, "invalid ClientInfo request body: "+err.Error())
+		}
+	}
+
+	db := rc.Server.DB
+	song, err := resolveMediaID(db, mediaID, mediaType)
+	if err != nil {
+		return errNotFound("Media")
+	}
+	mf, err := database.GetMediaFile(db, song.MediaFileID)
+	if err != nil {
+		return errNotFound("Media")
+	}
+
+	source := &StreamDetails{
+		Protocol:        "http",
+		Container:       mf.Format,
+		Codec:           mf.Format,
+		AudioChannels:   mf.Channels,
+		AudioBitrate:    mf.BitrateKbps,
+		AudioSamplerate: mf.SampleRateHz,
+	}
+
+	decision := TranscodeDecision{SourceStream: source}
+
+	if matchesDirectPlay(mf, client.DirectPlayProfiles) {
+		decision.CanDirectPlay = true
+		rc.Response = newResponse()
+		rc.Response.TranscodeDecision = &decision
+		return nil
+	}
+	decision.TranscodeReason = []string{"source format not in client's direct-play profiles"}
+
+	plan := planTranscode(mf, client)
+	if plan == nil {
+		decision.CanTranscode = false
+		decision.ErrorReason = "no client transcoding profile is compatible with what this server can produce"
+		rc.Response = newResponse()
+		rc.Response.TranscodeDecision = &decision
+		return nil
+	}
+
+	decision.CanTranscode = true
+	decision.TranscodeStream = &StreamDetails{
+		Protocol:      "http",
+		Container:     plan.container,
+		Codec:         plan.codec,
+		AudioChannels: plan.channels,
+		AudioBitrate:  plan.bitrate,
+	}
+	decision.TranscodeParams = encodeTranscodeParams(plan)
+
+	rc.Response = newResponse()
+	rc.Response.TranscodeDecision = &decision
+	return nil
+}
+
+// encodeTranscodeParams and decodeTranscodeParams round-trip a
+// transcodePlan through an opaque server-specific token, per the spec's
+// instruction that clients must pass back exactly what
+// getTranscodeDecision gave them rather than reconstructing it.
+func encodeTranscodeParams(plan *transcodePlan) string {
+	v := url.Values{}
+	v.Set("format", string(plan.format))
+	v.Set("bitrate", strconv.Itoa(plan.bitrate))
+	v.Set("channels", strconv.Itoa(plan.channels))
+	return v.Encode()
+}
+
+func decodeTranscodeParams(s string) (transcoding.Options, error) {
+	v, err := url.ParseQuery(s)
+	if err != nil {
+		return transcoding.Options{}, err
+	}
+	format, ok := codecFormats[v.Get("format")]
+	if !ok {
+		return transcoding.Options{}, fmt.Errorf("unsupported transcodeParams format %q", v.Get("format"))
+	}
+	bitrate, _ := strconv.Atoi(v.Get("bitrate"))
+	return transcoding.Options{Format: format, MaxBitRateKbps: bitrate}, nil
+}
+
+func handleGetTranscodeStream(rc *Context) *apiError {
+	if !rc.User.Allows(rc.User.Permissions.Stream) {
+		return errUnauthorized("streaming")
+	}
+	mediaID, aerr := rc.Params.Required("mediaId")
+	if aerr != nil {
+		return aerr
+	}
+	mediaType, aerr := rc.Params.Required("mediaType")
+	if aerr != nil {
+		return aerr
+	}
+	transcodeParams, aerr := rc.Params.Required("transcodeParams")
+	if aerr != nil {
+		return aerr
+	}
+	offset := rc.Params.Int("offset", 0)
+
+	db := rc.Server.DB
+	song, err := resolveMediaID(db, mediaID, mediaType)
+	if err != nil {
+		return errNotFound("Media")
+	}
+	mf, folder, aerr := resolveSongFile(rc, song.ID)
+	if aerr != nil {
+		return aerr
+	}
+	absPath, perr := streaming.ResolvePath(folder.Path, mf.RelativePath)
+	if perr != nil {
+		return newError(ErrGeneric, perr.Error())
+	}
+
+	opts, err := decodeTranscodeParams(transcodeParams)
+	if err != nil {
+		return newError(ErrGeneric, err.Error())
+	}
+	opts.TimeOffset = float64(offset)
+
+	rc.Handled = true
+	rc.W.Header().Set("Content-Type", opts.Format.MimeType())
+	ctx := rc.R.Context()
+	if err := rc.Server.Transcoder.Stream(ctx, rc.W, absPath, opts); err != nil && ctx.Err() == nil {
+		rc.Server.Log.Warn("transcode failed", "err", err, "path", absPath)
+	}
+	return nil
+}
@@ -0,0 +1,177 @@
+package subsonic
+
+import (
+	"bytes"
+	"encoding/json"
+	"io"
+	"net/http"
+	"net/url"
+	"testing"
+
+	"git.noroutes.net/NoRoutes/vibetunes/internal/database"
+)
+
+func postTranscodeDecision(t *testing.T, srv string, q map[string]string, client ClientInfo) *http.Response {
+	t.Helper()
+	body, err := json.Marshal(client)
+	if err != nil {
+		t.Fatal(err)
+	}
+	format := q["f"]
+	if format == "" {
+		format = "xml"
+	}
+	url := srv + "/rest/getTranscodeDecision.view?u=" + q["u"] + "&p=" + q["p"] + "&v=1.16.1&c=vibesonic-test&f=" + format + "&mediaId=" + q["mediaId"] + "&mediaType=" + q["mediaType"]
+	req, err := http.NewRequest(http.MethodPost, url, bytes.NewReader(body))
+	if err != nil {
+		t.Fatal(err)
+	}
+	req.Header.Set("Content-Type", "application/json")
+	resp, err := http.DefaultClient.Do(req)
+	if err != nil {
+		t.Fatal(err)
+	}
+	return resp
+}
+
+func TestTranscodeDecisionDirectPlay(t *testing.T) {
+	s, srv := newTestRecorderServer(t)
+	createTestUser(t, s, "alice", "password123", false)
+	seedLibrary(t, s)
+	song := firstSong(t, s)
+
+	client := ClientInfo{
+		Name:     "TestClient",
+		Platform: "linux",
+		DirectPlayProfiles: []DirectPlayProfile{
+			{Containers: []string{"mp3"}, AudioCodecs: []string{"mp3"}, Protocols: []string{"http"}},
+		},
+	}
+	resp := postTranscodeDecision(t, srv.URL, map[string]string{"u": "alice", "p": "password123", "mediaId": song.ID, "mediaType": "song", "f": "json"}, client)
+	j := decodeJSON(t, resp)
+	d, ok := j["transcodeDecision"].(map[string]any)
+	if !ok {
+		t.Fatalf("transcodeDecision missing: %v", j)
+	}
+	if d["canDirectPlay"] != true {
+		t.Fatalf("canDirectPlay = %v, want true (client declared mp3 direct play, fixture songs are mp3)", d)
+	}
+	if _, has := d["transcodeStream"]; has {
+		t.Error("transcodeStream should be absent when direct play is possible")
+	}
+}
+
+func TestTranscodeDecisionFallsBackToTranscode(t *testing.T) {
+	s, srv := newTestRecorderServer(t)
+	createTestUser(t, s, "alice", "password123", false)
+	seedLibrary(t, s)
+	song := firstSong(t, s)
+
+	client := ClientInfo{
+		Name:     "TestClient",
+		Platform: "linux",
+		DirectPlayProfiles: []DirectPlayProfile{
+			{Containers: []string{"flac"}, AudioCodecs: []string{"flac"}, Protocols: []string{"http"}},
+		},
+		TranscodingProfiles: []TranscodingProfile{
+			{Container: "ogg", AudioCodec: "opus", Protocol: "http", MaxAudioChannels: 2},
+		},
+		MaxTranscodingAudioBitrate: 96,
+	}
+	resp := postTranscodeDecision(t, srv.URL, map[string]string{"u": "alice", "p": "password123", "mediaId": song.ID, "mediaType": "song", "f": "json"}, client)
+	j := decodeJSON(t, resp)
+	d := j["transcodeDecision"].(map[string]any)
+	if d["canDirectPlay"] != nil && d["canDirectPlay"] != false {
+		t.Errorf("canDirectPlay = %v, want false (source is mp3, client only direct-plays flac)", d["canDirectPlay"])
+	}
+	if d["canTranscode"] != true {
+		t.Fatalf("canTranscode = %v, want true", d)
+	}
+	ts, ok := d["transcodeStream"].(map[string]any)
+	if !ok {
+		t.Fatalf("transcodeStream missing: %v", d)
+	}
+	if ts["codec"] != "opus" || ts["audioBitrate"] != float64(96) {
+		t.Errorf("transcodeStream = %v, want opus @ 96kbps", ts)
+	}
+	params, ok := d["transcodeParams"].(string)
+	if !ok || params == "" {
+		t.Fatalf("transcodeParams missing: %v", d)
+	}
+
+	// getTranscodeStream must accept exactly the transcodeParams token
+	// getTranscodeDecision returned and actually produce transcoded
+	// audio bytes.
+	streamResp, err := http.Get(srv.URL + "/rest/getTranscodeStream.view?u=alice&p=password123&v=1.16.1&c=vibesonic-test&mediaId=" + song.ID + "&mediaType=song&transcodeParams=" + url.QueryEscape(params))
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer streamResp.Body.Close()
+	if streamResp.StatusCode != http.StatusOK {
+		t.Fatalf("getTranscodeStream status = %d", streamResp.StatusCode)
+	}
+	data, _ := io.ReadAll(streamResp.Body)
+	if len(data) == 0 {
+		t.Fatal("getTranscodeStream returned no bytes")
+	}
+	if ct := streamResp.Header.Get("Content-Type"); ct != "audio/ogg" {
+		t.Errorf("Content-Type = %q, want audio/ogg for an opus transcode", ct)
+	}
+}
+
+func TestTranscodeDecisionNoCompatibleProfile(t *testing.T) {
+	s, srv := newTestRecorderServer(t)
+	createTestUser(t, s, "alice", "password123", false)
+	seedLibrary(t, s)
+	song := firstSong(t, s)
+
+	client := ClientInfo{
+		Name:     "PickyClient",
+		Platform: "linux",
+		DirectPlayProfiles: []DirectPlayProfile{
+			{Containers: []string{"flac"}, AudioCodecs: []string{"flac"}, Protocols: []string{"http"}},
+		},
+		TranscodingProfiles: []TranscodingProfile{
+			{Container: "wma", AudioCodec: "wmav2", Protocol: "http"},
+		},
+	}
+	resp := postTranscodeDecision(t, srv.URL, map[string]string{"u": "alice", "p": "password123", "mediaId": song.ID, "mediaType": "song", "f": "json"}, client)
+	j := decodeJSON(t, resp)
+	d := j["transcodeDecision"].(map[string]any)
+	if d["canTranscode"] != nil && d["canTranscode"] != false {
+		t.Errorf("canTranscode = %v, want false (server has no wmav2 encoder)", d["canTranscode"])
+	}
+	if d["errorReason"] == nil || d["errorReason"] == "" {
+		t.Error("expected a non-empty errorReason when no profile is compatible")
+	}
+}
+
+func TestTranscodeDecisionRequiresStreamPermission(t *testing.T) {
+	s, db := newTestServer(t)
+	u := createTestUser(t, s, "alice", "password123", false)
+	perms := u.Permissions
+	perms.Stream = false
+	if err := database.SetUserPermissions(db, u.ID, perms); err != nil {
+		t.Fatal(err)
+	}
+	srv := newSubsonicTestServer(t, s)
+	seedLibrary(t, s)
+	song := firstSong(t, s)
+
+	resp := postTranscodeDecision(t, srv, map[string]string{"u": "alice", "p": "password123", "mediaId": song.ID, "mediaType": "song"}, ClientInfo{})
+	x := decodeXML(t, resp)
+	if x.Status != "failed" || x.Error == nil || x.Error.Code != ErrUnauthorized {
+		t.Fatalf("expected error 50, got %+v", x)
+	}
+}
+
+func TestTranscodeDecisionUnknownMedia(t *testing.T) {
+	s, srv := newTestRecorderServer(t)
+	createTestUser(t, s, "alice", "password123", false)
+
+	resp := postTranscodeDecision(t, srv.URL, map[string]string{"u": "alice", "p": "password123", "mediaId": "sg_doesnotexist", "mediaType": "song"}, ClientInfo{})
+	x := decodeXML(t, resp)
+	if x.Status != "failed" || x.Error == nil || x.Error.Code != ErrNotFound {
+		t.Fatalf("expected error 70, got %+v", x)
+	}
+}