Remove dead code

cmd/cli/readline/errors.go

@@ -5,11 +5,3 @@ import (
 )
 
 var ErrInterrupt = errors.New("Interrupt")
-
-type InterruptError struct {
-	Line []rune
-}
-
-func (*InterruptError) Error() string {
-	return "Interrupted"
-}

cmd/cli/readline/history.go

@@ -98,10 +98,6 @@ func (h *History) Compact() {
 	}
 }
 
-func (h *History) Clear() {
-	h.Buf.Clear()
-}
-
 func (h *History) Prev() (line string) {
 	if h.Pos > 0 {
 		h.Pos--

cmd/cli/readline/readline.go

@@ -241,10 +241,6 @@ func (i *Instance) Readline() (string, error) {
 	}
 }
 
-func (i *Instance) HistoryEnable() {
-	i.History.Enabled = true
-}
-
 func (i *Instance) HistoryDisable() {
 	i.History.Enabled = false
 }

pkg/distribution/builder/builder.go

@@ -455,33 +455,6 @@ func (b *Builder) WithChatTemplateFile(path string) (*Builder, error) {
 	}, nil
 }
 
-// WithConfigArchive adds a config archive (tar) file to the artifact.
-func (b *Builder) WithConfigArchive(path string) (*Builder, error) {
-	// Check if config archive already exists.
-	layers, err := b.model.Layers()
-	if err != nil {
-		return nil, fmt.Errorf("get model layers: %w", err)
-	}
-
-	for _, layer := range layers {
-		mediaType, mediaTypeErr := layer.MediaType()
-		if mediaTypeErr == nil && mediaType == types.MediaTypeVLLMConfigArchive {
-			return nil, fmt.Errorf("model already has a config archive layer")
-		}
-	}
-
-	mt := b.resolveLayerMediaType(types.MediaTypeVLLMConfigArchive)
-	configLayer, err := partial.NewLayer(path, mt)
-	if err != nil {
-		return nil, fmt.Errorf("config archive layer from %q: %w", path, err)
-	}
-	return &Builder{
-		model:          mutate.AppendLayers(b.model, configLayer),
-		originalLayers: b.originalLayers,
-		outputFormat:   b.outputFormat,
-	}, nil
-}
-
 // Target represents a build target
 type Target interface {
 	Write(context.Context, types.ModelArtifact, io.Writer) error

pkg/distribution/builder/from_directory.go

@@ -10,7 +10,6 @@ import (
 
 	"github.com/docker/model-runner/pkg/distribution/files"
 	"github.com/docker/model-runner/pkg/distribution/format"
-	"github.com/docker/model-runner/pkg/distribution/internal/mutate"
 	"github.com/docker/model-runner/pkg/distribution/internal/partial"
 	"github.com/docker/model-runner/pkg/distribution/modelpack"
 	"github.com/docker/model-runner/pkg/distribution/oci"
@@ -382,21 +381,3 @@ func fileTypeToMediaType(ft files.FileType) oci.MediaType {
 		return types.MediaTypeModelFile
 	}
 }
-
-// WithFileLayer adds an individual file layer with a relative path annotation.
-// This is useful for adding files that should be extracted to a specific path.
-func (b *Builder) WithFileLayer(absPath, relPath string) (*Builder, error) {
-	// Classify the file to determine media type
-	fileType := files.Classify(absPath)
-	mediaType := fileTypeToMediaType(fileType)
-
-	layer, err := partial.NewLayerWithRelativePath(absPath, relPath, mediaType)
-	if err != nil {
-		return nil, fmt.Errorf("file layer from %q: %w", absPath, err)
-	}
-
-	return &Builder{
-		model:          mutate.AppendLayers(b.model, layer),
-		originalLayers: b.originalLayers,
-	}, nil
-}

pkg/distribution/huggingface/client.go

@@ -72,15 +72,6 @@ func WithToken(token string) ClientOption {
 	}
 }
 
-// WithTransport sets the HTTP transport for the client
-func WithTransport(transport http.RoundTripper) ClientOption {
-	return func(c *Client) {
-		if transport != nil {
-			c.httpClient.Transport = transport
-		}
-	}
-}
-
 // WithUserAgent sets the User-Agent header for requests
 func WithUserAgent(userAgent string) ClientOption {
 	return func(c *Client) {

pkg/distribution/huggingface/downloader.go

@@ -214,8 +214,3 @@ func (pr *progressReader) Read(p []byte) (n int, err error) {
 	}
 	return n, err
 }
-
-// DownloadSingleFile downloads a single file and returns its local path
-func (d *Downloader) DownloadSingleFile(ctx context.Context, file RepoFile) (string, error) {
-	return d.downloadFileWithProgress(ctx, file, uint64(file.ActualSize()), nil)
-}

pkg/distribution/internal/mutate/mutate.go

@@ -25,14 +25,3 @@ func ContextSize(mdl types.ModelArtifact, cs int32) types.ModelArtifact {
 		contextSize: &cs,
 	}
 }
-
-// SetManifestOptions applies manifest-level metadata (config media type and
-// artifact type) to the given model. The options propagate through subsequent
-// mutate wrappers via GetManifestOptions.
-func SetManifestOptions(mdl types.ModelArtifact, configMT oci.MediaType, artifactType string) types.ModelArtifact {
-	return &model{
-		base:            mdl,
-		configMediaType: configMT,
-		artifactType:    artifactType,
-	}
-}

pkg/distribution/internal/progress/reporter.go

@@ -142,15 +142,6 @@ func WriteError(w io.Writer, message string, mode oci.Mode) error {
 	})
 }
 
-// WriteWarning writes a warning message
-func WriteWarning(w io.Writer, message string, mode oci.Mode) error {
-	return write(w, oci.ProgressMessage{
-		Type:    oci.TypeWarning,
-		Message: message,
-		Mode:    mode,
-	})
-}
-
 // write writes a JSON-formatted progress message to the writer
 func write(w io.Writer, msg oci.ProgressMessage) error {
 	if w == nil {

pkg/distribution/oci/remote/remote.go

@@ -46,7 +46,6 @@ type options struct {
 	userAgent string
 	auth      authn.Authenticator
 	keychain  authn.Keychain
-	progress  chan<- oci.Update
 	plainHTTP bool
 }
 
@@ -85,13 +84,6 @@ func WithAuthFromKeychain(kc authn.Keychain) Option {
 	}
 }
 
-// WithProgress sets a channel for receiving progress updates.
-func WithProgress(ch chan<- oci.Update) Option {
-	return func(o *options) {
-		o.progress = ch
-	}
-}
-
 // WithPlainHTTP allows connecting to registries using plain HTTP instead of HTTPS.
 func WithPlainHTTP(plain bool) Option {
 	return func(o *options) {

pkg/distribution/registry/client.go

@@ -74,17 +74,6 @@ func WithUserAgent(userAgent string) ClientOption {
 	}
 }
 
-func WithAuthConfig(username, password string) ClientOption {
-	return func(c *Client) {
-		if username != "" && password != "" {
-			c.auth = &authn.Basic{
-				Username: username,
-				Password: password,
-			}
-		}
-	}
-}
-
 // WithAuth sets a custom authenticator.
 func WithAuth(auth authn.Authenticator) ClientOption {
 	return func(c *Client) {

pkg/distribution/tarball/reader.go

@@ -19,19 +19,6 @@ type Reader struct {
 	done        bool
 }
 
-type Blob struct {
-	diffID oci.Hash
-	rc     io.ReadCloser
-}
-
-func (b Blob) DiffID() (oci.Hash, error) {
-	return b.diffID, nil
-}
-
-func (b Blob) Uncompressed() (io.ReadCloser, error) {
-	return b.rc, nil
-}
-
 func (r *Reader) Next() (oci.Hash, error) {
 	for {
 		hdr, err := r.tr.Next()

pkg/envconfig/envconfig.go

@@ -17,13 +17,6 @@ func Var(key string) string {
 	return strings.Trim(strings.TrimSpace(os.Getenv(key)), "\"'")
 }
 
-// String returns a lazy string accessor for the given environment variable.
-func String(key string) func() string {
-	return func() string {
-		return Var(key)
-	}
-}
-
 // BoolWithDefault returns a lazy bool accessor for the given environment variable,
 // allowing a caller-specified default. If the variable is set but cannot be parsed
 // as a bool, the defaultValue is returned.

pkg/inference/models/api.go

@@ -27,21 +27,6 @@ type ModelPushRequest struct {
 	BearerToken string `json:"bearer-token,omitempty"`
 }
 
-// SimpleModel is a wrapper that allows creating a model with modified configuration
-type SimpleModel struct {
-	types.Model
-	ConfigValue     types.ModelConfig
-	DescriptorValue types.Descriptor
-}
-
-func (s *SimpleModel) Config() (types.ModelConfig, error) {
-	return s.ConfigValue, nil
-}
-
-func (s *SimpleModel) Descriptor() (types.Descriptor, error) {
-	return s.DescriptorValue, nil
-}
-
 // ToOpenAIList converts the model list to its OpenAI API representation. This function never
 // returns a nil slice (though it may return an empty slice).
 func ToOpenAIList(l []types.Model) (*OpenAIModelList, error) {

pkg/inference/scheduling/http_handler.go

@@ -555,12 +555,6 @@ func (h *HTTPHandler) RebuildRoutes(allowedOrigins []string) {
 	h.httpHandler = middleware.CorsMiddleware(allowedOrigins, h.router)
 }
 
-// GetLlamaCppSocket delegates to the scheduler's business logic.
-// Required by metrics.SchedulerInterface.
-func (h *HTTPHandler) GetLlamaCppSocket() (string, error) {
-	return h.scheduler.GetLlamaCppSocket()
-}
-
 // GetAllActiveRunners delegates to the scheduler's business logic.
 // Required by metrics.SchedulerInterface.
 func (h *HTTPHandler) GetAllActiveRunners() []metrics.ActiveRunner {

pkg/inference/scheduling/scheduler.go

@@ -2,7 +2,6 @@ package scheduling
 
 import (
 	"context"
-	"errors"
 	"fmt"
 	"net/http"
 	"slices"
@@ -11,7 +10,6 @@ import (
 	"github.com/docker/model-runner/pkg/distribution/types"
 	"github.com/docker/model-runner/pkg/inference"
 	"github.com/docker/model-runner/pkg/inference/backends/diffusers"
-	"github.com/docker/model-runner/pkg/inference/backends/llamacpp"
 	"github.com/docker/model-runner/pkg/inference/backends/mlx"
 	"github.com/docker/model-runner/pkg/inference/backends/sglang"
 	"github.com/docker/model-runner/pkg/inference/backends/vllm"
@@ -317,36 +315,6 @@ func (s *Scheduler) GetAllActiveRunners() []metrics.ActiveRunner {
 	return activeRunners
 }
 
-// GetLlamaCppSocket returns the Unix socket path for an active llama.cpp runner
-func (s *Scheduler) GetLlamaCppSocket() (string, error) {
-	runningBackends := s.getLoaderStatus(context.Background())
-
-	if !s.loader.lock(context.Background()) {
-		return "", errors.New("failed to acquire loader lock")
-	}
-	defer s.loader.unlock()
-
-	// Look for an active llama.cpp backend
-	for _, backend := range runningBackends {
-		if backend.BackendName == llamacpp.Name {
-			mode, ok := inference.ParseBackendMode(backend.Mode)
-			if !ok {
-				s.log.Warn("Unknown backend mode, defaulting to completion", "mode", backend.Mode)
-			}
-			// Find the runner slot for this backend/model combination
-			// We iterate through all runners since we don't know the draftModelID
-			for key, runnerInfo := range s.loader.runners {
-				if key.backend == backend.BackendName && key.modelID == backend.ModelName && key.mode == mode {
-					// Use the RunnerSocketPath function to get the socket path
-					return RunnerSocketPath(runnerInfo.slot)
-				}
-			}
-		}
-	}
-
-	return "", errors.New("no active llama.cpp backend found")
-}
-
 // ConfigureRunner configures a runner for a specific model and backend.
 // It handles all the business logic of configuration including parsing flags,
 // determining mode, selecting backend, and setting runner configuration.

pkg/metrics/aggregated_handler.go

@@ -16,6 +16,20 @@ import (
 	"github.com/prometheus/common/model"
 )
 
+// SchedulerInterface defines the methods we need from the scheduler
+type SchedulerInterface interface {
+	GetRunningBackends(w http.ResponseWriter, r *http.Request)
+	GetAllActiveRunners() []ActiveRunner
+}
+
+// ActiveRunner contains information about an active runner
+type ActiveRunner struct {
+	BackendName string
+	ModelName   string
+	Mode        string
+	Socket      string
+}
+
 // AggregatedMetricsHandler collects metrics from all active runners and aggregates them with labels
 type AggregatedMetricsHandler struct {
 	log       logging.Logger