Merge branch 'main' into improve-pull-output

# Conflicts:
#	pkg/distribution/distribution/client.go

Author: ilopezluna

README.md

@@ -402,21 +402,10 @@ in the form of [a Helm chart and static YAML](charts/docker-model-runner/README.
 If you are interested in a specific Kubernetes use-case, please start a
 discussion on the issue tracker.
 
-<<<<<<< Updated upstream
-=======
 ## dmrlet: Container Orchestrator for AI Inference
 
 dmrlet is a purpose-built container orchestrator for AI inference workloads. Unlike Kubernetes, it focuses exclusively on running stateless inference containers with zero configuration overhead. Multi-GPU mapping "just works" without YAML, device plugins, or node selectors.
 
-### Key Features
-
-| Feature | Kubernetes | dmrlet |
-|---------|------------|--------|
-| Multi-GPU setup | Device plugins + node selectors + resource limits YAML | `dmrlet serve llama3 --gpus all` |
-| Config overhead | 50+ lines of YAML minimum | Zero YAML, CLI-only |
-| Time to first inference | Minutes (pod scheduling, image pull) | Seconds (model already local) |
-| Model management | External (mount PVCs, manage yourself) | Integrated with Docker Model Runner store |
-
 ### Building dmrlet
 
 ```bash
@@ -429,91 +418,12 @@ go build -o dmrlet ./cmd/dmrlet
 
 ### Usage
 
-**Start the daemon:**
-```bash
-# Start in foreground
-dmrlet daemon
-
-# With custom socket path
-dmrlet daemon --socket /tmp/dmrlet.sock
-```
-
 **Serve a model:**
 ```bash
 # Auto-detect backend and GPUs
-dmrlet serve llama3.2
-
-# Specify backend
-dmrlet serve llama3.2 --backend vllm
-
-# Specify GPU allocation
-dmrlet serve llama3.2 --gpus 0,1
-dmrlet serve llama3.2 --gpus all
-
-# Multiple replicas
-dmrlet serve llama3.2 --replicas 2
-
-# Backend-specific options
-dmrlet serve llama3.2 --ctx-size 4096      # llama.cpp context size
-dmrlet serve llama3.2 --gpu-memory 0.8     # vLLM GPU memory utilization
-```
-
-**List running models:**
-```bash
-dmrlet ps
-# MODEL          BACKEND    REPLICAS   GPUS      ENDPOINTS              STATUS
-# llama3.2       llama.cpp  1          [0,1,2,3] localhost:30000        healthy
-```
-
-**View logs:**
-```bash
-dmrlet logs llama3.2        # Last 100 lines
-dmrlet logs llama3.2 -f     # Follow logs
-```
-
-**Scale replicas:**
-```bash
-dmrlet scale llama3.2 4     # Scale to 4 replicas
-```
-
-**Stop a model:**
-```bash
-dmrlet stop llama3.2
-dmrlet stop --all           # Stop all models
-```
-
-**Check status:**
-```bash
-dmrlet status
-# DAEMON: running
-# SOCKET: /var/run/dmrlet.sock
-#
-# GPUs:
-#   GPU 0:  NVIDIA A100 80GB  81920MB  (in use: llama3.2)
-#   GPU 1:  NVIDIA A100 80GB  81920MB  (available)
-#
-# MODELS: 1 running
-```
-
-### Supported Backends
-
-- **llama.cpp** - Default backend for GGUF models
-- **vLLM** - High-throughput serving for safetensors models
-- **SGLang** - Fast serving with RadixAttention
-
-### Architecture
-
-```
-dmrlet daemon
-  ├── GPU Manager      - Auto-detect and allocate GPUs
-  ├── Container Manager - Docker-based container lifecycle
-  ├── Service Registry  - Endpoint discovery with load balancing
-  ├── Health Monitor    - Auto-restart unhealthy containers
-  ├── Auto-scaler       - Scale based on QPS/latency/GPU utilization
-  └── Log Aggregator    - Centralized log collection
+dmrlet serve gemma3
 ```
 
->>>>>>> Stashed changes
 ## Community
 
 For general questions and discussion, please use [Docker Model Runner's Slack channel](https://dockercommunity.slack.com/archives/C09H9P5E57B).

pkg/distribution/builder/builder_test.go

@@ -10,7 +10,7 @@ import (
 	"time"
 
 	"github.com/docker/model-runner/pkg/distribution/builder"
-	"github.com/docker/model-runner/pkg/distribution/oci"
+	"github.com/docker/model-runner/pkg/distribution/internal/testutil"
 	"github.com/docker/model-runner/pkg/distribution/types"
 )
 
@@ -398,8 +398,7 @@ func TestFromModelWithAdditionalLayers(t *testing.T) {
 
 // TestFromModelErrorHandling tests that FromModel properly handles and surfaces errors from mdl.Layers()
 func TestFromModelErrorHandling(t *testing.T) {
-	// Create a mock model that fails when Layers() is called
-	mockModel := &mockFailingModel{}
+	mockModel := testutil.WithLayersError(testutil.NewGGUFArtifact(t, filepath.Join("..", "assets", "dummy.gguf")), fmt.Errorf("simulated layers error"))
 
 	// Attempt to create a builder from the failing model
 	_, err := builder.FromModel(mockModel)
@@ -424,12 +423,3 @@ func (ft *fakeTarget) Write(ctx context.Context, artifact types.ModelArtifact, w
 	ft.artifact = artifact
 	return nil
 }
-
-// mockFailingModel is a mock that fails when Layers() is called
-type mockFailingModel struct {
-	types.ModelArtifact
-}
-
-func (m *mockFailingModel) Layers() ([]oci.Layer, error) {
-	return nil, fmt.Errorf("simulated layers error")
-}

pkg/distribution/distribution/bundle_test.go

@@ -6,9 +6,7 @@ import (
 	"path/filepath"
 	"testing"
 
-	"github.com/docker/model-runner/pkg/distribution/builder"
-	"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/internal/testutil"
 	"github.com/docker/model-runner/pkg/distribution/types"
 )
 
@@ -22,12 +20,7 @@ func TestBundle(t *testing.T) {
 		t.Fatalf("Failed to create client: %v", err)
 	}
 
-	// Load dummy model from assets directory
-	b, err := builder.FromPath(filepath.Join("..", "assets", "dummy.gguf"))
-	if err != nil {
-		t.Fatalf("Failed to create model: %v", err)
-	}
-	mdl := b.Model()
+	mdl := testutil.NewGGUFArtifact(t, filepath.Join("..", "assets", "dummy.gguf"))
 	singleGGUFID, err := mdl.ID()
 	if err != nil {
 		t.Fatalf("Failed to get model ID: %v", err)
@@ -36,12 +29,11 @@ func TestBundle(t *testing.T) {
 		t.Fatalf("Failed to write model to store: %v", err)
 	}
 
-	// Load model with multi-modal projector file
-	mmprojLayer, err := partial.NewLayer(filepath.Join("..", "assets", "dummy.mmproj"), types.MediaTypeMultimodalProjector)
-	if err != nil {
-		t.Fatalf("Failed to create mmproj layer: %v", err)
-	}
-	mmprojMdl := mutate.AppendLayers(mdl, mmprojLayer)
+	mmprojMdl := testutil.NewGGUFArtifact(
+		t,
+		filepath.Join("..", "assets", "dummy.gguf"),
+		testutil.Layer(filepath.Join("..", "assets", "dummy.mmproj"), types.MediaTypeMultimodalProjector),
+	)
 	mmprojMdlID, err := mmprojMdl.ID()
 	if err != nil {
 		t.Fatalf("Failed to get model ID: %v", err)
@@ -50,12 +42,11 @@ func TestBundle(t *testing.T) {
 		t.Fatalf("Failed to write model to store: %v", err)
 	}
 
-	// Load model with template file
-	templateLayer, err := partial.NewLayer(filepath.Join("..", "assets", "template.jinja"), types.MediaTypeChatTemplate)
-	if err != nil {
-		t.Fatalf("Failed to create chat template layer: %v", err)
-	}
-	templateMdl := mutate.AppendLayers(mdl, templateLayer)
+	templateMdl := testutil.NewGGUFArtifact(
+		t,
+		filepath.Join("..", "assets", "dummy.gguf"),
+		testutil.Layer(filepath.Join("..", "assets", "template.jinja"), types.MediaTypeChatTemplate),
+	)
 	templateMdlID, err := templateMdl.ID()
 	if err != nil {
 		t.Fatalf("Failed to get model ID: %v", err)
@@ -64,12 +55,12 @@ func TestBundle(t *testing.T) {
 		t.Fatalf("Failed to write model to store: %v", err)
 	}
 
-	// Load sharded dummy model from asset directory
-	shardedB, err := builder.FromPath(filepath.Join("..", "assets", "dummy-00001-of-00002.gguf"))
-	if err != nil {
-		t.Fatalf("Failed to create model: %v", err)
-	}
-	shardedMdl := shardedB.Model()
+	shardedMdl := testutil.NewDockerArtifact(
+		t,
+		types.Config{Format: types.FormatGGUF},
+		testutil.Layer(filepath.Join("..", "assets", "dummy-00001-of-00002.gguf"), types.MediaTypeGGUF),
+		testutil.Layer(filepath.Join("..", "assets", "dummy-00002-of-00002.gguf"), types.MediaTypeGGUF),
+	)
 	shardedGGUFID, err := shardedMdl.ID()
 	if err != nil {
 		t.Fatalf("Failed to get model ID: %v", err)

pkg/distribution/distribution/client.go

@@ -8,21 +8,20 @@ import (
 	"log/slog"
 	"os"
 	"path/filepath"
-	"slices"
 	"strings"
 
 	"github.com/docker/model-runner/pkg/distribution/huggingface"
 	"github.com/docker/model-runner/pkg/distribution/internal/bundle"
 	"github.com/docker/model-runner/pkg/distribution/internal/mutate"
 	"github.com/docker/model-runner/pkg/distribution/internal/progress"
 	"github.com/docker/model-runner/pkg/distribution/internal/store"
+	"github.com/docker/model-runner/pkg/distribution/modelpack"
 	"github.com/docker/model-runner/pkg/distribution/oci"
 	"github.com/docker/model-runner/pkg/distribution/oci/authn"
 	"github.com/docker/model-runner/pkg/distribution/oci/remote"
 	"github.com/docker/model-runner/pkg/distribution/registry"
 	"github.com/docker/model-runner/pkg/distribution/tarball"
 	"github.com/docker/model-runner/pkg/distribution/types"
-	"github.com/docker/model-runner/pkg/inference/platform"
 	"github.com/docker/model-runner/pkg/internal/utils"
 )
 
@@ -774,19 +773,14 @@ func (c *Client) GetBundle(ref string) (types.ModelBundle, error) {
 	return c.store.BundleForModel(normalizedRef)
 }
 
-func GetSupportedFormats() []types.Format {
-	if platform.SupportsVLLM() {
-		return []types.Format{types.FormatGGUF, types.FormatSafetensors, types.FormatDiffusers}
-	}
-	return []types.Format{types.FormatGGUF, types.FormatDiffusers}
-}
-
 func checkCompat(image types.ModelArtifact, log *slog.Logger, reference string, progressWriter io.Writer) error {
 	manifest, err := image.Manifest()
 	if err != nil {
 		return err
 	}
-	if manifest.Config.MediaType != types.MediaTypeModelConfigV01 && manifest.Config.MediaType != types.MediaTypeModelConfigV02 {
+	if manifest.Config.MediaType != types.MediaTypeModelConfigV01 &&
+		manifest.Config.MediaType != types.MediaTypeModelConfigV02 &&
+		manifest.Config.MediaType != modelpack.MediaTypeModelConfigV1 {
 		return fmt.Errorf(
 			"config type %q is not supported (supported: %q, %q)"+
 				" - try upgrading: %w",
@@ -804,14 +798,7 @@ func checkCompat(image types.ModelArtifact, log *slog.Logger, reference string,
 	}
 
 	if config.GetFormat() == "" {
-		log.Warn("Model format field is empty for , unable to verify format compatibility", "model", utils.SanitizeForLog(reference))
-	} else if !slices.Contains(GetSupportedFormats(), config.GetFormat()) {
-		// Write warning but continue with pull
-		log.Warn(warnUnsupportedFormat)
-		if err := progress.WriteWarning(progressWriter, warnUnsupportedFormat, oci.ModePull); err != nil {
-			log.Warn("Failed to write warning message", "error", err)
-		}
-		// Don't return an error - allow the pull to continue
+		log.Warn("Model format field is empty; unable to verify format compatibility", "model", utils.SanitizeForLog(reference))
 	}
 
 	return nil