feat: Implement better dummy fetchers and integrate into BOM generation + update integration to do dummy fetching for json and xml

Author: wiebe-vandendriessche

.github/workflows/integration.yml

@@ -12,9 +12,27 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        fixture: [ repo-10, repo-basic ]
-        format: [ json, xml ]
-        hf-mode: [ dummy, online ]
+        include:
+          # Dummy mode: only once per format (fixture doesn't matter)
+          - fixture: repo-basic
+            format: json
+            hf-mode: dummy
+          - fixture: repo-basic
+            format: xml
+            hf-mode: dummy
+          # Online mode: both fixtures, both formats
+          - fixture: repo-10
+            format: json
+            hf-mode: online
+          - fixture: repo-10
+            format: xml
+            hf-mode: online
+          - fixture: repo-basic
+            format: json
+            hf-mode: online
+          - fixture: repo-basic
+            format: xml
+            hf-mode: online
     steps:
       - name: Checkout code
         uses: actions/checkout@v4

cmd/generate.go

@@ -110,28 +110,12 @@ var generateCmd = &cobra.Command{
 
 		var discoveredBOMs []generator.DiscoveredBOM
 		if mode == "dummy" {
-			// dummy mode: per discovery: store + build (no network).
-			// TODO: wire a real dummy method for mode==dummy.
-			bomBuilder := builder.NewBOMBuilder(builder.DefaultOptions())
-			discoveredBOMs = make([]generator.DiscoveredBOM, 0, len(discoveries))
-
-			for _, d := range discoveries {
-				modelID := discoveryModelID(d)
-
-				ctx := builder.BuildContext{
-					ModelID: modelID,
-					Scan:    d,
-					HF:      nil,
-				}
-
-				bom, err := bomBuilder.Build(ctx)
-				if err != nil {
-					return err
-				}
-				discoveredBOMs = append(discoveredBOMs, generator.DiscoveredBOM{
-					Discovery: d,
-					BOM:       bom,
-				})
+			// Dummy mode: do not look at the input (Scanner discoveries)
+			// Just build one fixed dummy BOM with all fields included.
+			// Uses dummy_model_api_fetcher.go and dummy_model_readme_fetcher.go
+			discoveredBOMs, err = generator.BuildDummyBOM()
+			if err != nil {
+				return err
 			}
 		} else {
 			// Online mode: per discovery: store + fetch + map + build (inside generator).

internal/.dummy/dummy_fetcher.go

@@ -0,0 +1,55 @@
+package fetcher
+
+import (
+	"context"
+)
+
+// DummyModelAPIFetcher returns a fixed ModelAPIResponse for testing/demo purposes
+// without making any HTTP requests.
+type DummyModelAPIFetcher struct{}
+
+// Fetch returns a comprehensive dummy ModelAPIResponse with all fields populated.
+func (f *DummyModelAPIFetcher) Fetch(ctx context.Context, modelID string) (*ModelAPIResponse, error) {
+	// Create a fixed comprehensive response with all fields populated
+	gatedBool := false
+	gated := BoolOrString{Bool: &gatedBool}
+
+	return &ModelAPIResponse{
+		ID:          "dummy-org/dummy-model",
+		ModelID:     "dummy-org/dummy-model",
+		Author:      "dummy-org",
+		PipelineTag: "text-generation",
+		LibraryName: "transformers",
+		Tags: []string{
+			"pytorch",
+			"gpt2",
+			"text-generation",
+			"en",
+			"license:mit",
+			"autotrain_compatible",
+			"endpoints_compatible",
+		},
+		License:     "mit",
+		SHA:         "1234567890abcdef1234567890abcdef12345678",
+		Downloads:   1234567,
+		Likes:       890,
+		LastMod:     "2024-01-15T10:30:00.000Z",
+		CreatedAt:   "2023-06-01T08:00:00.000Z",
+		Gated:       gated,
+		Private:     false,
+		Inference:   "enabled",
+		UsedStorage: 523456789,
+		CardData: map[string]any{
+			"language": "en",
+			"license":  "mit",
+			"tags":     []string{"text-generation", "pytorch"},
+		},
+		Config: struct {
+			ModelType     string   `json:"model_type"`
+			Architectures []string `json:"architectures"`
+		}{
+			ModelType:     "gpt2",
+			Architectures: []string{"GPT2LMHeadModel"},
+		},
+	}, nil
+}