add purl types to cataloger info cmd (#4984)

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

internal/capabilities/generate/io.go

@@ -297,6 +297,11 @@ func addCatalogerFieldComment(keyNode, valueNode *yaml.Node, catalogerName strin
 		if keyNode.LineComment == "" {
 			keyNode.LineComment = autoGeneratedComment
 		}
+	case "purl_types":
+		// cataloger-level purl_types (for custom catalogers) are AUTO-GENERATED
+		if keyNode.LineComment == "" {
+			keyNode.LineComment = autoGeneratedComment
+		}
 	case "json_schema_types":
 		// json_schema_types are AUTO-GENERATED
 		if keyNode.LineComment == "" {
@@ -367,7 +372,7 @@ func addParserComments(parsersNode *yaml.Node) {
 			valueNode := parserNode.Content[i+1]
 
 			switch keyNode.Value {
-			case "parser_function", "metadata_types", "package_types", "json_schema_types":
+			case "parser_function", "metadata_types", "package_types", "purl_types", "json_schema_types":
 				// add AUTO-GENERATED comment to these fields
 				if keyNode.LineComment == "" {
 					keyNode.LineComment = autoGeneratedComment

internal/capabilities/generate/merge.go

@@ -3,6 +3,7 @@ package main
 
 import (
 	"fmt"
+	"sort"
 	"strings"
 
 	"github.com/scylladb/go-set/strset"
@@ -10,6 +11,7 @@ import (
 	"github.com/anchore/syft/internal/capabilities"
 	"github.com/anchore/syft/internal/capabilities/internal"
 	"github.com/anchore/syft/internal/packagemetadata"
+	"github.com/anchore/syft/syft/pkg"
 	"github.com/anchore/syft/syft/pkg/cataloger/binary"
 )
 
@@ -192,6 +194,9 @@ func RegenerateCapabilities(catalogerDir string, repoRoot string) (*Statistics,
 			formatOrphans(orphans))
 	}
 
+	// 5a. Derive PURL types from the discovered package types (purely a function of pkg.Type)
+	derivePURLTypes(updated)
+
 	// 6. Write back to YAML files with comments, preserving existing node trees
 	fmt.Print("  → Writing updated capabilities files...")
 	if err := saveCapabilities(catalogerDir, repoRoot, updated, existingNodes); err != nil {
@@ -427,6 +432,39 @@ func (e *EnrichmentData) GetMetadataTypes(catalogerName string) ([]string, bool)
 	return types, ok
 }
 
+// derivePURLTypes populates the PURLTypes field on every cataloger and parser by mapping each
+// observed package type to its canonical PURL type via pkg.Type.PackageURLType. this is a pure
+// function of the already-discovered package_types, so it needs no additional test observations.
+func derivePURLTypes(doc *capabilities.Document) {
+	for i := range doc.Catalogers {
+		cat := &doc.Catalogers[i]
+		cat.PURLTypes = purlTypesFromPackageTypes(cat.PackageTypes)
+		for j := range cat.Parsers {
+			cat.Parsers[j].PURLTypes = purlTypesFromPackageTypes(cat.Parsers[j].PackageTypes)
+		}
+	}
+}
+
+// purlTypesFromPackageTypes maps syft package type strings (e.g. "go-module") to their PURL types
+// (e.g. "golang"), returning a deduplicated, sorted slice. package types with no PURL mapping are skipped.
+func purlTypesFromPackageTypes(packageTypes []string) []string {
+	if len(packageTypes) == 0 {
+		return nil
+	}
+	set := strset.New()
+	for _, pt := range packageTypes {
+		if purlType := pkg.Type(pt).PackageURLType(); purlType != "" {
+			set.Add(purlType)
+		}
+	}
+	if set.IsEmpty() {
+		return nil
+	}
+	result := set.List()
+	sort.Strings(result)
+	return result
+}
+
 // GetPackageTypes returns the package types for the given cataloger name
 func (e *EnrichmentData) GetPackageTypes(catalogerName string) ([]string, bool) {
 	types, ok := e.packageTypes[catalogerName]

internal/capabilities/generate/merge_test.go

@@ -552,3 +552,48 @@ func TestConvertToJSONSchemaTypesFromMetadata(t *testing.T) {
 		})
 	}
 }
+
+func TestPURLTypesFromPackageTypes(t *testing.T) {
+	tests := []struct {
+		name         string
+		packageTypes []string
+		want         []string
+	}{
+		{
+			name:         "nil slice returns nil",
+			packageTypes: nil,
+			want:         nil,
+		},
+		{
+			name:         "single package type maps to purl type",
+			packageTypes: []string{"go-module"},
+			want:         []string{"golang"},
+		},
+		{
+			name:         "multiple package types are sorted",
+			packageTypes: []string{"npm", "go-module"},
+			want:         []string{"golang", "npm"},
+		},
+		{
+			// java and jenkins-plugin both map to maven, so the result is deduplicated
+			name:         "distinct package types collapsing to one purl type are deduplicated",
+			packageTypes: []string{"java-archive", "jenkins-plugin"},
+			want:         []string{"maven"},
+		},
+		{
+			// unknown package types have no purl mapping and are skipped
+			name:         "package type without a purl mapping is skipped",
+			packageTypes: []string{"UnknownPackage"},
+			want:         nil,
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			got := purlTypesFromPackageTypes(tt.packageTypes)
+			if diff := cmp.Diff(tt.want, got); diff != "" {
+				t.Errorf("purlTypesFromPackageTypes() mismatch (-want +got):\n%s", diff)
+			}
+		})
+	}
+}