Fix for errors+failures parsing package.json

closes: #230

Signed-off-by: Toure Dunnon <toure.dunnon@anchore.com>

go.mod

@@ -8,6 +8,7 @@ require (
 	github.com/anchore/go-testutils v0.0.0-20200925183923-d5f45b0d3c04
 	github.com/anchore/go-version v1.2.2-0.20200701162849-18adb9c92b9b
 	github.com/anchore/stereoscope v0.0.0-20200925184903-c82da54e98fe
+	github.com/apex/log v1.3.0
 	github.com/bmatcuk/doublestar v1.3.1
 	github.com/docker/docker v17.12.0-ce-rc1.0.20200309214505-aa6a9891b09c+incompatible
 	github.com/dustin/go-humanize v1.0.0

syft/cataloger/javascript/parse_package_json.go

@@ -27,6 +27,7 @@ type PackageJSON struct {
 	Homepage     string            `json:"homepage"`
 	Description  string            `json:"description"`
 	Dependencies map[string]string `json:"dependencies"`
+	Repository   Repository        `json:"repository"`
 }
 
 type Author struct {
@@ -35,37 +36,40 @@ type Author struct {
 	URL   string `json:"url" mapstruct:"url"`
 }
 
+type Repository struct {
+	Type string `json:"type"`
+	URL  string `json:"url"`
+}
+
+// match example: "author": "Isaac Z. Schlueter <i@izs.me> (http://blog.izs.me)"
+// ---> name: "Isaac Z. Schlueter" email: "i@izs.me" url: "http://blog.izs.me"
 var authorPattern = regexp.MustCompile(`^\s*(?P<name>[^<(]*)(\s+<(?P<email>.*)>)?(\s\((?P<url>.*)\))?\s*$`)
 
+// This method implements the UnmarshalJSON interface to help normalize
+// the json structure.
 func (a *Author) UnmarshalJSON(b []byte) error {
 	var authorStr string
+	var fields map[string]string
+	var author Author
+
 	if err := json.Unmarshal(b, &authorStr); err != nil {
 		// string parsing did not work, assume a map was given
 		// for more information: https://docs.npmjs.com/files/package.json#people-fields-author-contributors
-		var fields map[string]string
-		var author Author
 		if err := json.Unmarshal(b, &fields); err != nil {
 			return fmt.Errorf("unable to parse package.json author: %w", err)
 		}
-		// translate the map into a structure
-		if err := mapstructure.Decode(fields, &author); err != nil {
-			return fmt.Errorf("unable to decode package.json author: %w", err)
-		}
-		*a = author
 	} else {
 		// parse out "name <email> (url)" into an Author struct
-		var fields = internal.MatchCaptureGroups(authorPattern, authorStr)
-		*a = Author{
-			Name:  fields["name"],
-			Email: fields["email"],
-			URL:   fields["url"],
-		}
+		fields = internal.MatchCaptureGroups(authorPattern, authorStr)
 	}
 
-	if a.Name == "" {
-		return fmt.Errorf("package.json author name is empty")
+	// translate the map into a structure
+	if err := mapstructure.Decode(fields, &author); err != nil {
+		return fmt.Errorf("unable to decode package.json author: %w", err)
 	}
 
+	*a = author
+
 	return nil
 }
 
@@ -102,6 +106,7 @@ func parsePackageJSON(_ string, reader io.Reader) ([]pkg.Package, error) {
 			Metadata: pkg.NpmMetadata{
 				Author:   p.Author.String(),
 				Homepage: p.Homepage,
+				URL:      p.Repository.URL,
 			},
 		})
 	}

syft/cataloger/javascript/parse_package_json_test.go

@@ -24,6 +24,7 @@ func TestParsePackageJSON(t *testing.T) {
 				Metadata: pkg.NpmMetadata{
 					Author:   "Isaac Z. Schlueter <i@izs.me> (http://blog.izs.me)",
 					Homepage: "https://docs.npmjs.com/",
+					URL:      "https://github.com/npm/cli",
 				},
 			},
 		},
@@ -38,6 +39,7 @@ func TestParsePackageJSON(t *testing.T) {
 				Metadata: pkg.NpmMetadata{
 					Author:   "Isaac Z. Schlueter <i@izs.me> (http://blog.izs.me)",
 					Homepage: "https://docs.npmjs.com/",
+					URL:      "https://github.com/npm/cli",
 				},
 			},
 		},
@@ -62,6 +64,7 @@ func TestParsePackageJSON(t *testing.T) {
 			}
 
 			for _, d := range deep.Equal(actual[0], test.ExpectedPkg) {
+
 				t.Errorf("diff: %+v", d)
 			}
 		})

syft/cataloger/javascript/test-fixtures/pkg-json/package-nested-author.json

@@ -8,5 +8,9 @@
     "email": "i@izs.me",
     "url": "http://blog.izs.me"
   },
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/npm/cli"
+  },
   "license": "Artistic-2.0"
 }

syft/pkg/metadata.go

@@ -1,11 +0,0 @@
-package pkg
-
-type MetadataType string
-
-const (
-	UnknownMetadataType MetadataType = "UnknownMetadata"
-	ApkMetadataType     MetadataType = "apk-metadata"
-	DpkgMetadataType    MetadataType = "dpkg-metadata"
-	GemgMetadataType    MetadataType = "gem-metadata"
-	RpmdbMetadataType   MetadataType = "rpmdb-metadata"
-)

syft/pkg/npm_metadata.go

@@ -9,4 +9,5 @@ type NpmMetadata struct {
 	License     string   `mapstructure:"license" json:"license"`
 	Homepage    string   `mapstructure:"homepage" json:"homepage"`
 	Description string   `mapstructure:"description" json:"description"`
+	URL         string   `mapstructure:"url" json:"url"`
 }