feat: detect name/version from directory scans

Signed-off-by: Keith Zantow <kzantow@gmail.com>

syft/source/directorysource/alias/detector.go

@@ -0,0 +1,13 @@
+package alias
+
+import "github.com/anchore/syft/syft/source"
+
+// Identifier is used by certain sources (directory, file) to attempt to identify the name and version of a scan target
+type Identifier func(src source.Source) *source.Alias
+
+func DefaultIdentifiers() []Identifier {
+	return []Identifier{
+		NPMPackageAliasIdentifier,
+		MavenProjectDirIdentifier,
+	}
+}

syft/source/directorysource/alias/maven_pom_xml.go

@@ -0,0 +1,77 @@
+package alias
+
+import (
+	"encoding/xml"
+
+	"github.com/anchore/syft/internal"
+	"github.com/anchore/syft/internal/log"
+	"github.com/anchore/syft/syft/source"
+)
+
+// MavenProjectDirIdentifier augments name and version with what's found in a root pom.xml
+func MavenProjectDirIdentifier(src source.Source) *source.Alias {
+	type pomXML struct {
+		Parent  *pomXML `xml:"parent"`
+		Name    string  `xml:"name"`
+		Version string  `xml:"version"`
+	}
+
+	// it's possible older layers would have a pom.xml that gets removed,
+	// but we can probably skip identifying a directory as those
+	r, err := src.FileResolver(source.SquashedScope)
+	if err != nil {
+		log.Debugf("error getting file resolver: %v", err)
+		return nil
+	}
+
+	locs, err := r.FilesByPath("pom.xml")
+	if err != nil {
+		log.Debugf("error getting pom.xml: %v", err)
+		return nil
+	}
+
+	// if we don't have exactly 1 pom.xml in the root directory, we can't guess which is the right one to use
+	if len(locs) == 0 {
+		// expected, not found
+		return nil
+	}
+	if len(locs) > 1 {
+		log.Debugf("multiple pom.xml files found: %v", locs)
+		return nil
+	}
+
+	contents, err := r.FileContentsByLocation(locs[0])
+	if err != nil {
+		log.Tracef("error getting pom.xml contents: %v", err)
+		return nil
+	}
+	defer internal.CloseAndLogError(contents, locs[0].RealPath)
+
+	dec := xml.NewDecoder(contents)
+	project := pomXML{}
+	err = dec.Decode(&project)
+	if err != nil {
+		log.Tracef("error decoding pom.xml contents: %v", err)
+		return nil
+	}
+
+	parent := pomXML{}
+	if project.Parent != nil {
+		parent = *project.Parent
+	}
+
+	return &source.Alias{
+		Name:    project.Name,
+		Version: nonEmpty(project.Version, parent.Version),
+	}
+}
+
+// nonEmpty returns the first non-empty string provided
+func nonEmpty(values ...string) string {
+	for _, v := range values {
+		if v != "" {
+			return v
+		}
+	}
+	return ""
+}

syft/source/directorysource/alias/package_json.go

@@ -0,0 +1,59 @@
+package alias
+
+import (
+	"encoding/json"
+
+	"github.com/anchore/syft/internal"
+	"github.com/anchore/syft/internal/log"
+	"github.com/anchore/syft/syft/source"
+)
+
+// NPMPackageAliasIdentifier augments name and version with what's found in a root package.json
+func NPMPackageAliasIdentifier(src source.Source) *source.Alias {
+	type js struct {
+		Name    string `json:"name"`
+		Version string `json:"version"`
+	}
+
+	// it's possible older layers would have a package.json that gets removed,
+	// but we can probably skip identifying a directory as those
+	r, err := src.FileResolver(source.SquashedScope)
+	if err != nil {
+		log.Debugf("error getting file resolver: %v", err)
+		return nil
+	}
+	locs, err := r.FilesByPath("package.json")
+	if err != nil {
+		log.Debugf("error getting package.json: %v", err)
+		return nil
+	}
+	// if we don't have exactly 1 package.json in the root directory, we can't guess which is the right one to use
+	if len(locs) == 0 {
+		// expected, not found
+		return nil
+	}
+	if len(locs) > 1 {
+		log.Debugf("multiple package.json files found: %v", locs)
+		return nil
+	}
+
+	contents, err := r.FileContentsByLocation(locs[0])
+	if err != nil {
+		log.Tracef("error getting package.json contents: %v", err)
+		return nil
+	}
+	defer internal.CloseAndLogError(contents, locs[0].RealPath)
+
+	dec := json.NewDecoder(contents)
+	project := js{}
+	err = dec.Decode(&project)
+	if err != nil {
+		log.Tracef("error decoding package.json contents: %v", err)
+		return nil
+	}
+
+	return &source.Alias{
+		Name:    project.Name,
+		Version: project.Version,
+	}
+}

syft/source/directorysource/directory_source.go

@@ -15,6 +15,7 @@ import (
 	"github.com/anchore/syft/syft/file"
 	"github.com/anchore/syft/syft/internal/fileresolver"
 	"github.com/anchore/syft/syft/source"
+	"github.com/anchore/syft/syft/source/directorysource/alias"
 	"github.com/anchore/syft/syft/source/internal"
 )
 
@@ -25,6 +26,7 @@ type Config struct {
 	Base        string
 	Exclude     source.ExcludeConfig
 	Alias       source.Alias
+	Identifiers []alias.Identifier
 }
 
 type directorySource struct {
@@ -51,11 +53,22 @@ func New(cfg Config) (source.Source, error) {
 		return nil, fmt.Errorf("given path is not a directory: %q", cfg.Path)
 	}
 
-	return &directorySource{
-		id:     deriveIDFromDirectory(cfg),
+	src := &directorySource{
 		config: cfg,
 		mutex:  &sync.Mutex{},
-	}, nil
+	}
+
+	for _, identifier := range cfg.Identifiers {
+		id := identifier(src)
+		if !id.IsEmpty() {
+			src.config.Alias = *id
+			break
+		}
+	}
+
+	src.id = deriveIDFromDirectory(src.config)
+
+	return src, nil
 }
 
 // deriveIDFromDirectory generates an artifact ID from the given directory config. If an alias is provided, then

syft/source/directorysource/directory_source_provider.go

@@ -8,6 +8,7 @@ import (
 	"github.com/spf13/afero"
 
 	"github.com/anchore/syft/syft/source"
+	"github.com/anchore/syft/syft/source/directorysource/alias"
 )
 
 func NewSourceProvider(path string, exclude source.ExcludeConfig, alias source.Alias, basePath string) source.Provider {
@@ -52,6 +53,7 @@ func (l directorySourceProvider) Provide(_ context.Context) (source.Source, erro
 			Base:        basePath(l.basePath, location),
 			Exclude:     l.exclude,
 			Alias:       l.alias,
+			Identifiers: alias.DefaultIdentifiers(),
 		},
 	)
 }