Update doublestar to include fix for open dirs issue (#240)

Signed-off-by: Dan Luhring <dan.luhring@anchore.com>

.github/ISSUE_TEMPLATE/bug_report.md

@@ -16,5 +16,5 @@ assignees: ''
 **Anything else we need to know?**:
 
 **Environment**:
-- syft version (use `syft version -v`):
+- Output of `syft version`:
 - OS (e.g: `cat /etc/os-release` or similar):

.github/workflows/acceptance-test.yaml

@@ -35,8 +35,11 @@ jobs:
             ${{ runner.os }}-go-${{ env.GO_VERSION }}-${{ hashFiles('**/go.sum') }}-
             ${{ runner.os }}-go-${{ env.GO_VERSION }}-
 
-      - name: Bootstrap dependencies
-        if: steps.cache.outputs.cache-hit != 'true'
+      - name: Bootstrap project dependencies
+        if: steps.bootstrap-cache.outputs.cache-hit != 'true'
+        run: make bootstrap
+
+      - name: Bootstrap CI dependencies
         run: make ci-bootstrap
 
       - name: Import GPG key

.github/workflows/codeql-analysis.yml

@@ -0,0 +1,71 @@
+# For most projects, this workflow file will not need changing; you simply need
+# to commit it to your repository.
+#
+# You may wish to alter this file to override the set of languages analyzed,
+# or to provide custom queries or build logic.
+name: "CodeQL Security Scan"
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+    # The branches below must be a subset of the branches above
+    branches: [main]
+  schedule:
+    - cron: '0 0 * * 3'
+
+jobs:
+  analyze:
+    name: Analyze
+    runs-on: ubuntu-latest
+
+    strategy:
+      fail-fast: false
+      matrix:
+        # Override automatic language detection by changing the below list
+        # Supported options are ['csharp', 'cpp', 'go', 'java', 'javascript', 'python']
+        language: ['go', 'python']
+        # Learn more...
+        # https://docs.github.com/en/github/finding-security-vulnerabilities-and-errors-in-your-code/configuring-code-scanning#overriding-automatic-language-detection
+
+    steps:
+    - name: Checkout repository
+      uses: actions/checkout@v2
+      with:
+        # We must fetch at least the immediate parents so that if this is
+        # a pull request then we can checkout the head.
+        fetch-depth: 2
+
+    # If this run was triggered by a pull request event, then checkout
+    # the head of the pull request instead of the merge commit.
+    - run: git checkout HEAD^2
+      if: ${{ github.event_name == 'pull_request' }}
+
+    # Initializes the CodeQL tools for scanning.
+    - name: Initialize CodeQL
+      uses: github/codeql-action/init@v1
+      with:
+        languages: ${{ matrix.language }}
+        # If you wish to specify custom queries, you can do so here or in a config file.
+        # By default, queries listed here will override any specified in a config file. 
+        # Prefix the list here with "+" to use these queries and those in the config file.
+        # queries: ./path/to/local/query, your-org/your-repo/queries@main
+
+    # Autobuild attempts to build any compiled languages  (C/C++, C#, or Java).
+    # If this step fails, then you should remove it and run the build manually (see below)
+    - name: Autobuild
+      uses: github/codeql-action/autobuild@v1
+
+    # ℹ️ Command-line programs to run using the OS shell.
+    # 📚 https://git.io/JvXDl
+
+    # ✏️ If the Autobuild fails above, remove it and uncomment the following three lines
+    #    and modify them (or add more) to build your code if your project
+    #    uses a compiled language
+
+    #- run: |
+    #   make bootstrap
+    #   make release
+
+    - name: Perform CodeQL Analysis
+      uses: github/codeql-action/analyze@v1

.github/workflows/release.yaml

@@ -126,7 +126,7 @@ jobs:
       - name: Build & publish release artifacts
         run: make release
         env:
-          GITHUB_TOKEN: ${{ secrets.TEMP_RELEASE_TOKEN }}
+          GITHUB_TOKEN: ${{ secrets.ANCHORE_GIT_READ_TOKEN }}
           GPG_PRIVATE_KEY: ${{ secrets.SIGNING_GPG_PRIVATE_KEY }}
           PASSPHRASE: ${{ secrets.SIGNING_GPG_PASSPHRASE }}
           SIGNING_FINGERPRINT: ${{ steps.import_gpg.outputs.fingerprint }}

.github/workflows/static-unit-integration.yaml

@@ -2,6 +2,7 @@ name: 'Static Analysis + Unit + Integration'
 on:
   workflow_dispatch:
   push:
+  pull_request:
 jobs:
 
   Static-Analysis:

Makefile

@@ -122,7 +122,7 @@ validate-cyclonedx-schema:
 .PHONY: unit
 unit: fixtures ## Run unit tests (with coverage)
 	$(call title,Running unit tests)
-	go test -coverprofile $(COVER_REPORT) ./...
+	go test -coverprofile $(COVER_REPORT) $(shell go list ./... | grep -v anchore/syft/test)
 	@go tool cover -func $(COVER_REPORT) | grep total |  awk '{print substr($$3, 1, length($$3)-1)}' > $(COVER_TOTAL)
 	@echo "Coverage: $$(cat $(COVER_TOTAL))"
 	@if [ $$(echo "$$(cat $(COVER_TOTAL)) >= $(COVERAGE_THRESHOLD)" | bc -l) -ne 1 ]; then echo "$(RED)$(BOLD)Failed coverage quality gate (> $(COVERAGE_THRESHOLD)%)$(RESET)" && false; fi
@@ -130,7 +130,9 @@ unit: fixtures ## Run unit tests (with coverage)
 .PHONY: integration
 integration: ## Run integration tests
 	$(call title,Running integration tests)
-	go test -tags=integration ./test/integration
+
+	go test -v ./test/integration
+
 
 # note: this is used by CI to determine if the integration test fixture cache (docker image tars) should be busted
 integration-fingerprint:

README.md

@@ -1,6 +1,7 @@
 # syft
 
-[![CircleCI](https://circleci.com/gh/anchore/syft.svg?style=svg)](https://circleci.com/gh/anchore/syft)
+[![Static Analysis + Unit + Integration](https://github.com/anchore/syft/workflows/Static%20Analysis%20+%20Unit%20+%20Integration/badge.svg)](https://github.com/anchore/syft/actions?query=workflow%3A%22Static+Analysis+%2B+Unit+%2B+Integration%22)
+[![Acceptance](https://github.com/anchore/syft/workflows/Acceptance/badge.svg)](https://github.com/anchore/syft/actions?query=workflow%3AAcceptance)
 [![Go Report Card](https://goreportcard.com/badge/github.com/anchore/syft)](https://goreportcard.com/report/github.com/anchore/syft)
 [![GitHub release](https://img.shields.io/github/release/anchore/syft.svg)](https://github.com/anchore/syft/releases/latest)
 [![License: Apache-2.0](https://img.shields.io/badge/License-Apache%202.0-blue.svg)](https://github.com/anchore/syft/blob/main/LICENSE)
@@ -15,7 +16,7 @@ A CLI tool and go library for generating a Software Bill of Materials (SBOM) fro
 - Linux distribution identification (supports Alpine, BusyBox, CentOS/RedHat, Debian/Ubuntu flavored distributions)
 - Supports Docker and OCI image formats
 
-> :warning: **This is pre-release software** and it may not work as expected. If you encounter an issue, please [let us know using the issue tracker](https://github.com/anchore/syft/issues).
+If you encounter an issue, please [let us know using the issue tracker](https://github.com/anchore/syft/issues).
 
 ## Getting started
 
@@ -59,7 +60,7 @@ Where the `format`s available are:
 curl -sSfL https://raw.githubusercontent.com/anchore/syft/main/install.sh | sh -s -- -b /usr/local/bin
 
 # install a specific version into a specific dir
-curl -sSfL https://raw.githubusercontent.com/anchore/syft/main/install.sh | sh -s <RELEASE_VERSION> -b <SOME_BIN_PATH>
+curl -sSfL https://raw.githubusercontent.com/anchore/syft/main/install.sh | sh -s -- -b <SOME_BIN_PATH> <RELEASE_VERSION>
 ```
 
 **macOS**
@@ -111,5 +112,4 @@ check-for-app-update: true
 ## Future plans
 
 The following areas of potential development are currently being investigated:
-- Add CycloneDX to list of output formats
 - Establish a stable interchange format w/Grype

go.mod

@@ -8,7 +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/bmatcuk/doublestar v1.3.1
+	github.com/bmatcuk/doublestar v1.3.3
 	github.com/docker/docker v17.12.0-ce-rc1.0.20200309214505-aa6a9891b09c+incompatible
 	github.com/dustin/go-humanize v1.0.0
 	github.com/go-test/deep v1.0.6

go.sum

@@ -162,6 +162,8 @@ github.com/blakesmith/ar v0.0.0-20190502131153-809d4375e1fb/go.mod h1:PkYb9DJNAw
 github.com/blang/semver v3.5.0+incompatible/go.mod h1:kRBLl5iJ+tD4TcOOxsy/0fnwebNt5EWlYSAyrTnjyyk=
 github.com/bmatcuk/doublestar v1.3.1 h1:rT8rxDPsavp9G+4ZULzqhhUSaI/OPsTZNG88Z3i0xvY=
 github.com/bmatcuk/doublestar v1.3.1/go.mod h1:wiQtGV+rzVYxB7WIlirSN++5HPtPlXEo9MEoZQC/PmE=
+github.com/bmatcuk/doublestar v1.3.3 h1:pVP1d49CcQQaNOl+PI6sPybIrIOD/6sux31PFdmhTH0=
+github.com/bmatcuk/doublestar v1.3.3/go.mod h1:wiQtGV+rzVYxB7WIlirSN++5HPtPlXEo9MEoZQC/PmE=
 github.com/bombsimon/wsl/v2 v2.0.0/go.mod h1:mf25kr/SqFEPhhcxW1+7pxzGlW+hIl/hYTKY95VwV8U=
 github.com/bombsimon/wsl/v2 v2.2.0/go.mod h1:Azh8c3XGEJl9LyX0/sFC+CKMc7Ssgua0g+6abzXN4Pg=
 github.com/bombsimon/wsl/v3 v3.0.0/go.mod h1:st10JtZYLE4D5sC7b8xV4zTKZwAQjCH/Hy2Pm1FNZIc=

internal/file/opener.go

@@ -0,0 +1,14 @@
+package file
+
+import (
+	"io"
+	"os"
+)
+
+type Opener struct {
+	path string
+}
+
+func (o Opener) Open() (io.ReadCloser, error) {
+	return os.Open(o.path)
+}

internal/file/zip_file_traversal.go

@@ -63,8 +63,8 @@ func TraverseFilesInZip(archivePath string, visitor func(*zip.File) error, paths
 	return nil
 }
 
-func ExtractFromZipToUniqueTempFile(archivePath, dir string, paths ...string) (map[string]io.Reader, error) {
-	results := make(map[string]io.Reader)
+func ExtractFromZipToUniqueTempFile(archivePath, dir string, paths ...string) (map[string]Opener, error) {
+	results := make(map[string]Opener)
 
 	// don't allow for full traversal, only select traversal from given paths
 	if len(paths) == 0 {
@@ -78,11 +78,21 @@ func ExtractFromZipToUniqueTempFile(archivePath, dir string, paths ...string) (m
 		if err != nil {
 			return fmt.Errorf("unable to create temp file: %w", err)
 		}
+		// we shouldn't try and keep the tempfile open as the returned result may have several files, which takes up
+		// resources (leading to "too many open files"). Instead we'll return a file opener to the caller which
+		// provides a ReadCloser. It is up to the caller to handle closing the file explicitly.
+		defer tempFile.Close()
 
 		zippedFile, err := file.Open()
 		if err != nil {
 			return fmt.Errorf("unable to read file=%q from zip=%q: %w", file.Name, archivePath, err)
 		}
+		defer func() {
+			err := zippedFile.Close()
+			if err != nil {
+				log.Errorf("unable to close source file=%q from zip=%q: %+v", file.Name, archivePath, err)
+			}
+		}()
 
 		if file.FileInfo().IsDir() {
 			return fmt.Errorf("unable to extract directories, only files: %s", file.Name)
@@ -103,12 +113,8 @@ func ExtractFromZipToUniqueTempFile(archivePath, dir string, paths ...string) (m
 			return fmt.Errorf("unable to reset file pointer (%s): %w", tempFile.Name(), err)
 		}
 
-		results[file.Name] = tempFile
+		results[file.Name] = Opener{path: tempFile.Name()}
 
-		err = zippedFile.Close()
-		if err != nil {
-			return fmt.Errorf("unable to close source file=%q from zip=%q: %w", file.Name, archivePath, err)
-		}
 		return nil
 	}
 

internal/parse.go

@@ -0,0 +1,15 @@
+package internal
+
+import "regexp"
+
+// MatchCaptureGroups takes a regular expression and string and returns all of the named capture group results in a map.
+func MatchCaptureGroups(regEx *regexp.Regexp, str string) map[string]string {
+	match := regEx.FindStringSubmatch(str)
+	results := make(map[string]string)
+	for i, name := range regEx.SubexpNames() {
+		if i > 0 && i <= len(match) {
+			results[name] = match[i]
+		}
+	}
+	return results
+}

schema/README.md

@@ -0,0 +1,9 @@
+## Updating the JSON schema
+Today the JSON schema is generated from integration test data. Specifically, when integration tests are run, the `/schema/json/examples` directory is populated with syft JSON output data. This examples directory is used to drive automatically generating the JSON schema.
+The caveats with this approach is:
+1) the JSON schema is only as good as the examples provided
+2) there is an integration test that ensures that the JSON schema is valid relative to what the code currently generates.
+This means to update the JSON schema you need to
+1) Open up `test/integration/json_schema_test.go` and comment out invocations of the `validateAgainstV1Schema` function.
+2) From the root of the repo run `generate-json-schema`. Now there should be a new schema generated at `/schema/json/schema.json`
+3) Uncomment the `validateAgainstV1Schema` function.

schema/json/schema.json

@@ -40,8 +40,11 @@
                             "architecture": {
                                 "type": "string"
                             },
-                            "authors": {
-                                "type": "null"
+                            "author": {
+                                "type": "string"
+                            },
+                            "authorEmail": {
+                                "type": "string"
                             },
                             "description": {
                                 "type": "string"
@@ -60,6 +63,21 @@
                                                 "checksum": {
                                                     "type": "string"
                                                 },
+                                                "digest": {
+                                                    "properties": {
+                                                        "algorithm": {
+                                                            "type": "string"
+                                                        },
+                                                        "value": {
+                                                            "type": "string"
+                                                        }
+                                                    },
+                                                    "required": [
+                                                        "algorithm",
+                                                        "value"
+                                                    ],
+                                                    "type": "object"
+                                                },
                                                 "ownerGid": {
                                                     "type": "string"
                                                 },
@@ -71,14 +89,13 @@
                                                 },
                                                 "permissions": {
                                                     "type": "string"
+                                                },
+                                                "size": {
+                                                    "type": "string"
                                                 }
                                             },
                                             "required": [
-                                                "checksum",
-                                                "ownerGid",
-                                                "ownerUid",
-                                                "path",
-                                                "permissions"
+                                                "path"
                                             ],
                                             "type": "object"
                                         }
@@ -89,6 +106,9 @@
                             "gitCommitOfApkPort": {
                                 "type": "string"
                             },
+                            "homepage": {
+                                "type": "string"
+                            },
                             "installedSize": {
                                 "type": "integer"
                             },
@@ -393,6 +413,9 @@
                                                 ],
                                                 "type": "object"
                                             },
+                                            "metadataType": {
+                                                "type": "string"
+                                            },
                                             "sources": {
                                                 "type": "null"
                                             },
@@ -409,6 +432,7 @@
                                             "licenses",
                                             "manifest",
                                             "metadata",
+                                            "metadataType",
                                             "sources",
                                             "type",
                                             "version"
@@ -417,6 +441,9 @@
                                     }
                                 ]
                             },
+                            "platform": {
+                                "type": "string"
+                            },
                             "pomProperties": {
                                 "properties": {
                                     "Path": {
@@ -457,6 +484,9 @@
                             "release": {
                                 "type": "string"
                             },
+                            "sitePackagesRootPath": {
+                                "type": "string"
+                            },
                             "size": {
                                 "type": "integer"
                             },
@@ -466,6 +496,12 @@
                             "sourceRpm": {
                                 "type": "string"
                             },
+                            "topLevelPackages": {
+                                "items": {
+                                    "type": "string"
+                                },
+                                "type": "array"
+                            },
                             "url": {
                                 "type": "string"
                             },

syft/cataloger/apkdb/parse_apk_db.go

@@ -20,9 +20,14 @@ var _ common.ParserFn = parseApkDB
 // parseApkDb parses individual packages from a given Alpine DB file. For more information on specific fields
 // see https://wiki.alpinelinux.org/wiki/Apk_spec .
 func parseApkDB(_ string, reader io.Reader) ([]pkg.Package, error) {
+	// larger capacity for the scanner.
+	const maxScannerCapacity = 1024 * 1024
+	// a new larger buffer for the scanner
+	bufScan := make([]byte, maxScannerCapacity)
 	packages := make([]pkg.Package, 0)
 
 	scanner := bufio.NewScanner(reader)
+	scanner.Buffer(bufScan, maxScannerCapacity)
 	onDoubleLF := func(data []byte, atEOF bool) (advance int, token []byte, err error) {
 		for i := 0; i < len(data); i++ {
 			if i > 0 && data[i-1] == '\n' && data[i] == '\n' {
@@ -44,11 +49,12 @@ func parseApkDB(_ string, reader io.Reader) ([]pkg.Package, error) {
 		}
 		if metadata != nil {
 			packages = append(packages, pkg.Package{
-				Name:     metadata.Package,
-				Version:  metadata.Version,
-				Licenses: strings.Split(metadata.License, " "),
-				Type:     pkg.ApkPkg,
-				Metadata: *metadata,
+				Name:         metadata.Package,
+				Version:      metadata.Version,
+				Licenses:     strings.Split(metadata.License, " "),
+				Type:         pkg.ApkPkg,
+				MetadataType: pkg.ApkMetadataType,
+				Metadata:     *metadata,
 			})
 		}
 	}

syft/cataloger/apkdb/parse_apk_db_test.go

@@ -160,10 +160,11 @@ func TestMultiplePackages(t *testing.T) {
 			fixture: "test-fixtures/multiple",
 			expected: []pkg.Package{
 				{
-					Name:     "libc-utils",
-					Version:  "0.7.2-r0",
-					Licenses: []string{"BSD"},
-					Type:     pkg.ApkPkg,
+					Name:         "libc-utils",
+					Version:      "0.7.2-r0",
+					Licenses:     []string{"BSD"},
+					Type:         pkg.ApkPkg,
+					MetadataType: pkg.ApkMetadataType,
 					Metadata: pkg.ApkMetadata{
 						Package:          "libc-utils",
 						OriginPackage:    "libc-dev",
@@ -182,10 +183,11 @@ func TestMultiplePackages(t *testing.T) {
 					},
 				},
 				{
-					Name:     "musl-utils",
-					Version:  "1.1.24-r2",
-					Licenses: []string{"MIT", "BSD", "GPL2+"},
-					Type:     pkg.ApkPkg,
+					Name:         "musl-utils",
+					Version:      "1.1.24-r2",
+					Licenses:     []string{"MIT", "BSD", "GPL2+"},
+					Type:         pkg.ApkPkg,
+					MetadataType: pkg.ApkMetadataType,
 					Metadata: pkg.ApkMetadata{
 						Package:          "musl-utils",
 						OriginPackage:    "musl",

syft/cataloger/catalog.go

@@ -1,7 +1,6 @@
 package cataloger
 
 import (
-	"github.com/anchore/stereoscope/pkg/file"
 	"github.com/anchore/syft/internal/bus"
 	"github.com/anchore/syft/internal/log"
 	"github.com/anchore/syft/syft/event"
@@ -39,36 +38,21 @@ func newMonitor() (*progress.Manual, *progress.Manual) {
 // request.
 func Catalog(resolver scope.Resolver, catalogers ...Cataloger) (*pkg.Catalog, error) {
 	catalog := pkg.NewCatalog()
-	fileSelection := make([]file.Reference, 0)
-
 	filesProcessed, packagesDiscovered := newMonitor()
 
-	// ask catalogers for files to extract from the image tar
-	for _, a := range catalogers {
-		fileSelection = append(fileSelection, a.SelectFiles(resolver)...)
-		log.Debugf("cataloger '%s' selected '%d' files", a.Name(), len(fileSelection))
-		filesProcessed.N += int64(len(fileSelection))
-	}
-
-	// fetch contents for requested selection by catalogers
-	// TODO: we should consider refactoring to return a set of io.Readers instead of the full contents themselves (allow for optional buffering).
-	contents, err := resolver.MultipleFileContentsByRef(fileSelection...)
-	if err != nil {
-		return nil, err
-	}
-
 	// perform analysis, accumulating errors for each failed analysis
 	var errs error
-	for _, a := range catalogers {
-		// TODO: check for multiple rounds of analyses by Iterate error
-		packages, err := a.Catalog(contents)
+	for _, theCataloger := range catalogers {
+		packages, err := theCataloger.Catalog(resolver)
 		if err != nil {
 			errs = multierror.Append(errs, err)
 			continue
 		}
 
-		log.Debugf("cataloger '%s' discovered '%d' packages", a.Name(), len(packages))
-		packagesDiscovered.N += int64(len(packages))
+		catalogedPackages := len(packages)
+
+		log.Debugf("cataloger '%s' discovered '%d' packages", theCataloger.Name(), catalogedPackages)
+		packagesDiscovered.N += int64(catalogedPackages)
 
 		for _, p := range packages {
 			catalog.Add(p)

syft/cataloger/cataloger.go

@@ -6,7 +6,6 @@ catalogers defined in child packages as well as the interface definition to impl
 package cataloger
 
 import (
-	"github.com/anchore/stereoscope/pkg/file"
 	"github.com/anchore/syft/syft/cataloger/apkdb"
 	"github.com/anchore/syft/syft/cataloger/deb"
 	"github.com/anchore/syft/syft/cataloger/golang"
@@ -25,20 +24,16 @@ import (
 type Cataloger interface {
 	// Name returns a string that uniquely describes a cataloger
 	Name() string
-	// SelectFiles discovers and returns specific files that the cataloger would like to inspect the contents of.
-	SelectFiles(scope.FileResolver) []file.Reference
-	// Catalog is given the file contents and should return any discovered Packages after analyzing the contents.
-	Catalog(map[file.Reference]string) ([]pkg.Package, error)
-	// TODO: add "IterationNeeded" error to indicate to the driver to continue with another Select/Catalog pass
-	// TODO: we should consider refactoring to return a set of io.Readers instead of the full contents themselves (allow for optional buffering).
+	// Catalog is given an object to resolve file references and content, this function returns any discovered Packages after analyzing the catalog source.
+	Catalog(resolver scope.Resolver) ([]pkg.Package, error)
 }
 
 // ImageCatalogers returns a slice of locally implemented catalogers that are fit for detecting installations of packages.
 func ImageCatalogers() []Cataloger {
 	return []Cataloger{
 		ruby.NewGemSpecCataloger(),
-		python.NewPythonCataloger(),         // TODO: split and replace me
-		javascript.NewJavascriptCataloger(), // TODO: split and replace me
+		python.NewPythonPackageCataloger(),
+		javascript.NewJavascriptPackageCataloger(),
 		deb.NewDpkgdbCataloger(),
 		rpmdb.NewRpmdbCataloger(),
 		java.NewJavaCataloger(),
@@ -51,8 +46,9 @@ func ImageCatalogers() []Cataloger {
 func DirectoryCatalogers() []Cataloger {
 	return []Cataloger{
 		ruby.NewGemFileLockCataloger(),
-		python.NewPythonCataloger(),         // TODO: split and replace me
-		javascript.NewJavascriptCataloger(), // TODO: split and replace me
+		python.NewPythonIndexCataloger(),
+		python.NewPythonPackageCataloger(),
+		javascript.NewJavascriptLockCataloger(),
 		deb.NewDpkgdbCataloger(),
 		rpmdb.NewRpmdbCataloger(),
 		java.NewJavaCataloger(),

syft/cataloger/common/generic_cataloger.go

@@ -34,73 +34,83 @@ func NewGenericCataloger(pathParsers map[string]ParserFn, globParsers map[string
 }
 
 // Name returns a string that uniquely describes the upstream cataloger that this Generic Cataloger represents.
-func (a *GenericCataloger) Name() string {
-	return a.upstreamCataloger
+func (c *GenericCataloger) Name() string {
+	return c.upstreamCataloger
 }
 
 // register pairs a set of file references with a parser function for future cataloging (when the file contents are resolved)
-func (a *GenericCataloger) register(files []file.Reference, parser ParserFn) {
-	a.selectedFiles = append(a.selectedFiles, files...)
+func (c *GenericCataloger) register(files []file.Reference, parser ParserFn) {
+	c.selectedFiles = append(c.selectedFiles, files...)
 	for _, f := range files {
-		a.parsers[f] = parser
+		c.parsers[f] = parser
 	}
 }
 
 // clear deletes all registered file-reference-to-parser-function pairings from former SelectFiles() and register() calls
-func (a *GenericCataloger) clear() {
-	a.selectedFiles = make([]file.Reference, 0)
-	a.parsers = make(map[file.Reference]ParserFn)
+func (c *GenericCataloger) clear() {
+	c.selectedFiles = make([]file.Reference, 0)
+	c.parsers = make(map[file.Reference]ParserFn)
+}
+
+// Catalog is given an object to resolve file references and content, this function returns any discovered Packages after analyzing the catalog source.
+func (c *GenericCataloger) Catalog(resolver scope.Resolver) ([]pkg.Package, error) {
+	fileSelection := c.selectFiles(resolver)
+	contents, err := resolver.MultipleFileContentsByRef(fileSelection...)
+	if err != nil {
+		return nil, err
+	}
+	return c.catalog(contents)
 }
 
 // SelectFiles takes a set of file trees and resolves and file references of interest for future cataloging
-func (a *GenericCataloger) SelectFiles(resolver scope.FileResolver) []file.Reference {
+func (c *GenericCataloger) selectFiles(resolver scope.FileResolver) []file.Reference {
 	// select by exact path
-	for path, parser := range a.pathParsers {
+	for path, parser := range c.pathParsers {
 		files, err := resolver.FilesByPath(file.Path(path))
 		if err != nil {
 			log.Errorf("cataloger failed to select files by path: %+v", err)
 		}
 		if files != nil {
-			a.register(files, parser)
+			c.register(files, parser)
 		}
 	}
 
 	// select by glob pattern
-	for globPattern, parser := range a.globParsers {
+	for globPattern, parser := range c.globParsers {
 		fileMatches, err := resolver.FilesByGlob(globPattern)
 		if err != nil {
 			log.Errorf("failed to find files by glob: %s", globPattern)
 		}
 		if fileMatches != nil {
-			a.register(fileMatches, parser)
+			c.register(fileMatches, parser)
 		}
 	}
 
-	return a.selectedFiles
+	return c.selectedFiles
 }
 
-// Catalog takes a set of file contents and uses any configured parser functions to resolve and return discovered packages
-func (a *GenericCataloger) Catalog(contents map[file.Reference]string) ([]pkg.Package, error) {
-	defer a.clear()
+// catalog takes a set of file contents and uses any configured parser functions to resolve and return discovered packages
+func (c *GenericCataloger) catalog(contents map[file.Reference]string) ([]pkg.Package, error) {
+	defer c.clear()
 
 	packages := make([]pkg.Package, 0)
 
-	for reference, parser := range a.parsers {
+	for reference, parser := range c.parsers {
 		content, ok := contents[reference]
 		if !ok {
-			log.Errorf("cataloger '%s' missing file content: %+v", a.upstreamCataloger, reference)
+			log.Errorf("cataloger '%s' missing file content: %+v", c.upstreamCataloger, reference)
 			continue
 		}
 
 		entries, err := parser(string(reference.Path), strings.NewReader(content))
 		if err != nil {
 			// TODO: should we fail? or only log?
-			log.Errorf("cataloger '%s' failed to parse entries (reference=%+v): %+v", a.upstreamCataloger, reference, err)
+			log.Errorf("cataloger '%s' failed to parse entries (reference=%+v): %+v", c.upstreamCataloger, reference, err)
 			continue
 		}
 
 		for _, entry := range entries {
-			entry.FoundBy = a.upstreamCataloger
+			entry.FoundBy = c.upstreamCataloger
 			entry.Source = []file.Reference{reference}
 
 			packages = append(packages, entry)

syft/cataloger/common/generic_cataloger_test.go

@@ -7,21 +7,28 @@ import (
 	"testing"
 
 	"github.com/anchore/stereoscope/pkg/file"
-	"github.com/anchore/syft/internal"
 	"github.com/anchore/syft/syft/pkg"
 )
 
-type testResolver struct {
+type testResolverMock struct {
 	contents map[file.Reference]string
 }
 
-func newTestResolver() *testResolver {
-	return &testResolver{
+func newTestResolver() *testResolverMock {
+	return &testResolverMock{
 		contents: make(map[file.Reference]string),
 	}
 }
 
-func (r *testResolver) FilesByPath(paths ...file.Path) ([]file.Reference, error) {
+func (r *testResolverMock) FileContentsByRef(_ file.Reference) (string, error) {
+	return "", fmt.Errorf("not implemented")
+}
+
+func (r *testResolverMock) MultipleFileContentsByRef(_ ...file.Reference) (map[file.Reference]string, error) {
+	return r.contents, nil
+}
+
+func (r *testResolverMock) FilesByPath(paths ...file.Path) ([]file.Reference, error) {
 	results := make([]file.Reference, len(paths))
 
 	for idx, p := range paths {
@@ -32,13 +39,17 @@ func (r *testResolver) FilesByPath(paths ...file.Path) ([]file.Reference, error)
 	return results, nil
 }
 
-func (r *testResolver) FilesByGlob(patterns ...string) ([]file.Reference, error) {
+func (r *testResolverMock) FilesByGlob(_ ...string) ([]file.Reference, error) {
 	path := "/a-path.txt"
 	ref := file.NewFileReference(file.Path(path))
 	r.contents[ref] = fmt.Sprintf("%s file contents!", path)
 	return []file.Reference{ref}, nil
 }
 
+func (r *testResolverMock) RelativeFileByPath(_ file.Reference, _ string) (*file.Reference, error) {
+	return nil, fmt.Errorf("not implemented")
+}
+
 func parser(_ string, reader io.Reader) ([]pkg.Package, error) {
 	contents, err := ioutil.ReadAll(reader)
 	if err != nil {
@@ -64,31 +75,16 @@ func TestGenericCataloger(t *testing.T) {
 	resolver := newTestResolver()
 	cataloger := NewGenericCataloger(pathParsers, globParsers, upstream)
 
-	selected := cataloger.SelectFiles(resolver)
-
-	if len(selected) != 3 {
-		t.Fatalf("unexpected selection length: %d", len(selected))
-	}
-
-	expectedSelection := internal.NewStringSetFromSlice([]string{"/last/path.txt", "/another-path.txt", "/a-path.txt"})
-	selectionByPath := make(map[string]file.Reference)
-	for _, s := range selected {
-		if !expectedSelection.Contains(string(s.Path)) {
-			t.Errorf("unexpected selection path: %+v", s.Path)
-		}
-		selectionByPath[string(s.Path)] = s
-	}
-
-	expectedPkgs := make(map[file.Reference]pkg.Package)
-	for path, ref := range selectionByPath {
-		expectedPkgs[ref] = pkg.Package{
+	expectedSelection := []string{"/last/path.txt", "/another-path.txt", "/a-path.txt"}
+	expectedPkgs := make(map[string]pkg.Package)
+	for _, path := range expectedSelection {
+		expectedPkgs[path] = pkg.Package{
 			FoundBy: upstream,
-			Source:  []file.Reference{ref},
 			Name:    fmt.Sprintf("%s file contents!", path),
 		}
 	}
 
-	actualPkgs, err := cataloger.Catalog(resolver.contents)
+	actualPkgs, err := cataloger.Catalog(resolver)
 	if err != nil {
 		t.Fatalf("cataloger catalog action failed: %+v", err)
 	}
@@ -99,7 +95,7 @@ func TestGenericCataloger(t *testing.T) {
 
 	for _, p := range actualPkgs {
 		ref := p.Source[0]
-		exP, ok := expectedPkgs[ref]
+		exP, ok := expectedPkgs[string(ref.Path)]
 		if !ok {
 			t.Errorf("missing expected pkg: ref=%+v", ref)
 			continue

syft/cataloger/deb/parse_dpkg_status.go

@@ -30,10 +30,11 @@ func parseDpkgStatus(_ string, reader io.Reader) ([]pkg.Package, error) {
 			return nil, err
 		}
 		packages = append(packages, pkg.Package{
-			Name:     entry.Package,
-			Version:  entry.Version,
-			Type:     pkg.DebPkg,
-			Metadata: entry,
+			Name:         entry.Package,
+			Version:      entry.Version,
+			Type:         pkg.DebPkg,
+			MetadataType: pkg.DpkgMetadataType,
+			Metadata:     entry,
 		})
 	}
 

syft/cataloger/java/archive_parser.go

@@ -5,6 +5,8 @@ import (
 	"io"
 	"strings"
 
+	"github.com/anchore/syft/internal/log"
+
 	"github.com/anchore/syft/internal"
 	"github.com/anchore/syft/internal/file"
 	"github.com/anchore/syft/syft/cataloger/common"
@@ -140,10 +142,11 @@ func (j *archiveParser) discoverMainPackage() (*pkg.Package, error) {
 	}
 
 	return &pkg.Package{
-		Name:     selectName(manifest, j.fileInfo),
-		Version:  selectVersion(manifest, j.fileInfo),
-		Language: pkg.Java,
-		Type:     pkg.JavaPkg,
+		Name:         selectName(manifest, j.fileInfo),
+		Version:      selectVersion(manifest, j.fileInfo),
+		Language:     pkg.Java,
+		Type:         pkg.JavaPkg,
+		MetadataType: pkg.JavaMetadataType,
 		Metadata: pkg.JavaMetadata{
 			Manifest: manifest,
 		},
@@ -175,10 +178,11 @@ func (j *archiveParser) discoverPkgsFromPomProperties(parentPkg *pkg.Package) ([
 
 				// discovered props = new package
 				p := pkg.Package{
-					Name:     propsObj.ArtifactID,
-					Version:  propsObj.Version,
-					Language: pkg.Java,
-					Type:     pkg.JavaPkg,
+					Name:         propsObj.ArtifactID,
+					Version:      propsObj.Version,
+					Language:     pkg.Java,
+					Type:         pkg.JavaPkg,
+					MetadataType: pkg.JavaMetadataType,
 					Metadata: pkg.JavaMetadata{
 						PomProperties: propsObj,
 						Parent:        parentPkg,
@@ -214,18 +218,28 @@ func (j *archiveParser) discoverPkgsFromNestedArchives(parentPkg *pkg.Package) (
 	}
 
 	// search and parse pom.properties files & fetch the contents
-	readers, err := file.ExtractFromZipToUniqueTempFile(j.archivePath, j.contentPath, j.fileManifest.GlobMatch(archiveFormatGlobs...)...)
+	openers, err := file.ExtractFromZipToUniqueTempFile(j.archivePath, j.contentPath, j.fileManifest.GlobMatch(archiveFormatGlobs...)...)
 	if err != nil {
 		return nil, fmt.Errorf("unable to extract files from zip: %w", err)
 	}
 
 	// discover nested artifacts
-	for archivePath, archiveReader := range readers {
-		nestedPath := fmt.Sprintf("%s:%s", j.virtualPath, archivePath)
-		nestedPkgs, err := parseJavaArchive(nestedPath, archiveReader)
+	for archivePath, archiveOpener := range openers {
+		archiveReadCloser, err := archiveOpener.Open()
 		if err != nil {
+			return nil, fmt.Errorf("unable to open archived file from tempdir: %w", err)
+		}
+		nestedPath := fmt.Sprintf("%s:%s", j.virtualPath, archivePath)
+		nestedPkgs, err := parseJavaArchive(nestedPath, archiveReadCloser)
+		if err != nil {
+			if closeErr := archiveReadCloser.Close(); closeErr != nil {
+				log.Errorf("unable to close archived file from tempdir: %+v", closeErr)
+			}
 			return nil, fmt.Errorf("unable to process nested java archive (%s): %w", archivePath, err)
 		}
+		if err = archiveReadCloser.Close(); err != nil {
+			return nil, fmt.Errorf("unable to close archived file from tempdir: %w", err)
+		}
 
 		// attach the parent package to all discovered packages that are not already associated with a java archive
 		for _, p := range nestedPkgs {

syft/cataloger/java/archive_parser_test.go

@@ -137,10 +137,11 @@ func TestParseJar(t *testing.T) {
 			},
 			expected: map[string]pkg.Package{
 				"example-jenkins-plugin": {
-					Name:     "example-jenkins-plugin",
-					Version:  "1.0-SNAPSHOT",
-					Language: pkg.Java,
-					Type:     pkg.JenkinsPluginPkg,
+					Name:         "example-jenkins-plugin",
+					Version:      "1.0-SNAPSHOT",
+					Language:     pkg.Java,
+					Type:         pkg.JenkinsPluginPkg,
+					MetadataType: pkg.JavaMetadataType,
 					Metadata: pkg.JavaMetadata{
 						Manifest: &pkg.JavaManifest{
 							ManifestVersion: "1.0",
@@ -181,10 +182,11 @@ func TestParseJar(t *testing.T) {
 			fixture: "test-fixtures/java-builds/packages/example-java-app-gradle-0.1.0.jar",
 			expected: map[string]pkg.Package{
 				"example-java-app-gradle": {
-					Name:     "example-java-app-gradle",
-					Version:  "0.1.0",
-					Language: pkg.Java,
-					Type:     pkg.JavaPkg,
+					Name:         "example-java-app-gradle",
+					Version:      "0.1.0",
+					Language:     pkg.Java,
+					Type:         pkg.JavaPkg,
+					MetadataType: pkg.JavaMetadataType,
 					Metadata: pkg.JavaMetadata{
 						Manifest: &pkg.JavaManifest{
 							ManifestVersion: "1.0",
@@ -200,10 +202,11 @@ func TestParseJar(t *testing.T) {
 			},
 			expected: map[string]pkg.Package{
 				"example-java-app-maven": {
-					Name:     "example-java-app-maven",
-					Version:  "0.1.0",
-					Language: pkg.Java,
-					Type:     pkg.JavaPkg,
+					Name:         "example-java-app-maven",
+					Version:      "0.1.0",
+					Language:     pkg.Java,
+					Type:         pkg.JavaPkg,
+					MetadataType: pkg.JavaMetadataType,
 					Metadata: pkg.JavaMetadata{
 						Manifest: &pkg.JavaManifest{
 							ManifestVersion: "1.0",
@@ -224,10 +227,11 @@ func TestParseJar(t *testing.T) {
 					},
 				},
 				"joda-time": {
-					Name:     "joda-time",
-					Version:  "2.9.2",
-					Language: pkg.Java,
-					Type:     pkg.JavaPkg,
+					Name:         "joda-time",
+					Version:      "2.9.2",
+					Language:     pkg.Java,
+					Type:         pkg.JavaPkg,
+					MetadataType: pkg.JavaMetadataType,
 					Metadata: pkg.JavaMetadata{
 						PomProperties: &pkg.PomProperties{
 							Path:       "META-INF/maven/joda-time/joda-time/pom.properties",

syft/cataloger/java/save_archive_to_tmp.go

@@ -35,6 +35,7 @@ func saveArchiveToTmp(reader io.Reader) (string, string, func(), error) {
 	if err != nil {
 		return contentDir, "", cleanupFn, fmt.Errorf("unable to create archive: %w", err)
 	}
+	defer archiveFile.Close()
 
 	_, err = io.Copy(archiveFile, reader)
 	if err != nil {

syft/cataloger/javascript/cataloger.go

@@ -7,12 +7,21 @@ import (
 	"github.com/anchore/syft/syft/cataloger/common"
 )
 
-// NewJavascriptCataloger returns a new JavaScript cataloger object.
-func NewJavascriptCataloger() *common.GenericCataloger {
+// NewJavascriptPackageCataloger returns a new JavaScript cataloger object based on detection of npm based packages.
+func NewJavascriptPackageCataloger() *common.GenericCataloger {
+	globParsers := map[string]common.ParserFn{
+		"**/package.json": parsePackageJSON,
+	}
+
+	return common.NewGenericCataloger(nil, globParsers, "javascript-package-cataloger")
+}
+
+// NewJavascriptLockCataloger returns a new Javascript cataloger object base on package lock files.
+func NewJavascriptLockCataloger() *common.GenericCataloger {
 	globParsers := map[string]common.ParserFn{
 		"**/package-lock.json": parsePackageLock,
 		"**/yarn.lock":         parseYarnLock,
 	}
 
-	return common.NewGenericCataloger(nil, globParsers, "javascript-cataloger")
+	return common.NewGenericCataloger(nil, globParsers, "javascript-lock-cataloger")
 }

syft/cataloger/javascript/parse_package_json.go

@@ -0,0 +1,140 @@
+package javascript
+
+import (
+	"encoding/json"
+	"fmt"
+	"io"
+	"regexp"
+
+	"github.com/anchore/syft/internal"
+
+	"github.com/mitchellh/mapstructure"
+
+	"github.com/anchore/syft/syft/cataloger/common"
+	"github.com/anchore/syft/syft/pkg"
+)
+
+// integrity check
+var _ common.ParserFn = parsePackageLock
+
+// PackageJSON represents a JavaScript package.json file
+type PackageJSON struct {
+	Version      string            `json:"version"`
+	Latest       []string          `json:"latest"`
+	Author       Author            `json:"author"`
+	License      string            `json:"license"`
+	Name         string            `json:"name"`
+	Homepage     string            `json:"homepage"`
+	Description  string            `json:"description"`
+	Dependencies map[string]string `json:"dependencies"`
+	Repository   Repository        `json:"repository"`
+}
+
+type Author struct {
+	Name  string `json:"name" mapstruct:"name"`
+	Email string `json:"email" mapstruct:"email"`
+	URL   string `json:"url" mapstruct:"url"`
+}
+
+type Repository struct {
+	Type string `json:"type" mapstructure:"type"`
+	URL  string `json:"url" mapstructure:"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*$`)
+
+// Exports Author.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
+		if err := json.Unmarshal(b, &fields); err != nil {
+			return fmt.Errorf("unable to parse package.json author: %w", err)
+		}
+	} else {
+		// parse out "name <email> (url)" into an Author struct
+		fields = internal.MatchCaptureGroups(authorPattern, authorStr)
+	}
+
+	// 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
+}
+
+func (a *Author) AuthorString() string {
+	result := a.Name
+	if a.Email != "" {
+		result += fmt.Sprintf(" <%s>", a.Email)
+	}
+	if a.URL != "" {
+		result += fmt.Sprintf(" (%s)", a.URL)
+	}
+	return result
+}
+
+func (r *Repository) UnmarshalJSON(b []byte) error {
+	var repositoryStr string
+	var fields map[string]string
+	var repository Repository
+
+	if err := json.Unmarshal(b, &repositoryStr); 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
+		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, &repository); err != nil {
+			return fmt.Errorf("unable to decode package.json author: %w", err)
+		}
+
+		*r = repository
+	} else {
+		r.URL = repositoryStr
+	}
+
+	return nil
+}
+
+// parsePackageJson parses a package.json and returns the discovered JavaScript packages.
+func parsePackageJSON(_ string, reader io.Reader) ([]pkg.Package, error) {
+	packages := make([]pkg.Package, 0)
+	dec := json.NewDecoder(reader)
+
+	for {
+		var p PackageJSON
+		if err := dec.Decode(&p); err == io.EOF {
+			break
+		} else if err != nil {
+			return nil, fmt.Errorf("failed to parse package.json file: %w", err)
+		}
+
+		packages = append(packages, pkg.Package{
+			Name:         p.Name,
+			Version:      p.Version,
+			Licenses:     []string{p.License},
+			Language:     pkg.JavaScript,
+			Type:         pkg.NpmPkg,
+			MetadataType: pkg.NpmPackageJSONMetadataType,
+			Metadata: pkg.NpmPackageJSONMetadata{
+				Author:   p.Author.AuthorString(),
+				Homepage: p.Homepage,
+				URL:      p.Repository.URL,
+				Licenses: []string{p.License},
+			},
+		})
+	}
+
+	return packages, nil
+}

syft/cataloger/javascript/parse_package_json_test.go

@@ -0,0 +1,93 @@
+package javascript
+
+import (
+	"os"
+	"testing"
+
+	"github.com/anchore/syft/syft/pkg"
+	"github.com/go-test/deep"
+)
+
+func TestParsePackageJSON(t *testing.T) {
+	tests := []struct {
+		Fixture     string
+		ExpectedPkg pkg.Package
+	}{
+		{
+			Fixture: "test-fixtures/pkg-json/package.json",
+			ExpectedPkg: pkg.Package{
+				Name:         "npm",
+				Version:      "6.14.6",
+				Type:         pkg.NpmPkg,
+				Licenses:     []string{"Artistic-2.0"},
+				Language:     pkg.JavaScript,
+				MetadataType: pkg.NpmPackageJSONMetadataType,
+				Metadata: pkg.NpmPackageJSONMetadata{
+					Author:   "Isaac Z. Schlueter <i@izs.me> (http://blog.izs.me)",
+					Homepage: "https://docs.npmjs.com/",
+					URL:      "https://github.com/npm/cli",
+					Licenses: []string{"Artistic-2.0"},
+				},
+			},
+		},
+		{
+			Fixture: "test-fixtures/pkg-json/package-nested-author.json",
+			ExpectedPkg: pkg.Package{
+				Name:         "npm",
+				Version:      "6.14.6",
+				Type:         pkg.NpmPkg,
+				Licenses:     []string{"Artistic-2.0"},
+				Language:     pkg.JavaScript,
+				MetadataType: pkg.NpmPackageJSONMetadataType,
+				Metadata: pkg.NpmPackageJSONMetadata{
+					Author:   "Isaac Z. Schlueter <i@izs.me> (http://blog.izs.me)",
+					Homepage: "https://docs.npmjs.com/",
+					URL:      "https://github.com/npm/cli",
+					Licenses: []string{"Artistic-2.0"},
+				},
+			},
+		},
+		{
+			Fixture: "test-fixtures/pkg-json/package-repo-string.json",
+			ExpectedPkg: pkg.Package{
+				Name:         "function-bind",
+				Version:      "1.1.1",
+				Type:         pkg.NpmPkg,
+				Licenses:     []string{"MIT"},
+				Language:     pkg.JavaScript,
+				MetadataType: pkg.NpmPackageJSONMetadataType,
+				Metadata: pkg.NpmPackageJSONMetadata{
+					Author:   "Raynos <raynos2@gmail.com>",
+					Homepage: "https://github.com/Raynos/function-bind",
+					URL:      "git://github.com/Raynos/function-bind.git",
+					Licenses: []string{"MIT"},
+				},
+			},
+		},
+	}
+
+	for _, test := range tests {
+		t.Run(test.Fixture, func(t *testing.T) {
+			fixture, err := os.Open(test.Fixture)
+			if err != nil {
+				t.Fatalf("failed to open fixture: %+v", err)
+			}
+
+			actual, err := parsePackageJSON(fixture.Name(), fixture)
+			if err != nil {
+				t.Fatalf("failed to parse package-lock.json: %+v", err)
+			}
+			if len(actual) != 1 {
+				for _, a := range actual {
+					t.Log("   ", a)
+				}
+				t.Fatalf("unexpected package count: %d!=1", len(actual))
+			}
+
+			for _, d := range deep.Equal(actual[0], test.ExpectedPkg) {
+
+				t.Errorf("diff: %+v", d)
+			}
+		})
+	}
+}

syft/cataloger/javascript/parse_yarn_lock.go

@@ -66,7 +66,7 @@ func parseYarnLock(_ string, reader io.Reader) ([]pkg.Package, error) {
 				Name:     currentName,
 				Version:  strings.Trim(version, "\""),
 				Language: pkg.JavaScript,
-				Type:     pkg.YarnPkg,
+				Type:     pkg.NpmPkg,
 			})
 		}
 	}

syft/cataloger/javascript/parse_yarn_lock_test.go

@@ -13,43 +13,43 @@ func TestParseYarnLock(t *testing.T) {
 			Name:     "@babel/code-frame",
 			Version:  "7.10.4",
 			Language: pkg.JavaScript,
-			Type:     pkg.YarnPkg,
+			Type:     pkg.NpmPkg,
 		},
 		"@types/minimatch": {
 			Name:     "@types/minimatch",
 			Version:  "3.0.3",
 			Language: pkg.JavaScript,
-			Type:     pkg.YarnPkg,
+			Type:     pkg.NpmPkg,
 		},
 		"@types/qs": {
 			Name:     "@types/qs",
 			Version:  "6.9.4",
 			Language: pkg.JavaScript,
-			Type:     pkg.YarnPkg,
+			Type:     pkg.NpmPkg,
 		},
 		"ajv": {
 			Name:     "ajv",
 			Version:  "6.12.3",
 			Language: pkg.JavaScript,
-			Type:     pkg.YarnPkg,
+			Type:     pkg.NpmPkg,
 		},
 		"atob": {
 			Name:     "atob",
 			Version:  "2.1.2",
 			Language: pkg.JavaScript,
-			Type:     pkg.YarnPkg,
+			Type:     pkg.NpmPkg,
 		},
 		"aws-sdk": {
 			Name:     "aws-sdk",
 			Version:  "2.706.0",
 			Language: pkg.JavaScript,
-			Type:     pkg.YarnPkg,
+			Type:     pkg.NpmPkg,
 		},
 		"jhipster-core": {
 			Name:     "jhipster-core",
 			Version:  "7.3.4",
 			Language: pkg.JavaScript,
-			Type:     pkg.YarnPkg,
+			Type:     pkg.NpmPkg,
 		},
 	}
 	fixture, err := os.Open("test-fixtures/yarn/yarn.lock")

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

@@ -0,0 +1,16 @@
+{
+  "version": "6.14.6",
+  "name": "npm",
+  "description": "a package manager for JavaScript",
+  "homepage": "https://docs.npmjs.com/",
+  "author": {
+    "name": "Isaac Z. Schlueter",
+    "email": "i@izs.me",
+    "url": "http://blog.izs.me"
+  },
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/npm/cli"
+  },
+  "license": "Artistic-2.0"
+}

syft/cataloger/javascript/test-fixtures/pkg-json/package-repo-string.json

@@ -0,0 +1,66 @@
+{
+  "name": "function-bind",
+  "version": "1.1.1",
+  "description": "Implementation of Function.prototype.bind",
+  "keywords": [
+    "function",
+    "bind",
+    "shim",
+    "es5"
+  ],
+  "author": "Raynos <raynos2@gmail.com>",
+  "repository": "git://github.com/Raynos/function-bind.git",
+  "main": "index",
+  "homepage": "https://github.com/Raynos/function-bind",
+  "contributors": [
+    {
+      "name": "Raynos"
+    },
+    {
+      "name": "Jordan Harband",
+      "url": "https://github.com/ljharb"
+    }
+  ],
+  "bugs": {
+    "url": "https://github.com/Raynos/function-bind/issues",
+    "email": "raynos2@gmail.com"
+  },
+  "dependencies": {},
+  "devDependencies": {
+    "@ljharb/eslint-config": "^12.2.1",
+    "covert": "^1.1.0",
+    "eslint": "^4.5.0",
+    "jscs": "^3.0.7",
+    "tape": "^4.8.0"
+  },
+  "license": "MIT",
+  "scripts": {
+    "pretest": "npm run lint",
+    "test": "npm run tests-only",
+    "posttest": "npm run coverage -- --quiet",
+    "tests-only": "node test",
+    "coverage": "covert test/*.js",
+    "lint": "npm run jscs && npm run eslint",
+    "jscs": "jscs *.js */*.js",
+    "eslint": "eslint *.js */*.js"
+  },
+  "testling": {
+    "files": "test/index.js",
+    "browsers": [
+      "ie/8..latest",
+      "firefox/16..latest",
+      "firefox/nightly",
+      "chrome/22..latest",
+      "chrome/canary",
+      "opera/12..latest",
+      "opera/next",
+      "safari/5.1..latest",
+      "ipad/6.0..latest",
+      "iphone/6.0..latest",
+      "android-browser/4.2..latest"
+    ]
+  }
+,"_resolved": "https://registry.npmjs.org/function-bind/-/function-bind-1.1.1.tgz"
+,"_integrity": "sha512-yIovAzMX49sF8Yl58fSCWJ5svSLuaibPxXQJFLmBObTuCr0Mf1KiPopGM9NiFjiYBCbfaa2Fh6breQ6ANVTI0A=="
+,"_from": "function-bind@1.1.1"
+}

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

@@ -0,0 +1,314 @@
+{
+  "version": "6.14.6",
+  "name": "npm",
+  "description": "a package manager for JavaScript",
+  "keywords": [
+    "install",
+    "modules",
+    "package manager",
+    "package.json"
+  ],
+  "preferGlobal": true,
+  "config": {
+    "publishtest": false
+  },
+  "homepage": "https://docs.npmjs.com/",
+  "author": "Isaac Z. Schlueter <i@izs.me> (http://blog.izs.me)",
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/npm/cli"
+  },
+  "bugs": {
+    "url": "https://npm.community/c/bugs"
+  },
+  "directories": {
+    "bin": "./bin",
+    "doc": "./doc",
+    "lib": "./lib",
+    "man": "./man"
+  },
+  "main": "./lib/npm.js",
+  "bin": {
+    "npm": "./bin/npm-cli.js",
+    "npx": "./bin/npx-cli.js"
+  },
+  "dependencies": {
+    "JSONStream": "^1.3.5",
+    "abbrev": "~1.1.1",
+    "ansicolors": "~0.3.2",
+    "ansistyles": "~0.1.3",
+    "aproba": "^2.0.0",
+    "archy": "~1.0.0",
+    "bin-links": "^1.1.7",
+    "bluebird": "^3.5.5",
+    "byte-size": "^5.0.1",
+    "cacache": "^12.0.3",
+    "call-limit": "^1.1.1",
+    "chownr": "^1.1.4",
+    "ci-info": "^2.0.0",
+    "cli-columns": "^3.1.2",
+    "cli-table3": "^0.5.1",
+    "cmd-shim": "^3.0.3",
+    "columnify": "~1.5.4",
+    "config-chain": "^1.1.12",
+    "detect-indent": "~5.0.0",
+    "detect-newline": "^2.1.0",
+    "dezalgo": "~1.0.3",
+    "editor": "~1.0.0",
+    "figgy-pudding": "^3.5.1",
+    "find-npm-prefix": "^1.0.2",
+    "fs-vacuum": "~1.2.10",
+    "fs-write-stream-atomic": "~1.0.10",
+    "gentle-fs": "^2.3.0",
+    "glob": "^7.1.6",
+    "graceful-fs": "^4.2.4",
+    "has-unicode": "~2.0.1",
+    "hosted-git-info": "^2.8.8",
+    "iferr": "^1.0.2",
+    "infer-owner": "^1.0.4",
+    "inflight": "~1.0.6",
+    "inherits": "^2.0.4",
+    "ini": "^1.3.5",
+    "init-package-json": "^1.10.3",
+    "is-cidr": "^3.0.0",
+    "json-parse-better-errors": "^1.0.2",
+    "lazy-property": "~1.0.0",
+    "libcipm": "^4.0.7",
+    "libnpm": "^3.0.1",
+    "libnpmaccess": "^3.0.2",
+    "libnpmhook": "^5.0.3",
+    "libnpmorg": "^1.0.1",
+    "libnpmsearch": "^2.0.2",
+    "libnpmteam": "^1.0.2",
+    "libnpx": "^10.2.2",
+    "lock-verify": "^2.1.0",
+    "lockfile": "^1.0.4",
+    "lodash._baseuniq": "~4.6.0",
+    "lodash.clonedeep": "~4.5.0",
+    "lodash.union": "~4.6.0",
+    "lodash.uniq": "~4.5.0",
+    "lodash.without": "~4.4.0",
+    "lru-cache": "^5.1.1",
+    "meant": "~1.0.1",
+    "mississippi": "^3.0.0",
+    "mkdirp": "^0.5.5",
+    "move-concurrently": "^1.0.1",
+    "node-gyp": "^5.1.0",
+    "nopt": "^4.0.3",
+    "normalize-package-data": "^2.5.0",
+    "npm-audit-report": "^1.3.2",
+    "npm-cache-filename": "~1.0.2",
+    "npm-install-checks": "^3.0.2",
+    "npm-lifecycle": "^3.1.4",
+    "npm-package-arg": "^6.1.1",
+    "npm-packlist": "^1.4.8",
+    "npm-pick-manifest": "^3.0.2",
+    "npm-profile": "^4.0.4",
+    "npm-registry-fetch": "^4.0.5",
+    "npm-user-validate": "~1.0.0",
+    "npmlog": "~4.1.2",
+    "once": "~1.4.0",
+    "opener": "^1.5.1",
+    "osenv": "^0.1.5",
+    "pacote": "^9.5.12",
+    "path-is-inside": "~1.0.2",
+    "promise-inflight": "~1.0.1",
+    "qrcode-terminal": "^0.12.0",
+    "query-string": "^6.8.2",
+    "qw": "~1.0.1",
+    "read": "~1.0.7",
+    "read-cmd-shim": "^1.0.5",
+    "read-installed": "~4.0.3",
+    "read-package-json": "^2.1.1",
+    "read-package-tree": "^5.3.1",
+    "readable-stream": "^3.6.0",
+    "readdir-scoped-modules": "^1.1.0",
+    "request": "^2.88.0",
+    "retry": "^0.12.0",
+    "rimraf": "^2.7.1",
+    "safe-buffer": "^5.1.2",
+    "semver": "^5.7.1",
+    "sha": "^3.0.0",
+    "slide": "~1.1.6",
+    "sorted-object": "~2.0.1",
+    "sorted-union-stream": "~2.1.3",
+    "ssri": "^6.0.1",
+    "stringify-package": "^1.0.1",
+    "tar": "^4.4.13",
+    "text-table": "~0.2.0",
+    "tiny-relative-date": "^1.3.0",
+    "uid-number": "0.0.6",
+    "umask": "~1.1.0",
+    "unique-filename": "^1.1.1",
+    "unpipe": "~1.0.0",
+    "update-notifier": "^2.5.0",
+    "uuid": "^3.3.3",
+    "validate-npm-package-license": "^3.0.4",
+    "validate-npm-package-name": "~3.0.0",
+    "which": "^1.3.1",
+    "worker-farm": "^1.7.0",
+    "write-file-atomic": "^2.4.3"
+  },
+  "bundleDependencies": [
+    "abbrev",
+    "ansicolors",
+    "ansistyles",
+    "aproba",
+    "archy",
+    "bin-links",
+    "bluebird",
+    "byte-size",
+    "cacache",
+    "call-limit",
+    "chownr",
+    "ci-info",
+    "cli-columns",
+    "cli-table3",
+    "cmd-shim",
+    "columnify",
+    "config-chain",
+    "debuglog",
+    "detect-indent",
+    "detect-newline",
+    "dezalgo",
+    "editor",
+    "figgy-pudding",
+    "find-npm-prefix",
+    "fs-vacuum",
+    "fs-write-stream-atomic",
+    "gentle-fs",
+    "glob",
+    "graceful-fs",
+    "has-unicode",
+    "hosted-git-info",
+    "iferr",
+    "imurmurhash",
+    "infer-owner",
+    "inflight",
+    "inherits",
+    "ini",
+    "init-package-json",
+    "is-cidr",
+    "json-parse-better-errors",
+    "JSONStream",
+    "lazy-property",
+    "libcipm",
+    "libnpm",
+    "libnpmaccess",
+    "libnpmhook",
+    "libnpmorg",
+    "libnpmsearch",
+    "libnpmteam",
+    "libnpx",
+    "lock-verify",
+    "lockfile",
+    "lodash._baseindexof",
+    "lodash._baseuniq",
+    "lodash._bindcallback",
+    "lodash._cacheindexof",
+    "lodash._createcache",
+    "lodash._getnative",
+    "lodash.clonedeep",
+    "lodash.restparam",
+    "lodash.union",
+    "lodash.uniq",
+    "lodash.without",
+    "lru-cache",
+    "meant",
+    "mississippi",
+    "mkdirp",
+    "move-concurrently",
+    "node-gyp",
+    "nopt",
+    "normalize-package-data",
+    "npm-audit-report",
+    "npm-cache-filename",
+    "npm-install-checks",
+    "npm-lifecycle",
+    "npm-package-arg",
+    "npm-packlist",
+    "npm-pick-manifest",
+    "npm-profile",
+    "npm-registry-fetch",
+    "npm-user-validate",
+    "npmlog",
+    "once",
+    "opener",
+    "osenv",
+    "pacote",
+    "path-is-inside",
+    "promise-inflight",
+    "qrcode-terminal",
+    "query-string",
+    "qw",
+    "read-cmd-shim",
+    "read-installed",
+    "read-package-json",
+    "read-package-tree",
+    "read",
+    "readable-stream",
+    "readdir-scoped-modules",
+    "request",
+    "retry",
+    "rimraf",
+    "safe-buffer",
+    "semver",
+    "sha",
+    "slide",
+    "sorted-object",
+    "sorted-union-stream",
+    "ssri",
+    "stringify-package",
+    "tar",
+    "text-table",
+    "tiny-relative-date",
+    "uid-number",
+    "umask",
+    "unique-filename",
+    "unpipe",
+    "update-notifier",
+    "uuid",
+    "validate-npm-package-license",
+    "validate-npm-package-name",
+    "which",
+    "worker-farm",
+    "write-file-atomic"
+  ],
+  "devDependencies": {
+    "deep-equal": "^1.0.1",
+    "get-stream": "^4.1.0",
+    "licensee": "^7.0.3",
+    "marked": "^0.6.3",
+    "marked-man": "^0.6.0",
+    "npm-registry-couchapp": "^2.7.4",
+    "npm-registry-mock": "^1.3.1",
+    "require-inject": "^1.4.4",
+    "sprintf-js": "^1.1.2",
+    "standard": "^11.0.1",
+    "tacks": "^1.3.0",
+    "tap": "^12.7.0",
+    "tar-stream": "^2.1.0"
+  },
+  "scripts": {
+    "dumpconf": "env | grep npm | sort | uniq",
+    "prepare": "node bin/npm-cli.js rebuild && node bin/npm-cli.js --no-audit --no-timing prune --prefix=. --no-global && rimraf test/*/*/node_modules && make -j4 mandocs",
+    "preversion": "bash scripts/update-authors.sh && git add AUTHORS && git commit -m \"update AUTHORS\" || true",
+    "licenses": "licensee --production --errors-only",
+    "tap": "tap -J --timeout 300 --no-esm",
+    "tap-cover": "tap -J --nyc-arg=--cache --coverage --timeout 600 --no-esm",
+    "lint": "standard",
+    "pretest": "npm run lint",
+    "test": "npm run test-tap --",
+    "test:nocleanup": "NO_TEST_CLEANUP=1 npm run test --",
+    "sudotest": "sudo npm run tap -- \"test/tap/*.js\"",
+    "sudotest:nocleanup": "sudo NO_TEST_CLEANUP=1 npm run tap -- \"test/tap/*.js\"",
+    "posttest": "rimraf test/npm_cache*",
+    "test-coverage": "npm run tap-cover -- \"test/tap/*.js\" \"test/network/*.js\"",
+    "test-tap": "npm run tap -- \"test/tap/*.js\" \"test/network/*.js\"",
+    "test-node": "tap --timeout 240 \"test/tap/*.js\" \"test/network/*.js\""
+  },
+  "license": "Artistic-2.0",
+  "engines": {
+    "node": "6 >=6.2.0 || 8 || >=9.3.0"
+  }
+}

syft/cataloger/python/cataloger.go

@@ -1,21 +0,0 @@
-/*
-Package python provides a concrete Cataloger implementation for Python ecosystem files (egg, wheel, requirements.txt).
-*/
-package python
-
-import (
-	"github.com/anchore/syft/syft/cataloger/common"
-)
-
-// NewPythonCataloger returns a new Python cataloger object.
-func NewPythonCataloger() *common.GenericCataloger {
-	globParsers := map[string]common.ParserFn{
-		"**/*egg-info/PKG-INFO":  parseEggMetadata,
-		"**/*dist-info/METADATA": parseWheelMetadata,
-		"**/*requirements*.txt":  parseRequirementsTxt,
-		"**/poetry.lock":         parsePoetryLock,
-		"**/setup.py":            parseSetup,
-	}
-
-	return common.NewGenericCataloger(nil, globParsers, "python-cataloger")
-}

syft/cataloger/python/index_cataloger.go

@@ -0,0 +1,19 @@
+/*
+Package python provides a concrete Cataloger implementation for Python ecosystem files (egg, wheel, requirements.txt).
+*/
+package python
+
+import (
+	"github.com/anchore/syft/syft/cataloger/common"
+)
+
+// NewPythonIndexCataloger returns a new cataloger for python packages referenced from poetry lock files, requirements.txt files, and setup.py files.
+func NewPythonIndexCataloger() *common.GenericCataloger {
+	globParsers := map[string]common.ParserFn{
+		"**/*requirements*.txt": parseRequirementsTxt,
+		"**/poetry.lock":        parsePoetryLock,
+		"**/setup.py":           parseSetup,
+	}
+
+	return common.NewGenericCataloger(nil, globParsers, "python-index-cataloger")
+}

syft/cataloger/python/package_cataloger.go

@@ -0,0 +1,179 @@
+package python
+
+import (
+	"bufio"
+	"fmt"
+	"path/filepath"
+	"strings"
+
+	"github.com/anchore/stereoscope/pkg/file"
+
+	"github.com/anchore/syft/syft/pkg"
+
+	"github.com/anchore/syft/syft/scope"
+)
+
+const (
+	eggMetadataGlob   = "**/*egg-info/PKG-INFO"
+	wheelMetadataGlob = "**/*dist-info/METADATA"
+)
+
+type PackageCataloger struct{}
+
+// NewPythonPackageCataloger returns a new cataloger for python packages within egg or wheel installation directories.
+func NewPythonPackageCataloger() *PackageCataloger {
+	return &PackageCataloger{}
+}
+
+// Name returns a string that uniquely describes a cataloger
+func (c *PackageCataloger) Name() string {
+	return "python-package-cataloger"
+}
+
+// Catalog is given an object to resolve file references and content, this function returns any discovered Packages after analyzing python egg and wheel installations.
+func (c *PackageCataloger) Catalog(resolver scope.Resolver) ([]pkg.Package, error) {
+	// nolint:prealloc
+	var fileMatches []file.Reference
+
+	for _, glob := range []string{eggMetadataGlob, wheelMetadataGlob} {
+		matches, err := resolver.FilesByGlob(glob)
+		if err != nil {
+			return nil, fmt.Errorf("failed to find files by glob: %s", glob)
+		}
+		fileMatches = append(fileMatches, matches...)
+	}
+
+	var pkgs []pkg.Package
+	for _, ref := range fileMatches {
+		p, err := c.catalogEggOrWheel(resolver, ref)
+		if err != nil {
+			return nil, fmt.Errorf("unable to catalog python package=%+v: %w", ref.Path, err)
+		}
+		if p != nil {
+			pkgs = append(pkgs, *p)
+		}
+	}
+	return pkgs, nil
+}
+
+// catalogEggOrWheel takes the primary metadata file reference and returns the python package it represents.
+func (c *PackageCataloger) catalogEggOrWheel(resolver scope.Resolver, metadataRef file.Reference) (*pkg.Package, error) {
+	metadata, sources, err := c.assembleEggOrWheelMetadata(resolver, metadataRef)
+	if err != nil {
+		return nil, err
+	}
+
+	var licenses []string
+	if metadata.License != "" {
+		licenses = []string{metadata.License}
+	}
+
+	return &pkg.Package{
+		Name:         metadata.Name,
+		Version:      metadata.Version,
+		FoundBy:      c.Name(),
+		Source:       sources,
+		Licenses:     licenses,
+		Language:     pkg.Python,
+		Type:         pkg.PythonPkg,
+		MetadataType: pkg.PythonPackageMetadataType,
+		Metadata:     *metadata,
+	}, nil
+}
+
+// fetchRecordFiles finds a corresponding RECORD file for the given python package metadata file and returns the set of file records contained.
+func (c *PackageCataloger) fetchRecordFiles(resolver scope.Resolver, metadataRef file.Reference) (files []pkg.PythonFileRecord, sources []file.Reference, err error) {
+	// we've been given a file reference to a specific wheel METADATA file. note: this may be for a directory
+	// or for an image... for an image the METADATA file may be present within multiple layers, so it is important
+	// to reconcile the RECORD path to the same layer (or the next adjacent lower layer).
+
+	// lets find the RECORD file relative to the directory where the METADATA file resides (in path AND layer structure)
+	recordPath := filepath.Join(filepath.Dir(string(metadataRef.Path)), "RECORD")
+	recordRef, err := resolver.RelativeFileByPath(metadataRef, recordPath)
+	if err != nil {
+		return nil, nil, err
+	}
+
+	if recordRef != nil {
+		sources = append(sources, *recordRef)
+
+		recordContents, err := resolver.FileContentsByRef(*recordRef)
+		if err != nil {
+			return nil, nil, err
+		}
+
+		// parse the record contents
+		records, err := parseWheelOrEggRecord(strings.NewReader(recordContents))
+		if err != nil {
+			return nil, nil, err
+		}
+
+		files = append(files, records...)
+	}
+	return files, sources, nil
+}
+
+// fetchTopLevelPackages finds a corresponding top_level.txt file for the given python package metadata file and returns the set of package names contained.
+func (c *PackageCataloger) fetchTopLevelPackages(resolver scope.Resolver, metadataRef file.Reference) (pkgs []string, sources []file.Reference, err error) {
+	// a top_level.txt file specifies the python top-level packages (provided by this python package) installed into site-packages
+	parentDir := filepath.Dir(string(metadataRef.Path))
+	topLevelPath := filepath.Join(parentDir, "top_level.txt")
+	topLevelRef, err := resolver.RelativeFileByPath(metadataRef, topLevelPath)
+	if err != nil {
+		return nil, nil, err
+	}
+	if topLevelRef == nil {
+		return nil, nil, fmt.Errorf("missing python package top_level.txt (package=%q)", string(metadataRef.Path))
+	}
+
+	sources = append(sources, *topLevelRef)
+
+	topLevelContents, err := resolver.FileContentsByRef(*topLevelRef)
+	if err != nil {
+		return nil, nil, err
+	}
+
+	scanner := bufio.NewScanner(strings.NewReader(topLevelContents))
+	for scanner.Scan() {
+		pkgs = append(pkgs, scanner.Text())
+	}
+
+	if err := scanner.Err(); err != nil {
+		return nil, nil, fmt.Errorf("could not read python package top_level.txt: %w", err)
+	}
+
+	return pkgs, sources, nil
+}
+
+// assembleEggOrWheelMetadata discovers and accumulates python package metadata from multiple file sources and returns a single metadata object as well as a list of files where the metadata was derived from.
+func (c *PackageCataloger) assembleEggOrWheelMetadata(resolver scope.Resolver, metadataRef file.Reference) (*pkg.PythonPackageMetadata, []file.Reference, error) {
+	var sources = []file.Reference{metadataRef}
+
+	metadataContents, err := resolver.FileContentsByRef(metadataRef)
+	if err != nil {
+		return nil, nil, err
+	}
+
+	metadata, err := parseWheelOrEggMetadata(metadataRef.Path, strings.NewReader(metadataContents))
+	if err != nil {
+		return nil, nil, err
+	}
+
+	// attach any python files found for the given wheel/egg installation
+	r, s, err := c.fetchRecordFiles(resolver, metadataRef)
+	if err != nil {
+		return nil, nil, err
+	}
+	sources = append(sources, s...)
+	metadata.Files = r
+
+	// attach any top-level package names found for the given wheel/egg installation
+	p, s, err := c.fetchTopLevelPackages(resolver, metadataRef)
+	if err != nil {
+		return nil, nil, err
+	}
+	sources = append(sources, s...)
+	metadata.TopLevelPackages = p
+
+	return &metadata, sources, nil
+}

syft/cataloger/python/package_cataloger_test.go

@@ -0,0 +1,241 @@
+package python
+
+import (
+	"fmt"
+	"io"
+	"io/ioutil"
+	"os"
+	"strings"
+	"testing"
+
+	"github.com/anchore/stereoscope/pkg/file"
+
+	"github.com/anchore/syft/syft/pkg"
+	"github.com/go-test/deep"
+)
+
+type pythonTestResolverMock struct {
+	metadataReader io.Reader
+	recordReader   io.Reader
+	topLevelReader io.Reader
+	metadataRef    *file.Reference
+	recordRef      *file.Reference
+	topLevelRef    *file.Reference
+	contents       map[file.Reference]string
+}
+
+func newTestResolver(metaPath, recordPath, topPath string) *pythonTestResolverMock {
+	metadataReader, err := os.Open(metaPath)
+	if err != nil {
+		panic(fmt.Errorf("failed to open metadata: %+v", err))
+	}
+
+	var recordReader io.Reader
+	if recordPath != "" {
+		recordReader, err = os.Open(recordPath)
+		if err != nil {
+			panic(fmt.Errorf("failed to open record: %+v", err))
+		}
+	}
+
+	var topLevelReader io.Reader
+	if topPath != "" {
+		topLevelReader, err = os.Open(topPath)
+		if err != nil {
+			panic(fmt.Errorf("failed to open top level: %+v", err))
+		}
+	}
+
+	var recordRef *file.Reference
+	if recordReader != nil {
+		ref := file.NewFileReference("test-fixtures/dist-info/RECORD")
+		recordRef = &ref
+	}
+	var topLevelRef *file.Reference
+	if topLevelReader != nil {
+		ref := file.NewFileReference("test-fixtures/dist-info/top_level.txt")
+		topLevelRef = &ref
+	}
+	metadataRef := file.NewFileReference("test-fixtures/dist-info/METADATA")
+	return &pythonTestResolverMock{
+		recordReader:   recordReader,
+		metadataReader: metadataReader,
+		topLevelReader: topLevelReader,
+		metadataRef:    &metadataRef,
+		recordRef:      recordRef,
+		topLevelRef:    topLevelRef,
+		contents:       make(map[file.Reference]string),
+	}
+}
+
+func (r *pythonTestResolverMock) FileContentsByRef(ref file.Reference) (string, error) {
+	switch ref.Path {
+	case r.topLevelRef.Path:
+		b, err := ioutil.ReadAll(r.topLevelReader)
+		if err != nil {
+			return "", err
+		}
+		return string(b), nil
+	case r.metadataRef.Path:
+		b, err := ioutil.ReadAll(r.metadataReader)
+		if err != nil {
+			return "", err
+		}
+		return string(b), nil
+	case r.recordRef.Path:
+		b, err := ioutil.ReadAll(r.recordReader)
+		if err != nil {
+			return "", err
+		}
+		return string(b), nil
+	}
+	return "", fmt.Errorf("invalid value given")
+}
+
+func (r *pythonTestResolverMock) MultipleFileContentsByRef(_ ...file.Reference) (map[file.Reference]string, error) {
+	return nil, fmt.Errorf("not implemented")
+}
+
+func (r *pythonTestResolverMock) FilesByPath(_ ...file.Path) ([]file.Reference, error) {
+	return nil, fmt.Errorf("not implemented")
+}
+
+func (r *pythonTestResolverMock) FilesByGlob(_ ...string) ([]file.Reference, error) {
+	return nil, fmt.Errorf("not implemented")
+}
+func (r *pythonTestResolverMock) RelativeFileByPath(_ file.Reference, path string) (*file.Reference, error) {
+	switch {
+	case strings.Contains(path, "RECORD"):
+		return r.recordRef, nil
+	case strings.Contains(path, "top_level.txt"):
+		return r.topLevelRef, nil
+	default:
+		return nil, fmt.Errorf("invalid RelativeFileByPath value given: %q", path)
+	}
+}
+
+func TestPythonPackageWheelCataloger(t *testing.T) {
+	tests := []struct {
+		MetadataFixture string
+		RecordFixture   string
+		TopLevelFixture string
+		ExpectedPackage pkg.Package
+	}{
+		{
+			MetadataFixture: "test-fixtures/egg-info/PKG-INFO",
+			RecordFixture:   "test-fixtures/egg-info/RECORD",
+			TopLevelFixture: "test-fixtures/egg-info/top_level.txt",
+			ExpectedPackage: pkg.Package{
+				Name:         "requests",
+				Version:      "2.22.0",
+				Type:         pkg.PythonPkg,
+				Language:     pkg.Python,
+				Licenses:     []string{"Apache 2.0"},
+				FoundBy:      "python-package-cataloger",
+				MetadataType: pkg.PythonPackageMetadataType,
+				Metadata: pkg.PythonPackageMetadata{
+					Name:                 "requests",
+					Version:              "2.22.0",
+					License:              "Apache 2.0",
+					Platform:             "UNKNOWN",
+					Author:               "Kenneth Reitz",
+					AuthorEmail:          "me@kennethreitz.org",
+					SitePackagesRootPath: "test-fixtures",
+					Files: []pkg.PythonFileRecord{
+						{Path: "requests-2.22.0.dist-info/INSTALLER", Digest: &pkg.Digest{"sha256", "zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg"}, Size: "4"},
+						{Path: "requests/__init__.py", Digest: &pkg.Digest{"sha256", "PnKCgjcTq44LaAMzB-7--B2FdewRrE8F_vjZeaG9NhA"}, Size: "3921"},
+						{Path: "requests/__pycache__/__version__.cpython-38.pyc"},
+						{Path: "requests/__pycache__/utils.cpython-38.pyc"},
+						{Path: "requests/__version__.py", Digest: &pkg.Digest{"sha256", "Bm-GFstQaFezsFlnmEMrJDe8JNROz9n2XXYtODdvjjc"}, Size: "436"},
+						{Path: "requests/utils.py", Digest: &pkg.Digest{"sha256", "LtPJ1db6mJff2TJSJWKi7rBpzjPS3mSOrjC9zRhoD3A"}, Size: "30049"},
+					},
+					TopLevelPackages: []string{"requests"},
+				},
+			},
+		},
+		{
+			MetadataFixture: "test-fixtures/dist-info/METADATA",
+			RecordFixture:   "test-fixtures/dist-info/RECORD",
+			TopLevelFixture: "test-fixtures/dist-info/top_level.txt",
+			ExpectedPackage: pkg.Package{
+				Name:         "Pygments",
+				Version:      "2.6.1",
+				Type:         pkg.PythonPkg,
+				Language:     pkg.Python,
+				Licenses:     []string{"BSD License"},
+				FoundBy:      "python-package-cataloger",
+				MetadataType: pkg.PythonPackageMetadataType,
+				Metadata: pkg.PythonPackageMetadata{
+					Name:                 "Pygments",
+					Version:              "2.6.1",
+					License:              "BSD License",
+					Platform:             "any",
+					Author:               "Georg Brandl",
+					AuthorEmail:          "georg@python.org",
+					SitePackagesRootPath: "test-fixtures",
+					Files: []pkg.PythonFileRecord{
+						{Path: "../../../bin/pygmentize", Digest: &pkg.Digest{"sha256", "dDhv_U2jiCpmFQwIRHpFRLAHUO4R1jIJPEvT_QYTFp8"}, Size: "220"},
+						{Path: "Pygments-2.6.1.dist-info/AUTHORS", Digest: &pkg.Digest{"sha256", "PVpa2_Oku6BGuiUvutvuPnWGpzxqFy2I8-NIrqCvqUY"}, Size: "8449"},
+						{Path: "Pygments-2.6.1.dist-info/RECORD"},
+						{Path: "pygments/__pycache__/__init__.cpython-38.pyc"},
+						{Path: "pygments/util.py", Digest: &pkg.Digest{"sha256", "586xXHiJGGZxqk5PMBu3vBhE68DLuAe5MBARWrSPGxA"}, Size: "10778"},
+					},
+					TopLevelPackages: []string{"pygments", "something_else"},
+				},
+			},
+		},
+		{
+			// in cases where the metadata file is available and the record is not we should still record there is a package
+			// additionally empty top_level.txt files should not result in an error
+			MetadataFixture: "test-fixtures/partial.dist-info/METADATA",
+			TopLevelFixture: "test-fixtures/partial.dist-info/top_level.txt",
+			ExpectedPackage: pkg.Package{
+				Name:         "Pygments",
+				Version:      "2.6.1",
+				Type:         pkg.PythonPkg,
+				Language:     pkg.Python,
+				Licenses:     []string{"BSD License"},
+				FoundBy:      "python-package-cataloger",
+				MetadataType: pkg.PythonPackageMetadataType,
+				Metadata: pkg.PythonPackageMetadata{
+					Name:                 "Pygments",
+					Version:              "2.6.1",
+					License:              "BSD License",
+					Platform:             "any",
+					Author:               "Georg Brandl",
+					AuthorEmail:          "georg@python.org",
+					SitePackagesRootPath: "test-fixtures",
+				},
+			},
+		},
+	}
+
+	for _, test := range tests {
+		t.Run(test.MetadataFixture, func(t *testing.T) {
+			resolver := newTestResolver(test.MetadataFixture, test.RecordFixture, test.TopLevelFixture)
+
+			// note that the source is the record ref created by the resolver mock... attach the expected values
+			test.ExpectedPackage.Source = []file.Reference{*resolver.metadataRef}
+			if resolver.recordRef != nil {
+				test.ExpectedPackage.Source = append(test.ExpectedPackage.Source, *resolver.recordRef)
+			}
+
+			if resolver.topLevelRef != nil {
+				test.ExpectedPackage.Source = append(test.ExpectedPackage.Source, *resolver.topLevelRef)
+			}
+			// end patching expected values with runtime data...
+
+			pyPkgCataloger := NewPythonPackageCataloger()
+
+			actual, err := pyPkgCataloger.catalogEggOrWheel(resolver, *resolver.metadataRef)
+			if err != nil {
+				t.Fatalf("failed to catalog python package: %+v", err)
+			}
+
+			for _, d := range deep.Equal(actual, &test.ExpectedPackage) {
+				t.Errorf("diff: %+v", d)
+			}
+		})
+	}
+
+}

syft/cataloger/python/parse_poetry_lock_test.go

@@ -1,10 +1,11 @@
 package python
 
 import (
-	"github.com/anchore/syft/syft/pkg"
-	"github.com/go-test/deep"
 	"os"
 	"testing"
+
+	"github.com/anchore/syft/syft/pkg"
+	"github.com/go-test/deep"
 )
 
 func TestParsePoetryLock(t *testing.T) {
@@ -13,28 +14,28 @@ func TestParsePoetryLock(t *testing.T) {
 			Name:     "added-value",
 			Version:  "0.14.2",
 			Language: pkg.Python,
-			Type:     pkg.PoetryPkg,
+			Type:     pkg.PythonPkg,
 			Licenses: nil,
 		},
 		{
 			Name:     "alabaster",
 			Version:  "0.7.12",
 			Language: pkg.Python,
-			Type:     pkg.PoetryPkg,
+			Type:     pkg.PythonPkg,
 			Licenses: nil,
 		},
 		{
 			Name:     "appnope",
 			Version:  "0.1.0",
 			Language: pkg.Python,
-			Type:     pkg.PoetryPkg,
+			Type:     pkg.PythonPkg,
 			Licenses: nil,
 		},
 		{
 			Name:     "asciitree",
 			Version:  "0.3.3",
 			Language: pkg.Python,
-			Type:     pkg.PoetryPkg,
+			Type:     pkg.PythonPkg,
 			Licenses: nil,
 		},
 	}

syft/cataloger/python/parse_requirements.go

@@ -47,7 +47,7 @@ func parseRequirementsTxt(_ string, reader io.Reader) ([]pkg.Package, error) {
 				Name:     name,
 				Version:  version,
 				Language: pkg.Python,
-				Type:     pkg.PythonRequirementsPkg,
+				Type:     pkg.PythonPkg,
 			})
 		default:
 			continue

syft/cataloger/python/parse_requirements_test.go

@@ -4,24 +4,45 @@ import (
 	"os"
 	"testing"
 
+	"github.com/go-test/deep"
+
 	"github.com/anchore/syft/syft/pkg"
 )
 
+func assertPackagesEqual(t *testing.T, actual []pkg.Package, expected map[string]pkg.Package) {
+	t.Helper()
+	if len(actual) != len(expected) {
+		for _, a := range actual {
+			t.Log("   ", a)
+		}
+		t.Fatalf("unexpected package count: %d!=%d", len(actual), len(expected))
+	}
+
+	for _, a := range actual {
+		expectedPkg, ok := expected[a.Name]
+		if !ok {
+			t.Errorf("unexpected package found: '%s'", a.Name)
+		}
+
+		for _, d := range deep.Equal(a, expectedPkg) {
+			t.Errorf("diff: %+v", d)
+		}
+	}
+}
+
 func TestParseRequirementsTxt(t *testing.T) {
 	expected := map[string]pkg.Package{
 		"foo": {
 			Name:     "foo",
 			Version:  "1.0.0",
 			Language: pkg.Python,
-			Type:     pkg.PythonRequirementsPkg,
-			Licenses: []string{},
+			Type:     pkg.PythonPkg,
 		},
 		"flask": {
 			Name:     "flask",
 			Version:  "4.0.0",
 			Language: pkg.Python,
-			Type:     pkg.PythonRequirementsPkg,
-			Licenses: []string{},
+			Type:     pkg.PythonPkg,
 		},
 	}
 	fixture, err := os.Open("test-fixtures/requires/requirements.txt")
@@ -34,6 +55,6 @@ func TestParseRequirementsTxt(t *testing.T) {
 		t.Fatalf("failed to parse requirements: %+v", err)
 	}
 
-	assertPkgsEqual(t, actual, expected)
+	assertPackagesEqual(t, actual, expected)
 
 }

syft/cataloger/python/parse_setup.go

@@ -41,7 +41,7 @@ func parseSetup(_ string, reader io.Reader) ([]pkg.Package, error) {
 				Name:     strings.Trim(name, "'\""),
 				Version:  strings.Trim(version, "'\""),
 				Language: pkg.Python,
-				Type:     pkg.PythonSetupPkg,
+				Type:     pkg.PythonPkg,
 			})
 		}
 	}

syft/cataloger/python/parse_setup_test.go

@@ -13,36 +13,31 @@ func TestParseSetup(t *testing.T) {
 			Name:     "pathlib3",
 			Version:  "2.2.0",
 			Language: pkg.Python,
-			Type:     pkg.PythonSetupPkg,
-			Licenses: []string{},
+			Type:     pkg.PythonPkg,
 		},
 		"mypy": {
 			Name:     "mypy",
 			Version:  "v0.770",
 			Language: pkg.Python,
-			Type:     pkg.PythonSetupPkg,
-			Licenses: []string{},
+			Type:     pkg.PythonPkg,
 		},
 		"mypy1": {
 			Name:     "mypy1",
 			Version:  "v0.770",
 			Language: pkg.Python,
-			Type:     pkg.PythonSetupPkg,
-			Licenses: []string{},
+			Type:     pkg.PythonPkg,
 		},
 		"mypy2": {
 			Name:     "mypy2",
 			Version:  "v0.770",
 			Language: pkg.Python,
-			Type:     pkg.PythonSetupPkg,
-			Licenses: []string{},
+			Type:     pkg.PythonPkg,
 		},
 		"mypy3": {
 			Name:     "mypy3",
 			Version:  "v0.770",
 			Language: pkg.Python,
-			Type:     pkg.PythonSetupPkg,
-			Licenses: []string{},
+			Type:     pkg.PythonPkg,
 		},
 	}
 	fixture, err := os.Open("test-fixtures/setup/setup.py")
@@ -55,6 +50,6 @@ func TestParseSetup(t *testing.T) {
 		t.Fatalf("failed to parse requirements: %+v", err)
 	}
 
-	assertPkgsEqual(t, actual, expected)
+	assertPackagesEqual(t, actual, expected)
 
 }

syft/cataloger/python/parse_wheel_egg.go

@@ -1,100 +0,0 @@
-package python
-
-import (
-	"bufio"
-	"fmt"
-	"io"
-	"strings"
-
-	"github.com/anchore/syft/syft/cataloger/common"
-	"github.com/anchore/syft/syft/pkg"
-)
-
-// integrity check
-var _ common.ParserFn = parseWheelMetadata
-var _ common.ParserFn = parseEggMetadata
-
-// parseWheelMetadata is a parser function for individual Python Wheel metadata file contents, returning all Python
-// packages listed.
-func parseWheelMetadata(_ string, reader io.Reader) ([]pkg.Package, error) {
-	packages, err := parseWheelOrEggMetadata(reader)
-	for idx := range packages {
-		packages[idx].Type = pkg.WheelPkg
-	}
-	return packages, err
-}
-
-// parseEggMetadata is a parser function for individual Python Egg metadata file contents, returning all Python
-// packages listed.
-func parseEggMetadata(_ string, reader io.Reader) ([]pkg.Package, error) {
-	packages, err := parseWheelOrEggMetadata(reader)
-	for idx := range packages {
-		packages[idx].Type = pkg.EggPkg
-	}
-	return packages, err
-}
-
-// parseWheelOrEggMetadata takes a Python Egg or Wheel (which share the same format and values for our purposes),
-// returning all Python packages listed.
-func parseWheelOrEggMetadata(reader io.Reader) ([]pkg.Package, error) {
-	fields := make(map[string]string)
-	var key string
-
-	scanner := bufio.NewScanner(reader)
-	for scanner.Scan() {
-		line := scanner.Text()
-
-		line = strings.TrimRight(line, "\n")
-
-		// empty line indicates end of entry
-		if len(line) == 0 {
-			// if the entry has not started, keep parsing lines
-			if len(fields) == 0 {
-				continue
-			}
-			break
-		}
-
-		switch {
-		case strings.HasPrefix(line, " "):
-			// a field-body continuation
-			if len(key) == 0 {
-				return nil, fmt.Errorf("no match for continuation: line: '%s'", line)
-			}
-
-			val, ok := fields[key]
-			if !ok {
-				return nil, fmt.Errorf("no previous key exists, expecting: %s", key)
-			}
-			// concatenate onto previous value
-			val = fmt.Sprintf("%s\n %s", val, strings.TrimSpace(line))
-			fields[key] = val
-		default:
-			// parse a new key (note, duplicate keys are overridden)
-			if i := strings.Index(line, ":"); i > 0 {
-				key = strings.TrimSpace(line[0:i])
-				val := strings.TrimSpace(line[i+1:])
-
-				fields[key] = val
-			} else {
-				return nil, fmt.Errorf("cannot parse field from line: '%s'", line)
-			}
-		}
-	}
-
-	if err := scanner.Err(); err != nil {
-		return nil, fmt.Errorf("failed to parse python wheel/egg: %w", err)
-	}
-
-	p := pkg.Package{
-		Name:     fields["Name"],
-		Version:  fields["Version"],
-		Language: pkg.Python,
-	}
-
-	if license, ok := fields["License"]; ok && license != "" {
-		p.Licenses = []string{license}
-	}
-
-	return []pkg.Package{p}, nil
-}

syft/cataloger/python/parse_wheel_egg_metadata.go

@@ -0,0 +1,80 @@
+package python
+
+import (
+	"bufio"
+	"fmt"
+	"io"
+	"path/filepath"
+	"strings"
+
+	"github.com/anchore/stereoscope/pkg/file"
+
+	"github.com/mitchellh/mapstructure"
+
+	"github.com/anchore/syft/syft/pkg"
+)
+
+// parseWheelOrEggMetadata takes a Python Egg or Wheel (which share the same format and values for our purposes),
+// returning all Python packages listed.
+func parseWheelOrEggMetadata(path file.Path, reader io.Reader) (pkg.PythonPackageMetadata, error) {
+	fields := make(map[string]string)
+	var key string
+
+	scanner := bufio.NewScanner(reader)
+	for scanner.Scan() {
+		line := scanner.Text()
+		line = strings.TrimRight(line, "\n")
+
+		// empty line indicates end of entry
+		if len(line) == 0 {
+			// if the entry has not started, keep parsing lines
+			if len(fields) == 0 {
+				continue
+			}
+			break
+		}
+
+		switch {
+		case strings.HasPrefix(line, " "):
+			// a field-body continuation
+			if len(key) == 0 {
+				return pkg.PythonPackageMetadata{}, fmt.Errorf("no match for continuation: line: '%s'", line)
+			}
+
+			val, ok := fields[key]
+			if !ok {
+				return pkg.PythonPackageMetadata{}, fmt.Errorf("no previous key exists, expecting: %s", key)
+			}
+			// concatenate onto previous value
+			val = fmt.Sprintf("%s\n %s", val, strings.TrimSpace(line))
+			fields[key] = val
+		default:
+			// parse a new key (note, duplicate keys are overridden)
+			if i := strings.Index(line, ":"); i > 0 {
+				// mapstruct cannot map keys with dashes, and we are expected to persist the "Author-email" field
+				key = strings.ReplaceAll(strings.TrimSpace(line[0:i]), "-", "")
+				val := strings.TrimSpace(line[i+1:])
+
+				fields[key] = val
+			} else {
+				return pkg.PythonPackageMetadata{}, fmt.Errorf("cannot parse field from line: '%s'", line)
+			}
+		}
+	}
+
+	if err := scanner.Err(); err != nil {
+		return pkg.PythonPackageMetadata{}, fmt.Errorf("failed to parse python wheel/egg: %w", err)
+	}
+
+	var metadata pkg.PythonPackageMetadata
+	if err := mapstructure.Decode(fields, &metadata); err != nil {
+		return pkg.PythonPackageMetadata{}, fmt.Errorf("unable to parse APK metadata: %w", err)
+	}
+
+	// add additional metadata not stored in the egg/wheel metadata file
+
+	sitePackagesRoot := filepath.Clean(filepath.Join(filepath.Dir(string(path)), ".."))
+	metadata.SitePackagesRootPath = sitePackagesRoot
+
+	return metadata, nil
+}

syft/cataloger/python/parse_wheel_egg_metadata_test.go

@@ -0,0 +1,62 @@
+package python
+
+import (
+	"os"
+	"testing"
+
+	"github.com/anchore/stereoscope/pkg/file"
+
+	"github.com/anchore/syft/syft/pkg"
+	"github.com/go-test/deep"
+)
+
+func TestParseWheelEggMetadata(t *testing.T) {
+	tests := []struct {
+		Fixture          string
+		ExpectedMetadata pkg.PythonPackageMetadata
+	}{
+		{
+			Fixture: "test-fixtures/egg-info/PKG-INFO",
+			ExpectedMetadata: pkg.PythonPackageMetadata{
+				Name:                 "requests",
+				Version:              "2.22.0",
+				License:              "Apache 2.0",
+				Platform:             "UNKNOWN",
+				Author:               "Kenneth Reitz",
+				AuthorEmail:          "me@kennethreitz.org",
+				SitePackagesRootPath: "test-fixtures",
+			},
+		},
+		{
+			Fixture: "test-fixtures/dist-info/METADATA",
+			ExpectedMetadata: pkg.PythonPackageMetadata{
+				Name:                 "Pygments",
+				Version:              "2.6.1",
+				License:              "BSD License",
+				Platform:             "any",
+				Author:               "Georg Brandl",
+				AuthorEmail:          "georg@python.org",
+				SitePackagesRootPath: "test-fixtures",
+			},
+		},
+	}
+
+	for _, test := range tests {
+		t.Run(test.Fixture, func(t *testing.T) {
+			fixture, err := os.Open(test.Fixture)
+			if err != nil {
+				t.Fatalf("failed to open fixture: %+v", err)
+			}
+
+			actual, err := parseWheelOrEggMetadata(file.Path(test.Fixture), fixture)
+			if err != nil {
+				t.Fatalf("failed to parse: %+v", err)
+			}
+
+			for _, d := range deep.Equal(actual, test.ExpectedMetadata) {
+				t.Errorf("diff: %+v", d)
+			}
+		})
+	}
+
+}

syft/cataloger/python/parse_wheel_egg_record.go

@@ -0,0 +1,60 @@
+package python
+
+import (
+	"encoding/csv"
+	"fmt"
+	"io"
+	"strings"
+
+	"github.com/anchore/syft/syft/pkg"
+)
+
+// parseWheelOrEggMetadata takes a Python Egg or Wheel (which share the same format and values for our purposes),
+// returning all Python packages listed.
+func parseWheelOrEggRecord(reader io.Reader) ([]pkg.PythonFileRecord, error) {
+	var records []pkg.PythonFileRecord
+	r := csv.NewReader(reader)
+
+	for {
+		recordList, err := r.Read()
+		if err == io.EOF {
+			break
+		}
+		if err != nil {
+			return nil, fmt.Errorf("unable to read python record file: %w", err)
+		}
+
+		if len(recordList) != 3 {
+			return nil, fmt.Errorf("python record an unexpected length=%d: %q", len(recordList), recordList)
+		}
+
+		var record pkg.PythonFileRecord
+
+		for idx, item := range recordList {
+			switch idx {
+			case 0:
+				record.Path = item
+			case 1:
+				if item == "" {
+					continue
+				}
+				fields := strings.Split(item, "=")
+
+				if len(fields) != 2 {
+					return nil, fmt.Errorf("unexpected python record digest: %q", item)
+				}
+
+				record.Digest = &pkg.Digest{
+					Algorithm: fields[0],
+					Value:     fields[1],
+				}
+			case 2:
+				record.Size = item
+			}
+		}
+
+		records = append(records, record)
+	}
+
+	return records, nil
+}

syft/cataloger/python/parse_wheel_egg_record_test.go

@@ -0,0 +1,57 @@
+package python
+
+import (
+	"os"
+	"testing"
+
+	"github.com/anchore/syft/syft/pkg"
+	"github.com/go-test/deep"
+)
+
+func TestParseWheelEggRecord(t *testing.T) {
+	tests := []struct {
+		Fixture          string
+		ExpectedMetadata []pkg.PythonFileRecord
+	}{
+		{
+			Fixture: "test-fixtures/egg-info/RECORD",
+			ExpectedMetadata: []pkg.PythonFileRecord{
+				{Path: "requests-2.22.0.dist-info/INSTALLER", Digest: &pkg.Digest{"sha256", "zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg"}, Size: "4"},
+				{Path: "requests/__init__.py", Digest: &pkg.Digest{"sha256", "PnKCgjcTq44LaAMzB-7--B2FdewRrE8F_vjZeaG9NhA"}, Size: "3921"},
+				{Path: "requests/__pycache__/__version__.cpython-38.pyc"},
+				{Path: "requests/__pycache__/utils.cpython-38.pyc"},
+				{Path: "requests/__version__.py", Digest: &pkg.Digest{"sha256", "Bm-GFstQaFezsFlnmEMrJDe8JNROz9n2XXYtODdvjjc"}, Size: "436"},
+				{Path: "requests/utils.py", Digest: &pkg.Digest{"sha256", "LtPJ1db6mJff2TJSJWKi7rBpzjPS3mSOrjC9zRhoD3A"}, Size: "30049"},
+			},
+		},
+		{
+			Fixture: "test-fixtures/dist-info/RECORD",
+			ExpectedMetadata: []pkg.PythonFileRecord{
+				{Path: "../../../bin/pygmentize", Digest: &pkg.Digest{"sha256", "dDhv_U2jiCpmFQwIRHpFRLAHUO4R1jIJPEvT_QYTFp8"}, Size: "220"},
+				{Path: "Pygments-2.6.1.dist-info/AUTHORS", Digest: &pkg.Digest{"sha256", "PVpa2_Oku6BGuiUvutvuPnWGpzxqFy2I8-NIrqCvqUY"}, Size: "8449"},
+				{Path: "Pygments-2.6.1.dist-info/RECORD"},
+				{Path: "pygments/__pycache__/__init__.cpython-38.pyc"},
+				{Path: "pygments/util.py", Digest: &pkg.Digest{"sha256", "586xXHiJGGZxqk5PMBu3vBhE68DLuAe5MBARWrSPGxA"}, Size: "10778"},
+			},
+		},
+	}
+
+	for _, test := range tests {
+		t.Run(test.Fixture, func(t *testing.T) {
+			fixture, err := os.Open(test.Fixture)
+			if err != nil {
+				t.Fatalf("failed to open fixture: %+v", err)
+			}
+
+			actual, err := parseWheelOrEggRecord(fixture)
+			if err != nil {
+				t.Fatalf("failed to parse: %+v", err)
+			}
+
+			for _, d := range deep.Equal(actual, test.ExpectedMetadata) {
+				t.Errorf("diff: %+v", d)
+			}
+		})
+	}
+
+}

syft/cataloger/python/parse_wheel_egg_test.go

@@ -1,93 +0,0 @@
-package python
-
-import (
-	"os"
-	"testing"
-
-	"github.com/anchore/syft/syft/pkg"
-)
-
-func assertPkgsEqual(t *testing.T, actual []pkg.Package, expected map[string]pkg.Package) {
-	t.Helper()
-	if len(actual) != len(expected) {
-		for _, a := range actual {
-			t.Log("   ", a)
-		}
-		t.Fatalf("unexpected package count: %d!=%d", len(actual), len(expected))
-	}
-
-	for _, a := range actual {
-		expectedPkg, ok := expected[a.Name]
-		if !ok {
-			t.Errorf("unexpected package found: '%s'", a.Name)
-		}
-
-		if expectedPkg.Version != a.Version {
-			t.Errorf("unexpected package version: '%s'", a.Version)
-		}
-
-		if a.Language != expectedPkg.Language {
-			t.Errorf("bad language: '%+v'", a.Language)
-		}
-
-		if a.Type != expectedPkg.Type {
-			t.Errorf("bad package type: %+v", a.Type)
-		}
-
-		if len(a.Licenses) < len(expectedPkg.Licenses) {
-			t.Errorf("bad package licenses count: '%+v'", a.Licenses)
-		}
-		if len(a.Licenses) > 0 {
-			if a.Licenses[0] != expectedPkg.Licenses[0] {
-				t.Errorf("bad package licenses: '%+v'", a.Licenses)
-			}
-		}
-
-	}
-}
-
-func TestParseEggMetadata(t *testing.T) {
-	expected := map[string]pkg.Package{
-		"requests": {
-			Name:     "requests",
-			Version:  "2.22.0",
-			Language: pkg.Python,
-			Type:     pkg.EggPkg,
-			Licenses: []string{"Apache 2.0"},
-		},
-	}
-	fixture, err := os.Open("test-fixtures/egg-info/PKG-INFO")
-	if err != nil {
-		t.Fatalf("failed to open fixture: %+v", err)
-	}
-
-	actual, err := parseEggMetadata(fixture.Name(), fixture)
-	if err != nil {
-		t.Fatalf("failed to parse egg-info: %+v", err)
-	}
-
-	assertPkgsEqual(t, actual, expected)
-}
-
-func TestParseWheelMetadata(t *testing.T) {
-	expected := map[string]pkg.Package{
-		"Pygments": {
-			Name:     "Pygments",
-			Version:  "2.6.1",
-			Language: pkg.Python,
-			Type:     pkg.WheelPkg,
-			Licenses: []string{"BSD License"},
-		},
-	}
-	fixture, err := os.Open("test-fixtures/dist-info/METADATA")
-	if err != nil {
-		t.Fatalf("failed to open fixture: %+v", err)
-	}
-
-	actual, err := parseWheelMetadata(fixture.Name(), fixture)
-	if err != nil {
-		t.Fatalf("failed to parse dist-info: %+v", err)
-	}
-
-	assertPkgsEqual(t, actual, expected)
-}

syft/cataloger/python/poetry_metadata_package.go

@@ -16,6 +16,6 @@ func (p PoetryMetadataPackage) Pkg() pkg.Package {
 		Name:     p.Name,
 		Version:  p.Version,
 		Language: pkg.Python,
-		Type:     pkg.PoetryPkg,
+		Type:     pkg.PythonPkg,
 	}
 }

syft/cataloger/python/test-fixtures/dist-info/RECORD

@@ -0,0 +1,5 @@
+../../../bin/pygmentize,sha256=dDhv_U2jiCpmFQwIRHpFRLAHUO4R1jIJPEvT_QYTFp8,220
+Pygments-2.6.1.dist-info/AUTHORS,sha256=PVpa2_Oku6BGuiUvutvuPnWGpzxqFy2I8-NIrqCvqUY,8449
+Pygments-2.6.1.dist-info/RECORD,,
+pygments/__pycache__/__init__.cpython-38.pyc,,
+pygments/util.py,sha256=586xXHiJGGZxqk5PMBu3vBhE68DLuAe5MBARWrSPGxA,10778

syft/cataloger/python/test-fixtures/dist-info/top_level.txt

@@ -0,0 +1,2 @@
+pygments
+something_else

syft/cataloger/python/test-fixtures/egg-info/RECORD

@@ -0,0 +1,6 @@
+requests-2.22.0.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+requests/__init__.py,sha256=PnKCgjcTq44LaAMzB-7--B2FdewRrE8F_vjZeaG9NhA,3921
+requests/__pycache__/__version__.cpython-38.pyc,,
+requests/__pycache__/utils.cpython-38.pyc,,
+requests/__version__.py,sha256=Bm-GFstQaFezsFlnmEMrJDe8JNROz9n2XXYtODdvjjc,436
+requests/utils.py,sha256=LtPJ1db6mJff2TJSJWKi7rBpzjPS3mSOrjC9zRhoD3A,30049

syft/cataloger/python/test-fixtures/egg-info/top_level.txt

@@ -0,0 +1 @@
+requests

syft/cataloger/python/test-fixtures/partial.dist-info/METADATA

@@ -0,0 +1,47 @@
+Metadata-Version: 2.1
+Name: Pygments
+Version: 2.6.1
+Summary: Pygments is a syntax highlighting package written in Python.
+Home-page: https://pygments.org/
+Author: Georg Brandl
+Author-email: georg@python.org
+License: BSD License
+Keywords: syntax highlighting
+Platform: any
+Classifier: License :: OSI Approved :: BSD License
+Classifier: Intended Audience :: Developers
+Classifier: Intended Audience :: End Users/Desktop
+Classifier: Intended Audience :: System Administrators
+Classifier: Development Status :: 6 - Mature
+Classifier: Programming Language :: Python
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.5
+Classifier: Programming Language :: Python :: 3.6
+Classifier: Programming Language :: Python :: 3.7
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: Implementation :: CPython
+Classifier: Programming Language :: Python :: Implementation :: PyPy
+Classifier: Operating System :: OS Independent
+Classifier: Topic :: Text Processing :: Filters
+Classifier: Topic :: Utilities
+Requires-Python: >=3.5
+
+
+Pygments
+~~~~~~~~
+
+Pygments is a syntax highlighting package written in Python.
+
+It is a generic syntax highlighter suitable for use in code hosting, forums,
+wikis or other applications that need to prettify source code.  Highlights
+are:
+
+* a wide range of over 500 languages and other text formats is supported
+* special attention is paid to details, increasing quality by a fair amount
+* support for new languages and formats are added easily
+* a number of output formats, presently HTML, LaTeX, RTF, SVG, all image     formats that PIL supports and ANSI sequences
+* it is usable as a command-line tool and as a library
+
+:copyright: Copyright 2006-2019 by the Pygments team, see AUTHORS.
+:license: BSD, see LICENSE for details.
+

syft/cataloger/rpmdb/parse_rpmdb.go

@@ -52,8 +52,9 @@ func parseRpmDB(_ string, reader io.Reader) ([]pkg.Package, error) {
 			Name:    entry.Name,
 			Version: fmt.Sprintf("%s-%s", entry.Version, entry.Release), // this is what engine does
 			//Version: fmt.Sprintf("%d:%s-%s.%s", entry.Epoch, entry.Version, entry.Release, entry.Arch),
-			Type: pkg.RpmPkg,
-			Metadata: pkg.RpmMetadata{
+			Type:         pkg.RpmPkg,
+			MetadataType: pkg.RpmdbMetadataType,
+			Metadata: pkg.RpmdbMetadata{
 				Name:      entry.Name,
 				Version:   entry.Version,
 				Epoch:     entry.Epoch,

syft/cataloger/rpmdb/parse_rpmdb_test.go

@@ -1,19 +1,21 @@
 package rpmdb
 
 import (
-	"github.com/anchore/syft/syft/pkg"
-	"github.com/go-test/deep"
 	"os"
 	"testing"
+
+	"github.com/anchore/syft/syft/pkg"
+	"github.com/go-test/deep"
 )
 
 func TestParseRpmDB(t *testing.T) {
 	expected := map[string]pkg.Package{
 		"dive": {
-			Name:    "dive",
-			Version: "0.9.2-1",
-			Type:    pkg.RpmPkg,
-			Metadata: pkg.RpmMetadata{
+			Name:         "dive",
+			Version:      "0.9.2-1",
+			Type:         pkg.RpmPkg,
+			MetadataType: pkg.RpmdbMetadataType,
+			Metadata: pkg.RpmdbMetadata{
 				Name:      "dive",
 				Epoch:     0,
 				Arch:      "x86_64",

syft/cataloger/ruby/parse_gemspec.go

@@ -8,6 +8,8 @@ import (
 	"regexp"
 	"strings"
 
+	"github.com/anchore/syft/internal"
+
 	"github.com/mitchellh/mapstructure"
 
 	"github.com/anchore/syft/syft/cataloger/common"
@@ -76,7 +78,7 @@ func parseGemSpecEntries(_ string, reader io.Reader) ([]pkg.Package, error) {
 		}
 
 		for field, pattern := range patterns {
-			matchMap := matchCaptureGroups(pattern, sanitizedLine)
+			matchMap := internal.MatchCaptureGroups(pattern, sanitizedLine)
 			if value := matchMap[field]; value != "" {
 				if postProcessor := postProcessors[field]; postProcessor != nil {
 					fields[field] = postProcessor(value)
@@ -96,12 +98,13 @@ func parseGemSpecEntries(_ string, reader io.Reader) ([]pkg.Package, error) {
 		}
 
 		pkgs = append(pkgs, pkg.Package{
-			Name:     metadata.Name,
-			Version:  metadata.Version,
-			Licenses: metadata.Licenses,
-			Language: pkg.Ruby,
-			Type:     pkg.GemPkg,
-			Metadata: metadata,
+			Name:         metadata.Name,
+			Version:      metadata.Version,
+			Licenses:     metadata.Licenses,
+			Language:     pkg.Ruby,
+			Type:         pkg.GemPkg,
+			MetadataType: pkg.GemMetadataType,
+			Metadata:     metadata,
 		})
 	}
 
@@ -121,15 +124,3 @@ func renderUtf8(s string) string {
 	})
 	return fullReplacement
 }
-
-// matchCaptureGroups takes a regular expression and string and returns all of the named capture group results in a map.
-func matchCaptureGroups(regEx *regexp.Regexp, str string) map[string]string {
-	match := regEx.FindStringSubmatch(str)
-	results := make(map[string]string)
-	for i, name := range regEx.SubexpNames() {
-		if i > 0 && i <= len(match) {
-			results[name] = match[i]
-		}
-	}
-	return results
-}

syft/cataloger/ruby/parse_gemspec_test.go

@@ -10,17 +10,19 @@ import (
 
 func TestParseGemspec(t *testing.T) {
 	var expectedPkg = pkg.Package{
-		Name:     "bundler",
-		Version:  "2.1.4",
-		Type:     pkg.GemPkg,
-		Licenses: []string{"MIT"},
-		Language: pkg.Ruby,
+		Name:         "bundler",
+		Version:      "2.1.4",
+		Type:         pkg.GemPkg,
+		Licenses:     []string{"MIT"},
+		Language:     pkg.Ruby,
+		MetadataType: pkg.GemMetadataType,
 		Metadata: pkg.GemMetadata{
 			Name:     "bundler",
 			Version:  "2.1.4",
 			Files:    []string{"exe/bundle", "exe/bundler"},
 			Authors:  []string{"André Arko", "Samuel Giddins", "Colby Swandale", "Hiroshi Shibata", "David Rodríguez", "Grey Baker", "Stephanie Morillo", "Chris Morris", "James Wen", "Tim Moore", "André Medeiros", "Jessica Lynn Suttles", "Terence Lee", "Carl Lerche", "Yehuda Katz"},
 			Licenses: []string{"MIT"},
+			Homepage: "https://bundler.io",
 		},
 	}
 

syft/pkg/gem_metadata.go

@@ -3,7 +3,8 @@ package pkg
 type GemMetadata struct {
 	Name     string   `mapstructure:"name" json:"name"`
 	Version  string   `mapstructure:"version" json:"version"`
-	Files    []string `mapstructure:"files" json:"files"`
-	Authors  []string `mapstructure:"authors" json:"authors"`
-	Licenses []string `mapstructure:"licenses" json:"licenses"`
+	Files    []string `mapstructure:"files" json:"files,omitempty"`
+	Authors  []string `mapstructure:"authors" json:"authors,omitempty"`
+	Licenses []string `mapstructure:"licenses" json:"licenses,omitempty"`
+	Homepage string   `mapstructure:"homepage" json:"homepage,omitempty"`
 }

syft/pkg/metadata.go

@@ -0,0 +1,14 @@
+package pkg
+
+type MetadataType string
+
+const (
+	UnknownMetadataType        MetadataType = "UnknownMetadata"
+	ApkMetadataType            MetadataType = "apk-metadata"
+	DpkgMetadataType           MetadataType = "dpkg-metadata"
+	GemMetadataType            MetadataType = "gem-metadata"
+	JavaMetadataType           MetadataType = "java-metadata"
+	NpmPackageJSONMetadataType MetadataType = "npm-package-json-metadata"
+	RpmdbMetadataType          MetadataType = "rpmdb-metadata"
+	PythonPackageMetadataType  MetadataType = "python-package-metadata"
+)

syft/pkg/npm_metadata.go

@@ -0,0 +1,11 @@
+package pkg
+
+// NpmPackageJSONMetadata holds extra information that is used in pkg.Package
+type NpmPackageJSONMetadata struct {
+	Files       []string `mapstructure:"files" json:"files,omitempty"`
+	Author      string   `mapstructure:"author" json:"author"`
+	Licenses    []string `mapstructure:"licenses" json:"licenses"`
+	Homepage    string   `mapstructure:"homepage" json:"homepage"`
+	Description string   `mapstructure:"description" json:"description"`
+	URL         string   `mapstructure:"url" json:"url"`
+}

syft/pkg/package.go

@@ -23,10 +23,11 @@ type Package struct {
 	FoundBy string           `json:"foundBy"`  // the specific cataloger that discovered this package
 	Source  []file.Reference `json:"sources"`  // the locations that lead to the discovery of this package (note: this is not necessarily the locations that make up this package)
 	// TODO: should we move licenses into metadata?
-	Licenses []string    `json:"licenses"`           // licenses discovered with the package metadata
-	Language Language    `json:"language"`           // the language ecosystem this package belongs to (e.g. JavaScript, Python, etc)
-	Type     Type        `json:"type"`               // the package type (e.g. Npm, Yarn, Egg, Wheel, Rpm, Deb, etc)
-	Metadata interface{} `json:"metadata,omitempty"` // additional data found while parsing the package source
+	Licenses     []string     `json:"licenses"`           // licenses discovered with the package metadata
+	Language     Language     `json:"language"`           // the language ecosystem this package belongs to (e.g. JavaScript, Python, etc)
+	Type         Type         `json:"type"`               // the package type (e.g. Npm, Yarn, Python, Rpm, Deb, etc)
+	MetadataType MetadataType `json:"metadataType"`       // the shape of the additional data in the "metadata" field
+	Metadata     interface{}  `json:"metadata,omitempty"` // additional data found while parsing the package source
 }
 
 // ID returns the package ID, which is unique relative to a package catalog.

syft/pkg/package_test.go

@@ -25,7 +25,7 @@ func TestPackage_pURL(t *testing.T) {
 			pkg: Package{
 				Name:    "name",
 				Version: "v0.1.0",
-				Type:    WheelPkg,
+				Type:    PythonPkg,
 			},
 			expected: "pkg:pypi/name@v0.1.0",
 		},
@@ -33,7 +33,7 @@ func TestPackage_pURL(t *testing.T) {
 			pkg: Package{
 				Name:    "name",
 				Version: "v0.1.0",
-				Type:    EggPkg,
+				Type:    PythonPkg,
 			},
 			expected: "pkg:pypi/name@v0.1.0",
 		},
@@ -41,7 +41,7 @@ func TestPackage_pURL(t *testing.T) {
 			pkg: Package{
 				Name:    "name",
 				Version: "v0.1.0",
-				Type:    PythonSetupPkg,
+				Type:    PythonPkg,
 			},
 			expected: "pkg:pypi/name@v0.1.0",
 		},
@@ -49,7 +49,7 @@ func TestPackage_pURL(t *testing.T) {
 			pkg: Package{
 				Name:    "name",
 				Version: "v0.1.0",
-				Type:    PythonRequirementsPkg,
+				Type:    PythonPkg,
 			},
 			expected: "pkg:pypi/name@v0.1.0",
 		},
@@ -69,14 +69,6 @@ func TestPackage_pURL(t *testing.T) {
 			},
 			expected: "pkg:npm/name@v0.1.0",
 		},
-		{
-			pkg: Package{
-				Name:    "name",
-				Version: "v0.1.0",
-				Type:    YarnPkg,
-			},
-			expected: "pkg:npm/name@v0.1.0",
-		},
 		{
 			distro: distro.Distro{
 				Type: distro.Ubuntu,
@@ -101,7 +93,7 @@ func TestPackage_pURL(t *testing.T) {
 				Name:    "bad-name",
 				Version: "bad-v0.1.0",
 				Type:    RpmPkg,
-				Metadata: RpmMetadata{
+				Metadata: RpmdbMetadata{
 					Name:    "name",
 					Version: "v0.1.0",
 					Epoch:   2,

syft/pkg/python_package_metadata.go

@@ -0,0 +1,26 @@
+package pkg
+
+type Digest struct {
+	Algorithm string `json:"algorithm"`
+	Value     string `json:"value"`
+}
+
+// PythonFileRecord represents a single entry within a RECORD file for a python wheel or egg package
+type PythonFileRecord struct {
+	Path   string  `json:"path"`
+	Digest *Digest `json:"digest,omitempty"`
+	Size   string  `json:"size,omitempty"`
+}
+
+// PythonPackageMetadata represents all captured data for a python egg or wheel package.
+type PythonPackageMetadata struct {
+	Name                 string             `json:"name" mapstruct:"Name"`
+	Version              string             `json:"version" mapstruct:"Version"`
+	License              string             `json:"license" mapstruct:"License"`
+	Author               string             `json:"author" mapstruct:"Author"`
+	AuthorEmail          string             `json:"authorEmail" mapstruct:"Authoremail"`
+	Platform             string             `json:"platform" mapstruct:"Platform"`
+	Files                []PythonFileRecord `json:"files,omitempty"`
+	SitePackagesRootPath string             `json:"sitePackagesRootPath"`
+	TopLevelPackages     []string           `json:"topLevelPackages,omitempty"`
+}

syft/pkg/rpmdb_metadata.go

@@ -7,8 +7,8 @@ import (
 	"github.com/package-url/packageurl-go"
 )
 
-// RpmMetadata represents all captured data for a RPM DB package entry.
-type RpmMetadata struct {
+// RpmdbMetadata represents all captured data for a RPM DB package entry.
+type RpmdbMetadata struct {
 	Name      string `json:"name"`
 	Version   string `json:"version"`
 	Epoch     int    `json:"epoch"`
@@ -20,7 +20,7 @@ type RpmMetadata struct {
 	Vendor    string `json:"vendor"`
 }
 
-func (m RpmMetadata) PackageURL(d distro.Distro) string {
+func (m RpmdbMetadata) PackageURL(d distro.Distro) string {
 	pURL := packageurl.NewPackageURL(
 		packageurl.TypeRPM,
 		d.Type.String(),

syft/pkg/rpmdb_metadata_test.go

@@ -1,22 +1,23 @@
 package pkg
 
 import (
+	"testing"
+
 	"github.com/anchore/syft/syft/distro"
 	"github.com/sergi/go-diff/diffmatchpatch"
-	"testing"
 )
 
 func TestRpmMetadata_pURL(t *testing.T) {
 	tests := []struct {
 		distro   distro.Distro
-		metadata RpmMetadata
+		metadata RpmdbMetadata
 		expected string
 	}{
 		{
 			distro: distro.Distro{
 				Type: distro.CentOS,
 			},
-			metadata: RpmMetadata{
+			metadata: RpmdbMetadata{
 				Name:    "p",
 				Version: "v",
 				Arch:    "a",
@@ -29,7 +30,7 @@ func TestRpmMetadata_pURL(t *testing.T) {
 			distro: distro.Distro{
 				Type: distro.RedHat,
 			},
-			metadata: RpmMetadata{
+			metadata: RpmdbMetadata{
 				Name:    "p",
 				Version: "v",
 				Arch:    "a",

syft/pkg/type.go

@@ -6,36 +6,25 @@ import "github.com/package-url/packageurl-go"
 type Type string
 
 const (
-	UnknownPkg Type = "UnknownPackage"
-	ApkPkg     Type = "apk"
-	GemPkg     Type = "gem"
-	DebPkg     Type = "deb"
-	EggPkg     Type = "egg"
-	// PacmanPkg Type = "pacman"
-	RpmPkg                Type = "rpm"
-	WheelPkg              Type = "wheel"
-	PoetryPkg             Type = "poetry"
-	NpmPkg                Type = "npm"
-	YarnPkg               Type = "yarn"
-	PythonRequirementsPkg Type = "python-requirements"
-	PythonSetupPkg        Type = "python-setup"
-	JavaPkg               Type = "java-archive"
-	JenkinsPluginPkg      Type = "jenkins-plugin"
-	GoModulePkg           Type = "go-module"
+	UnknownPkg       Type = "UnknownPackage"
+	ApkPkg           Type = "apk"
+	GemPkg           Type = "gem"
+	DebPkg           Type = "deb"
+	RpmPkg           Type = "rpm"
+	NpmPkg           Type = "npm"
+	PythonPkg        Type = "python"
+	JavaPkg          Type = "java-archive"
+	JenkinsPluginPkg Type = "jenkins-plugin"
+	GoModulePkg      Type = "go-module"
 )
 
 var AllPkgs = []Type{
 	ApkPkg,
 	GemPkg,
 	DebPkg,
-	EggPkg,
-	// PacmanPkg,
 	RpmPkg,
-	WheelPkg,
 	NpmPkg,
-	YarnPkg,
-	PythonRequirementsPkg,
-	PythonSetupPkg,
+	PythonPkg,
 	JavaPkg,
 	JenkinsPluginPkg,
 	GoModulePkg,
@@ -49,9 +38,9 @@ func (t Type) PackageURLType() string {
 		return packageurl.TypeGem
 	case DebPkg:
 		return "deb"
-	case EggPkg, WheelPkg, PythonRequirementsPkg, PythonSetupPkg:
+	case PythonPkg:
 		return packageurl.TypePyPi
-	case NpmPkg, YarnPkg:
+	case NpmPkg:
 		return packageurl.TypeNPM
 	case JavaPkg, JenkinsPluginPkg:
 		return packageurl.TypeMaven

syft/presenter/cyclonedx/presenter_test.go

@@ -3,11 +3,12 @@ package cyclonedx
 import (
 	"bytes"
 	"flag"
-	"github.com/anchore/stereoscope/pkg/imagetest"
-	"github.com/anchore/syft/syft/distro"
 	"regexp"
 	"testing"
 
+	"github.com/anchore/stereoscope/pkg/imagetest"
+	"github.com/anchore/syft/syft/distro"
+
 	"github.com/anchore/go-testutils"
 	"github.com/anchore/stereoscope/pkg/file"
 	"github.com/anchore/syft/syft/pkg"
@@ -109,7 +110,7 @@ func TestCycloneDxImgsPresenter(t *testing.T) {
 		},
 		Type:    pkg.RpmPkg,
 		FoundBy: "the-cataloger-1",
-		Metadata: pkg.RpmMetadata{
+		Metadata: pkg.RpmdbMetadata{
 			Name:      "package1",
 			Epoch:     0,
 			Arch:      "x86_64",
@@ -133,7 +134,7 @@ func TestCycloneDxImgsPresenter(t *testing.T) {
 			"MIT",
 			"Apache-v2",
 		},
-		Metadata: pkg.RpmMetadata{
+		Metadata: pkg.RpmdbMetadata{
 			Name:      "package2",
 			Epoch:     0,
 			Arch:      "x86_64",

syft/scope/resolver.go

@@ -16,13 +16,20 @@ type Resolver interface {
 
 // ContentResolver knows how to get file content for given file.References
 type ContentResolver interface {
+	FileContentsByRef(ref file.Reference) (string, error)
 	MultipleFileContentsByRef(f ...file.Reference) (map[file.Reference]string, error)
+	// TODO: we should consider refactoring to return a set of io.Readers or file.Openers instead of the full contents themselves (allow for optional buffering).
 }
 
 //  FileResolver knows how to get file.References for given string paths and globs
 type FileResolver interface {
+	// FilesByPath fetches a set of file references which have the given path (for an image, there may be multiple matches)
 	FilesByPath(paths ...file.Path) ([]file.Reference, error)
+	// FilesByGlob fetches a set of file references which the given glob matches
 	FilesByGlob(patterns ...string) ([]file.Reference, error)
+	// RelativeFileByPath fetches a single file at the given path relative to the layer squash of the given reference.
+	// This is helpful when attempting to find a file that is in the same layer or lower as another file.
+	RelativeFileByPath(reference file.Reference, path string) (*file.Reference, error)
 }
 
 // getImageResolver returns the appropriate resolve for a container image given the scope option

syft/scope/resolvers/all_layers_resolver.go

@@ -109,8 +109,23 @@ func (r *AllLayersResolver) FilesByGlob(patterns ...string) ([]file.Reference, e
 	return uniqueFiles, nil
 }
 
+func (r *AllLayersResolver) RelativeFileByPath(reference file.Reference, path string) (*file.Reference, error) {
+	entry, err := r.img.FileCatalog.Get(reference)
+	if err != nil {
+		return nil, err
+	}
+
+	return entry.Source.SquashedTree.File(file.Path(path)), nil
+}
+
 // MultipleFileContentsByRef returns the file contents for all file.References relative to the image. Note that a
 // file.Reference is a path relative to a particular layer.
 func (r *AllLayersResolver) MultipleFileContentsByRef(f ...file.Reference) (map[file.Reference]string, error) {
 	return r.img.MultipleFileContentsByRef(f...)
 }
+
+// FileContentsByRef fetches file contents for a single file reference, irregardless of the source layer.
+// If the path does not exist an error is returned.
+func (r *AllLayersResolver) FileContentsByRef(ref file.Reference) (string, error) {
+	return r.img.FileContentsByRef(ref)
+}

syft/scope/resolvers/directory_resolver.go

@@ -5,6 +5,7 @@ import (
 	"io/ioutil"
 	"os"
 	"path"
+	"path/filepath"
 
 	"github.com/anchore/stereoscope/pkg/file"
 	"github.com/anchore/syft/internal/log"
@@ -18,7 +19,7 @@ type DirectoryResolver struct {
 
 // Stringer to represent a directory path data source
 func (s DirectoryResolver) String() string {
-	return fmt.Sprintf("dir://%s", s.Path)
+	return fmt.Sprintf("dir:%s", s.Path)
 }
 
 // FilesByPath returns all file.References that match the given paths from the directory.
@@ -26,15 +27,19 @@ func (s DirectoryResolver) FilesByPath(userPaths ...file.Path) ([]file.Reference
 	var references = make([]file.Reference, 0)
 
 	for _, userPath := range userPaths {
-		resolvedPath := path.Join(s.Path, string(userPath))
-		_, err := os.Stat(resolvedPath)
+		userStrPath := string(userPath)
+
+		if filepath.IsAbs(userStrPath) {
+			// a path relative to root should be prefixed with the resolvers directory path, otherwise it should be left as is
+			userStrPath = path.Join(s.Path, userStrPath)
+		}
+		_, err := os.Stat(userStrPath)
 		if os.IsNotExist(err) {
 			continue
 		} else if err != nil {
-			log.Errorf("path (%s) is not valid: %v", resolvedPath, err)
+			log.Errorf("path (%s) is not valid: %v", userStrPath, err)
 		}
-		filePath := file.Path(resolvedPath)
-		references = append(references, file.NewFileReference(filePath))
+		references = append(references, file.NewFileReference(file.Path(userStrPath)))
 	}
 
 	return references, nil
@@ -75,6 +80,18 @@ func (s DirectoryResolver) FilesByGlob(patterns ...string) ([]file.Reference, er
 	return result, nil
 }
 
+func (s *DirectoryResolver) RelativeFileByPath(_ file.Reference, path string) (*file.Reference, error) {
+	paths, err := s.FilesByPath(file.Path(path))
+	if err != nil {
+		return nil, err
+	}
+	if len(paths) == 0 {
+		return nil, nil
+	}
+
+	return &paths[0], nil
+}
+
 // MultipleFileContentsByRef returns the file contents for all file.References relative a directory.
 func (s DirectoryResolver) MultipleFileContentsByRef(f ...file.Reference) (map[file.Reference]string, error) {
 	refContents := make(map[file.Reference]string)
@@ -82,9 +99,20 @@ func (s DirectoryResolver) MultipleFileContentsByRef(f ...file.Reference) (map[f
 		contents, err := fileContents(fileRef.Path)
 
 		if err != nil {
-			return refContents, fmt.Errorf("could not read contents of file: %s", fileRef.Path)
+			return nil, fmt.Errorf("could not read contents of file: %s", fileRef.Path)
 		}
 		refContents[fileRef] = string(contents)
 	}
 	return refContents, nil
 }
+
+// FileContentsByRef fetches file contents for a single file reference relative to a directory.
+// If the path does not exist an error is returned.
+func (s DirectoryResolver) FileContentsByRef(reference file.Reference) (string, error) {
+	contents, err := fileContents(reference.Path)
+	if err != nil {
+		return "", fmt.Errorf("could not read contents of file: %s", reference.Path)
+	}
+
+	return string(contents), nil
+}

syft/scope/resolvers/directory_resolver_test.go

@@ -1,7 +1,6 @@
 package resolvers
 
 import (
-	"path"
 	"testing"
 
 	"github.com/anchore/stereoscope/pkg/file"
@@ -10,24 +9,49 @@ import (
 func TestDirectoryResolver_FilesByPath(t *testing.T) {
 	cases := []struct {
 		name     string
+		root     string
 		input    string
+		expected string
 		refCount int
 	}{
 		{
-			name:     "finds a file",
-			input:    "image-symlinks/file-1.txt",
+			name:     "finds a file (relative)",
+			root:     "./test-fixtures/",
+			input:    "test-fixtures/image-symlinks/file-1.txt",
+			expected: "test-fixtures/image-symlinks/file-1.txt",
 			refCount: 1,
 		},
 		{
-			name:     "managed non-existing files",
-			input:    "image-symlinks/bogus.txt",
+			name:     "finds a file with relative indirection",
+			root:     "./test-fixtures/../test-fixtures",
+			input:    "test-fixtures/image-symlinks/file-1.txt",
+			expected: "test-fixtures/image-symlinks/file-1.txt",
+			refCount: 1,
+		},
+		{
+			// note: this is asserting the old behavior is not supported
+			name:     "relative lookup with wrong path fails",
+			root:     "./test-fixtures/",
+			input:    "image-symlinks/file-1.txt",
 			refCount: 0,
 		},
+		{
+			name:     "managed non-existing files (relative)",
+			root:     "./test-fixtures/",
+			input:    "test-fixtures/image-symlinks/bogus.txt",
+			refCount: 0,
+		},
+		{
+			name:     "finds a file (absolute)",
+			root:     "./test-fixtures/",
+			input:    "/image-symlinks/file-1.txt",
+			expected: "test-fixtures/image-symlinks/file-1.txt",
+			refCount: 1,
+		},
 	}
 	for _, c := range cases {
 		t.Run(c.name, func(t *testing.T) {
-			resolver := DirectoryResolver{"test-fixtures"}
-			expected := path.Join("test-fixtures", c.input)
+			resolver := DirectoryResolver{c.root}
 			refs, err := resolver.FilesByPath(file.Path(c.input))
 			if err != nil {
 				t.Fatalf("could not use resolver: %+v, %+v", err, refs)
@@ -38,8 +62,8 @@ func TestDirectoryResolver_FilesByPath(t *testing.T) {
 			}
 
 			for _, actual := range refs {
-				if actual.Path != file.Path(expected) {
-					t.Errorf("bad resolve path: '%s'!='%s'", actual.Path, c.input)
+				if actual.Path != file.Path(c.expected) {
+					t.Errorf("bad resolve path: '%s'!='%s'", actual.Path, c.expected)
 				}
 			}
 		})
@@ -54,17 +78,17 @@ func TestDirectoryResolver_MultipleFilesByPath(t *testing.T) {
 	}{
 		{
 			name:     "finds multiple files",
-			input:    []file.Path{file.Path("image-symlinks/file-1.txt"), file.Path("image-symlinks/file-2.txt")},
+			input:    []file.Path{file.Path("test-fixtures/image-symlinks/file-1.txt"), file.Path("test-fixtures/image-symlinks/file-2.txt")},
 			refCount: 2,
 		},
 		{
 			name:     "skips non-existing files",
-			input:    []file.Path{file.Path("image-symlinks/bogus.txt"), file.Path("image-symlinks/file-1.txt")},
+			input:    []file.Path{file.Path("test-fixtures/image-symlinks/bogus.txt"), file.Path("test-fixtures/image-symlinks/file-1.txt")},
 			refCount: 1,
 		},
 		{
 			name:     "does not return anything for non-existing directories",
-			input:    []file.Path{file.Path("non-existing/bogus.txt"), file.Path("non-existing/file-1.txt")},
+			input:    []file.Path{file.Path("test-fixtures/non-existing/bogus.txt"), file.Path("test-fixtures/non-existing/file-1.txt")},
 			refCount: 0,
 		},
 	}
@@ -93,17 +117,17 @@ func TestDirectoryResolver_MultipleFileContentsByRef(t *testing.T) {
 	}{
 		{
 			name:     "gets multiple file contents",
-			input:    []file.Path{file.Path("image-symlinks/file-1.txt"), file.Path("image-symlinks/file-2.txt")},
+			input:    []file.Path{file.Path("test-fixtures/image-symlinks/file-1.txt"), file.Path("test-fixtures/image-symlinks/file-2.txt")},
 			refCount: 2,
 		},
 		{
 			name:     "skips non-existing files",
-			input:    []file.Path{file.Path("image-symlinks/bogus.txt"), file.Path("image-symlinks/file-1.txt")},
+			input:    []file.Path{file.Path("test-fixtures/image-symlinks/bogus.txt"), file.Path("test-fixtures/image-symlinks/file-1.txt")},
 			refCount: 1,
 		},
 		{
 			name:     "does not return anything for non-existing directories",
-			input:    []file.Path{file.Path("non-existing/bogus.txt"), file.Path("non-existing/file-1.txt")},
+			input:    []file.Path{file.Path("test-fixtures/non-existing/bogus.txt"), file.Path("test-fixtures/non-existing/file-1.txt")},
 			refCount: 0,
 		},
 	}

syft/scope/resolvers/image_squash_resolver.go

@@ -73,8 +73,26 @@ func (r *ImageSquashResolver) FilesByGlob(patterns ...string) ([]file.Reference,
 	return uniqueFiles, nil
 }
 
+func (r *ImageSquashResolver) RelativeFileByPath(reference file.Reference, path string) (*file.Reference, error) {
+	paths, err := r.FilesByPath(file.Path(path))
+	if err != nil {
+		return nil, err
+	}
+	if len(paths) == 0 {
+		return nil, nil
+	}
+
+	return &paths[0], nil
+}
+
 // MultipleFileContentsByRef returns the file contents for all file.References relative to the image. Note that a
 // file.Reference is a path relative to a particular layer, in this case only from the squashed representation.
 func (r *ImageSquashResolver) MultipleFileContentsByRef(f ...file.Reference) (map[file.Reference]string, error) {
 	return r.img.MultipleFileContentsByRef(f...)
 }
+
+// FileContentsByRef fetches file contents for a single file reference, irregardless of the source layer.
+// If the path does not exist an error is returned.
+func (r *ImageSquashResolver) FileContentsByRef(ref file.Reference) (string, error) {
+	return r.img.FileContentsByRef(ref)
+}

syft/scope/scope_test.go

@@ -61,13 +61,13 @@ func TestDirectoryScope(t *testing.T) {
 		{
 			desc:       "path detected",
 			input:      "test-fixtures",
-			inputPaths: []file.Path{file.Path("path-detected")},
+			inputPaths: []file.Path{file.Path("test-fixtures/path-detected")},
 			expRefs:    1,
 		},
 		{
 			desc:       "no files-by-path detected",
 			input:      "test-fixtures",
-			inputPaths: []file.Path{file.Path("no-path-detected")},
+			inputPaths: []file.Path{file.Path("test-fixtures/no-path-detected")},
 			expRefs:    0,
 		},
 	}
@@ -105,13 +105,13 @@ func TestMultipleFileContentsByRefContents(t *testing.T) {
 		{
 			input:    "test-fixtures/path-detected",
 			desc:     "empty file",
-			path:     "empty",
+			path:     "test-fixtures/path-detected/empty",
 			expected: "",
 		},
 		{
 			input:    "test-fixtures/path-detected",
 			desc:     "file has contents",
-			path:     ".vimrc",
+			path:     "test-fixtures/path-detected/.vimrc",
 			expected: "\" A .vimrc file\n",
 		},
 	}
@@ -127,7 +127,7 @@ func TestMultipleFileContentsByRefContents(t *testing.T) {
 			}
 
 			if len(refs) != 1 {
-				t.Errorf("expected a single ref to be generated but got: %d", len(refs))
+				t.Fatalf("expected a single ref to be generated but got: %d", len(refs))
 			}
 			ref := refs[0]
 

test/integration/distro_test.go

@@ -1,5 +1,3 @@
-// +build integration
-
 package integration
 
 import (

test/integration/json_schema_test.go

@@ -1,5 +1,3 @@
-// +build integration
-
 package integration
 
 import (

test/integration/pkg_cases_test.go

@@ -1,5 +1,3 @@
-// +build integration
-
 package integration
 
 import "github.com/anchore/syft/syft/pkg"
@@ -20,6 +18,25 @@ var imageOnlyTestCases = []testCase{
 			"bundler": "2.1.4",
 		},
 	},
+	{
+		name:        "find npm package",
+		pkgType:     pkg.NpmPkg,
+		pkgLanguage: pkg.JavaScript,
+		pkgInfo: map[string]string{
+			"npm": "6.14.6",
+		},
+	},
+	{
+		name:        "find python egg & wheel packages",
+		pkgType:     pkg.PythonPkg,
+		pkgLanguage: pkg.Python,
+		pkgInfo: map[string]string{
+			"Pygments":     "2.6.1",
+			"requests":     "2.22.0",
+			"somerequests": "3.22.0",
+			"someotherpkg": "3.19.0",
+		},
+	},
 }
 
 var dirOnlyTestCases = []testCase{
@@ -81,6 +98,35 @@ var dirOnlyTestCases = []testCase{
 			"unicorn":              "4.8.3",
 		},
 	},
+	{
+		name:        "find javascript npm packages (yarn.lock & package-lock.json)",
+		pkgType:     pkg.NpmPkg,
+		pkgLanguage: pkg.JavaScript,
+		pkgInfo: map[string]string{
+			"@babel/code-frame": "7.10.4",
+			"get-stdin":         "8.0.0",
+		},
+	},
+	{
+		name:        "find python requirements.txt & setup.py package references",
+		pkgType:     pkg.PythonPkg,
+		pkgLanguage: pkg.Python,
+		pkgInfo: map[string]string{
+			// dir specific test cases
+			"flask":              "4.0.0",
+			"python-dateutil":    "2.8.1",
+			"python-swiftclient": "3.8.1",
+			"pytz":               "2019.3",
+			"jsonschema":         "2.6.0",
+			"passlib":            "1.7.2",
+			"mypy":               "v0.770",
+			// common to image and directory
+			"Pygments":     "2.6.1",
+			"requests":     "2.22.0",
+			"somerequests": "3.22.0",
+			"someotherpkg": "3.19.0",
+		},
+	},
 }
 
 var commonTestCases = []testCase{
@@ -116,62 +162,6 @@ var commonTestCases = []testCase{
 			"example-jenkins-plugin": "1.0-SNAPSHOT",
 		},
 	},
-	{
-		name:        "find python wheel packages",
-		pkgType:     pkg.WheelPkg,
-		pkgLanguage: pkg.Python,
-		pkgInfo: map[string]string{
-			"Pygments": "2.6.1",
-			"requests": "2.10.0",
-		},
-	},
-	{
-		name:        "find javascript npm packages",
-		pkgType:     pkg.NpmPkg,
-		pkgLanguage: pkg.JavaScript,
-		pkgInfo: map[string]string{
-			"get-stdin": "8.0.0",
-		},
-	},
-	{
-		name:        "find javascript yarn packages",
-		pkgType:     pkg.YarnPkg,
-		pkgLanguage: pkg.JavaScript,
-		pkgInfo: map[string]string{
-			"@babel/code-frame": "7.10.4",
-		},
-	},
-	{
-		name:        "find python egg packages",
-		pkgType:     pkg.EggPkg,
-		pkgLanguage: pkg.Python,
-		pkgInfo: map[string]string{
-			"requests": "2.22.0",
-			"otherpkg": "2.19.0",
-		},
-	},
-	{
-		name:        "find python requirements.txt packages",
-		pkgType:     pkg.PythonRequirementsPkg,
-		pkgLanguage: pkg.Python,
-		pkgInfo: map[string]string{
-			"flask":              "4.0.0",
-			"python-dateutil":    "2.8.1",
-			"python-swiftclient": "3.8.1",
-			"pytz":               "2019.3",
-			"jsonschema":         "2.6.0",
-			"passlib":            "1.7.2",
-			"pathlib":            "1.0.1",
-		},
-	},
-	{
-		name:        "find python setup.py packages",
-		pkgType:     pkg.PythonSetupPkg,
-		pkgLanguage: pkg.Python,
-		pkgInfo: map[string]string{
-			"mypy": "v0.770",
-		},
-	},
 	{
 
 		name:    "find apkdb packages",

test/integration/pkg_coverage_test.go

@@ -1,5 +1,3 @@
-// +build integration
-
 package integration
 
 import (
@@ -70,6 +68,7 @@ func TestPkgCoverageImage(t *testing.T) {
 			}
 
 			if pkgCount != len(c.pkgInfo) {
+				t.Logf("Discovered packages of type %+v", c.pkgType)
 				for a := range catalog.Enumerate(c.pkgType) {
 					t.Log("   ", a)
 				}
@@ -123,39 +122,39 @@ func TestPkgCoverageDirectory(t *testing.T) {
 	cases = append(cases, commonTestCases...)
 	cases = append(cases, dirOnlyTestCases...)
 
-	for _, c := range cases {
-		t.Run(c.name, func(t *testing.T) {
-			pkgCount := 0
+	for _, test := range cases {
+		t.Run(test.name, func(t *testing.T) {
+			actualPkgCount := 0
 
-			for a := range catalog.Enumerate(c.pkgType) {
+			for actualPkg := range catalog.Enumerate(test.pkgType) {
 
-				observedLanguages.Add(a.Language.String())
-				observedPkgs.Add(string(a.Type))
+				observedLanguages.Add(actualPkg.Language.String())
+				observedPkgs.Add(string(actualPkg.Type))
 
-				expectedVersion, ok := c.pkgInfo[a.Name]
+				expectedVersion, ok := test.pkgInfo[actualPkg.Name]
 				if !ok {
-					t.Errorf("unexpected package found: %s", a.Name)
+					t.Errorf("unexpected package found: %s", actualPkg.Name)
 				}
 
-				if expectedVersion != a.Version {
-					t.Errorf("unexpected package version (pkg=%s): %s", a.Name, a.Version)
+				if expectedVersion != actualPkg.Version {
+					t.Errorf("unexpected package version (pkg=%s): %s", actualPkg.Name, actualPkg.Version)
 				}
 
-				if a.Language != c.pkgLanguage {
-					t.Errorf("bad language (pkg=%+v): %+v", a.Name, a.Language)
+				if actualPkg.Language != test.pkgLanguage {
+					t.Errorf("bad language (pkg=%+v): %+v", actualPkg.Name, actualPkg.Language)
 				}
 
-				if a.Type != c.pkgType {
-					t.Errorf("bad package type (pkg=%+v): %+v", a.Name, a.Type)
+				if actualPkg.Type != test.pkgType {
+					t.Errorf("bad package type (pkg=%+v): %+v", actualPkg.Name, actualPkg.Type)
 				}
-				pkgCount++
+				actualPkgCount++
 			}
 
-			if pkgCount != len(c.pkgInfo) {
-				for a := range catalog.Enumerate(c.pkgType) {
-					t.Log("   ", a)
+			if actualPkgCount != len(test.pkgInfo) {
+				for actualPkg := range catalog.Enumerate(test.pkgType) {
+					t.Log("   ", actualPkg)
 				}
-				t.Fatalf("unexpected package count: %d!=%d", pkgCount, len(c.pkgInfo))
+				t.Fatalf("unexpected package count: %d!=%d", actualPkgCount, len(test.pkgInfo))
 			}
 
 		})

test/integration/regression_test.go

@@ -0,0 +1,38 @@
+// +build integration
+
+package integration
+
+import (
+	"testing"
+
+	"github.com/anchore/syft/syft/pkg"
+
+	"github.com/anchore/stereoscope/pkg/imagetest"
+	"github.com/anchore/syft/syft"
+	"github.com/anchore/syft/syft/scope"
+)
+
+func TestRegression212ApkBufferSize(t *testing.T) {
+	// This is a regression test for issue #212 (https://github.com/anchore/syft/issues/212) in which the apk db could
+	// not be processed due to a scanner buffer that was too small
+
+	fixtureImageName := "image-large-apk-data"
+	_, cleanup := imagetest.GetFixtureImage(t, "docker-archive", fixtureImageName)
+	tarPath := imagetest.GetFixtureImageTarPath(t, fixtureImageName)
+	defer cleanup()
+
+	catalog, _, _, err := syft.Catalog("docker-archive:"+tarPath, scope.SquashedScope)
+	if err != nil {
+		t.Fatalf("failed to catalog image: %+v", err)
+	}
+
+	expectedPkgs := 57
+	actualPkgs := 0
+	for range catalog.Enumerate(pkg.ApkPkg) {
+		actualPkgs += 1
+	}
+
+	if actualPkgs != expectedPkgs {
+		t.Errorf("unexpected number of APK packages: %d != %d", expectedPkgs, actualPkgs)
+	}
+}

test/integration/test-fixtures/image-large-apk-data/Dockerfile

@@ -0,0 +1,2 @@
+FROM alpine:latest
+RUN apk add tzdata vim alpine-sdk

test/integration/test-fixtures/image-pkg-coverage/javascript/package-json/package.json

@@ -0,0 +1,314 @@
+{
+  "version": "6.14.6",
+  "name": "npm",
+  "description": "a package manager for JavaScript",
+  "keywords": [
+    "install",
+    "modules",
+    "package manager",
+    "package.json"
+  ],
+  "preferGlobal": true,
+  "config": {
+    "publishtest": false
+  },
+  "homepage": "https://docs.npmjs.com/",
+  "author": "Isaac Z. Schlueter <i@izs.me> (http://blog.izs.me)",
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/npm/cli"
+  },
+  "bugs": {
+    "url": "https://npm.community/c/bugs"
+  },
+  "directories": {
+    "bin": "./bin",
+    "doc": "./doc",
+    "lib": "./lib",
+    "man": "./man"
+  },
+  "main": "./lib/npm.js",
+  "bin": {
+    "npm": "./bin/npm-cli.js",
+    "npx": "./bin/npx-cli.js"
+  },
+  "dependencies": {
+    "JSONStream": "^1.3.5",
+    "abbrev": "~1.1.1",
+    "ansicolors": "~0.3.2",
+    "ansistyles": "~0.1.3",
+    "aproba": "^2.0.0",
+    "archy": "~1.0.0",
+    "bin-links": "^1.1.7",
+    "bluebird": "^3.5.5",
+    "byte-size": "^5.0.1",
+    "cacache": "^12.0.3",
+    "call-limit": "^1.1.1",
+    "chownr": "^1.1.4",
+    "ci-info": "^2.0.0",
+    "cli-columns": "^3.1.2",
+    "cli-table3": "^0.5.1",
+    "cmd-shim": "^3.0.3",
+    "columnify": "~1.5.4",
+    "config-chain": "^1.1.12",
+    "detect-indent": "~5.0.0",
+    "detect-newline": "^2.1.0",
+    "dezalgo": "~1.0.3",
+    "editor": "~1.0.0",
+    "figgy-pudding": "^3.5.1",
+    "find-npm-prefix": "^1.0.2",
+    "fs-vacuum": "~1.2.10",
+    "fs-write-stream-atomic": "~1.0.10",
+    "gentle-fs": "^2.3.0",
+    "glob": "^7.1.6",
+    "graceful-fs": "^4.2.4",
+    "has-unicode": "~2.0.1",
+    "hosted-git-info": "^2.8.8",
+    "iferr": "^1.0.2",
+    "infer-owner": "^1.0.4",
+    "inflight": "~1.0.6",
+    "inherits": "^2.0.4",
+    "ini": "^1.3.5",
+    "init-package-json": "^1.10.3",
+    "is-cidr": "^3.0.0",
+    "json-parse-better-errors": "^1.0.2",
+    "lazy-property": "~1.0.0",
+    "libcipm": "^4.0.7",
+    "libnpm": "^3.0.1",
+    "libnpmaccess": "^3.0.2",
+    "libnpmhook": "^5.0.3",
+    "libnpmorg": "^1.0.1",
+    "libnpmsearch": "^2.0.2",
+    "libnpmteam": "^1.0.2",
+    "libnpx": "^10.2.2",
+    "lock-verify": "^2.1.0",
+    "lockfile": "^1.0.4",
+    "lodash._baseuniq": "~4.6.0",
+    "lodash.clonedeep": "~4.5.0",
+    "lodash.union": "~4.6.0",
+    "lodash.uniq": "~4.5.0",
+    "lodash.without": "~4.4.0",
+    "lru-cache": "^5.1.1",
+    "meant": "~1.0.1",
+    "mississippi": "^3.0.0",
+    "mkdirp": "^0.5.5",
+    "move-concurrently": "^1.0.1",
+    "node-gyp": "^5.1.0",
+    "nopt": "^4.0.3",
+    "normalize-package-data": "^2.5.0",
+    "npm-audit-report": "^1.3.2",
+    "npm-cache-filename": "~1.0.2",
+    "npm-install-checks": "^3.0.2",
+    "npm-lifecycle": "^3.1.4",
+    "npm-package-arg": "^6.1.1",
+    "npm-packlist": "^1.4.8",
+    "npm-pick-manifest": "^3.0.2",
+    "npm-profile": "^4.0.4",
+    "npm-registry-fetch": "^4.0.5",
+    "npm-user-validate": "~1.0.0",
+    "npmlog": "~4.1.2",
+    "once": "~1.4.0",
+    "opener": "^1.5.1",
+    "osenv": "^0.1.5",
+    "pacote": "^9.5.12",
+    "path-is-inside": "~1.0.2",
+    "promise-inflight": "~1.0.1",
+    "qrcode-terminal": "^0.12.0",
+    "query-string": "^6.8.2",
+    "qw": "~1.0.1",
+    "read": "~1.0.7",
+    "read-cmd-shim": "^1.0.5",
+    "read-installed": "~4.0.3",
+    "read-package-json": "^2.1.1",
+    "read-package-tree": "^5.3.1",
+    "readable-stream": "^3.6.0",
+    "readdir-scoped-modules": "^1.1.0",
+    "request": "^2.88.0",
+    "retry": "^0.12.0",
+    "rimraf": "^2.7.1",
+    "safe-buffer": "^5.1.2",
+    "semver": "^5.7.1",
+    "sha": "^3.0.0",
+    "slide": "~1.1.6",
+    "sorted-object": "~2.0.1",
+    "sorted-union-stream": "~2.1.3",
+    "ssri": "^6.0.1",
+    "stringify-package": "^1.0.1",
+    "tar": "^4.4.13",
+    "text-table": "~0.2.0",
+    "tiny-relative-date": "^1.3.0",
+    "uid-number": "0.0.6",
+    "umask": "~1.1.0",
+    "unique-filename": "^1.1.1",
+    "unpipe": "~1.0.0",
+    "update-notifier": "^2.5.0",
+    "uuid": "^3.3.3",
+    "validate-npm-package-license": "^3.0.4",
+    "validate-npm-package-name": "~3.0.0",
+    "which": "^1.3.1",
+    "worker-farm": "^1.7.0",
+    "write-file-atomic": "^2.4.3"
+  },
+  "bundleDependencies": [
+    "abbrev",
+    "ansicolors",
+    "ansistyles",
+    "aproba",
+    "archy",
+    "bin-links",
+    "bluebird",
+    "byte-size",
+    "cacache",
+    "call-limit",
+    "chownr",
+    "ci-info",
+    "cli-columns",
+    "cli-table3",
+    "cmd-shim",
+    "columnify",
+    "config-chain",
+    "debuglog",
+    "detect-indent",
+    "detect-newline",
+    "dezalgo",
+    "editor",
+    "figgy-pudding",
+    "find-npm-prefix",
+    "fs-vacuum",
+    "fs-write-stream-atomic",
+    "gentle-fs",
+    "glob",
+    "graceful-fs",
+    "has-unicode",
+    "hosted-git-info",
+    "iferr",
+    "imurmurhash",
+    "infer-owner",
+    "inflight",
+    "inherits",
+    "ini",
+    "init-package-json",
+    "is-cidr",
+    "json-parse-better-errors",
+    "JSONStream",
+    "lazy-property",
+    "libcipm",
+    "libnpm",
+    "libnpmaccess",
+    "libnpmhook",
+    "libnpmorg",
+    "libnpmsearch",
+    "libnpmteam",
+    "libnpx",
+    "lock-verify",
+    "lockfile",
+    "lodash._baseindexof",
+    "lodash._baseuniq",
+    "lodash._bindcallback",
+    "lodash._cacheindexof",
+    "lodash._createcache",
+    "lodash._getnative",
+    "lodash.clonedeep",
+    "lodash.restparam",
+    "lodash.union",
+    "lodash.uniq",
+    "lodash.without",
+    "lru-cache",
+    "meant",
+    "mississippi",
+    "mkdirp",
+    "move-concurrently",
+    "node-gyp",
+    "nopt",
+    "normalize-package-data",
+    "npm-audit-report",
+    "npm-cache-filename",
+    "npm-install-checks",
+    "npm-lifecycle",
+    "npm-package-arg",
+    "npm-packlist",
+    "npm-pick-manifest",
+    "npm-profile",
+    "npm-registry-fetch",
+    "npm-user-validate",
+    "npmlog",
+    "once",
+    "opener",
+    "osenv",
+    "pacote",
+    "path-is-inside",
+    "promise-inflight",
+    "qrcode-terminal",
+    "query-string",
+    "qw",
+    "read-cmd-shim",
+    "read-installed",
+    "read-package-json",
+    "read-package-tree",
+    "read",
+    "readable-stream",
+    "readdir-scoped-modules",
+    "request",
+    "retry",
+    "rimraf",
+    "safe-buffer",
+    "semver",
+    "sha",
+    "slide",
+    "sorted-object",
+    "sorted-union-stream",
+    "ssri",
+    "stringify-package",
+    "tar",
+    "text-table",
+    "tiny-relative-date",
+    "uid-number",
+    "umask",
+    "unique-filename",
+    "unpipe",
+    "update-notifier",
+    "uuid",
+    "validate-npm-package-license",
+    "validate-npm-package-name",
+    "which",
+    "worker-farm",
+    "write-file-atomic"
+  ],
+  "devDependencies": {
+    "deep-equal": "^1.0.1",
+    "get-stream": "^4.1.0",
+    "licensee": "^7.0.3",
+    "marked": "^0.6.3",
+    "marked-man": "^0.6.0",
+    "npm-registry-couchapp": "^2.7.4",
+    "npm-registry-mock": "^1.3.1",
+    "require-inject": "^1.4.4",
+    "sprintf-js": "^1.1.2",
+    "standard": "^11.0.1",
+    "tacks": "^1.3.0",
+    "tap": "^12.7.0",
+    "tar-stream": "^2.1.0"
+  },
+  "scripts": {
+    "dumpconf": "env | grep npm | sort | uniq",
+    "prepare": "node bin/npm-cli.js rebuild && node bin/npm-cli.js --no-audit --no-timing prune --prefix=. --no-global && rimraf test/*/*/node_modules && make -j4 mandocs",
+    "preversion": "bash scripts/update-authors.sh && git add AUTHORS && git commit -m \"update AUTHORS\" || true",
+    "licenses": "licensee --production --errors-only",
+    "tap": "tap -J --timeout 300 --no-esm",
+    "tap-cover": "tap -J --nyc-arg=--cache --coverage --timeout 600 --no-esm",
+    "lint": "standard",
+    "pretest": "npm run lint",
+    "test": "npm run test-tap --",
+    "test:nocleanup": "NO_TEST_CLEANUP=1 npm run test --",
+    "sudotest": "sudo npm run tap -- \"test/tap/*.js\"",
+    "sudotest:nocleanup": "sudo NO_TEST_CLEANUP=1 npm run tap -- \"test/tap/*.js\"",
+    "posttest": "rimraf test/npm_cache*",
+    "test-coverage": "npm run tap-cover -- \"test/tap/*.js\" \"test/network/*.js\"",
+    "test-tap": "npm run tap -- \"test/tap/*.js\" \"test/network/*.js\"",
+    "test-node": "tap --timeout 240 \"test/tap/*.js\" \"test/network/*.js\""
+  },
+  "license": "Artistic-2.0",
+  "engines": {
+    "node": "6 >=6.2.0 || 8 || >=9.3.0"
+  }
+}

test/integration/test-fixtures/image-pkg-coverage/python/dist-info/RECORD

@@ -0,0 +1,5 @@
+../../../bin/pygmentize,sha256=dDhv_U2jiCpmFQwIRHpFRLAHUO4R1jIJPEvT_QYTFp8,220
+Pygments-2.6.1.dist-info/AUTHORS,sha256=PVpa2_Oku6BGuiUvutvuPnWGpzxqFy2I8-NIrqCvqUY,8449
+Pygments-2.6.1.dist-info/RECORD,,
+pygments/__pycache__/__init__.cpython-38.pyc,,
+pygments/util.py,sha256=586xXHiJGGZxqk5PMBu3vBhE68DLuAe5MBARWrSPGxA,10778

test/integration/test-fixtures/image-pkg-coverage/python/dist-info/top_level.txt

@@ -0,0 +1 @@
+top-level-pkg

test/integration/test-fixtures/image-pkg-coverage/python/egg-info/top_level.txt

@@ -0,0 +1 @@
+top-level-pkg

test/integration/test-fixtures/image-pkg-coverage/python/requires/requirements-dev.txt

@@ -1,3 +1,2 @@
 jsonschema==2.6.0
-passlib==1.7.2
-pathlib==1.0.1
+passlib==1.7.2

test/integration/test-fixtures/image-pkg-coverage/python/someotherpkg-3.19.0-py3.8.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
-Name: otherpkg
-Version: 2.19.0
+Name: someotherpkg
+Version: 3.19.0
 Summary: Python HTTP for Humans.
 Home-page: http://python-requests.org
 Author: Kenneth Reitz

test/integration/test-fixtures/image-pkg-coverage/python/someotherpkg-3.19.0-py3.8.egg-info/top_level.txt

@@ -0,0 +1 @@
+top-level-pkg

test/integration/test-fixtures/image-pkg-coverage/python/somerequests-3.22.0.dist-info/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
-Name: requests
-Version: 2.10.0
+Name: somerequests
+Version: 3.22.0
 Summary: stuff
 Home-page: stuff
 Author: Georg Brandl

test/integration/test-fixtures/image-pkg-coverage/python/somerequests-3.22.0.dist-info/top_level.txt

@@ -0,0 +1 @@
+top-level-pkg