Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Introduce indexed embedded CPE dictionary #1897

Merged
merged 5 commits into from
Jul 21, 2023
Merged
Show file tree
Hide file tree
Changes from 3 commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
13 changes: 11 additions & 2 deletions Makefile
Original file line number Diff line number Diff line change
Expand Up @@ -312,6 +312,11 @@ generate-license-list: ## Generate an updated spdx license list

## Build-related targets #################################

.PHONY: cpe-index
cpe-index: ## Build the CPE index
$(call title,Building CPE index)
go run ./syft/pkg/cataloger/common/cpe/dictionary/index-generator -o ./syft/pkg/cataloger/common/cpe/cpe-index.json

.PHONY: build
build: $(SNAPSHOT_DIR) ## Build release snapshot binaries and packages

Expand All @@ -338,7 +343,7 @@ release:
@.github/scripts/trigger-release.sh

.PHONY: ci-release
ci-release: ci-check clean-dist $(CHANGELOG)
ci-release: ci-check clean-dist $(CHANGELOG) cpe-index
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

actually, generating new data during the release step isn't the right place for this. Though it answers the question of how this gets refreshed, it introduces potentially breaking changes right at the release, bypassing testing.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yeah, I think we can just take that step out, lean on the index that's checked into the repo, and periodically regenerate what's in the repo. I've seen other projects take a similar approach with generated/embedded data. How does that sound to you?

$(call title,Publishing release artifacts)

# create a config with the dist dir overridden
Expand All @@ -361,9 +366,13 @@ ci-check:
## Cleanup targets #################################

.PHONY: clean
clean: clean-dist clean-snapshot clean-test-image-cache ## Remove previous builds, result reports, and test cache
clean: clean-dist clean-snapshot clean-test-image-cache clean-cpe-index ## Remove previous builds, result reports, and test cache
$(call safe_rm_rf_children,$(TEMP_DIR))

.PHONY: clean-cpe-index
clean-cpe-index: ## Remove the CPE index
$(call safe_rm_rf,./syft/pkg/cataloger/common/cpe/cpe-index.json)

.PHONY: clean-snapshot
clean-snapshot:
$(call safe_rm_rf,$(SNAPSHOT_DIR))
Expand Down
9 changes: 8 additions & 1 deletion syft/pkg/cataloger/catalog.go
Original file line number Diff line number Diff line change
Expand Up @@ -76,7 +76,14 @@ func runCataloger(cataloger pkg.Cataloger, resolver file.Resolver) (catalogerRes
for _, p := range packages {
// generate CPEs (note: this is excluded from package ID, so is safe to mutate)
// we might have binary classified CPE already with the package so we want to append here
p.CPEs = append(p.CPEs, cpe.Generate(p)...)

dictionaryCPE, ok := cpe.DictionaryFind(p)
if ok {
log.Debugf("used CPE dictionary to find CPE for %s package %q: %s", p.Type, p.Name, dictionaryCPE.BindToFmtString())
p.CPEs = append(p.CPEs, dictionaryCPE)
} else {
p.CPEs = append(p.CPEs, cpe.Generate(p)...)
}

// if we were not able to identify the language we have an opportunity
// to try and get this value from the PURL. Worst case we assert that
Expand Down
1,293 changes: 1,293 additions & 0 deletions syft/pkg/cataloger/common/cpe/cpe-index.json

Large diffs are not rendered by default.

230 changes: 230 additions & 0 deletions syft/pkg/cataloger/common/cpe/dictionary/index-generator/generate.go
Original file line number Diff line number Diff line change
@@ -0,0 +1,230 @@
package main

import (
"compress/gzip"
"encoding/json"
"encoding/xml"
"fmt"
"io"
"log"
"strings"

"github.com/facebookincubator/nvdtools/wfn"
"golang.org/x/exp/slices"

"github.com/anchore/syft/syft/pkg/cataloger/common/cpe/dictionary"
)

func generateIndexedDictionaryJSON(rawGzipData io.Reader) ([]byte, error) {
gzipReader, err := gzip.NewReader(rawGzipData)
if err != nil {
return nil, fmt.Errorf("unable to decompress CPE dictionary: %w", err)
}
defer gzipReader.Close()

// Read XML data
data, err := io.ReadAll(gzipReader)
if err != nil {
return nil, fmt.Errorf("unable to read CPE dictionary: %w", err)
}

// Unmarshal XML
var cpeList CpeList
if err := xml.Unmarshal(data, &cpeList); err != nil {
return nil, fmt.Errorf("unable to unmarshal CPE dictionary XML: %w", err)
}

// Filter out data that's not applicable here
cpeList = filterCpeList(cpeList)

// Create indexed dictionary to help with looking up CPEs
indexedDictionary := indexCPEList(cpeList)

// Convert to JSON
jsonData, err := json.MarshalIndent(indexedDictionary, "", " ")
if err != nil {
return nil, fmt.Errorf("unable to marshal CPE dictionary to JSON: %w", err)
}
return jsonData, nil
}

// filterCpeList removes CPE items that are not applicable to software packages.
func filterCpeList(cpeList CpeList) CpeList {
var processedCpeList CpeList

seen := make(map[string]struct{})

for _, cpeItem := range cpeList.CpeItems {
// Skip CPE items that don't have any references.
if len(cpeItem.References) == 0 {
continue
}

// Skip CPE items where the CPE URI doesn't meet our criteria.
parsedName, err := wfn.Parse(cpeItem.Name)
if err != nil {
log.Printf("unable to parse CPE URI %q: %s", cpeItem.Name, err)
}

if slices.Contains([]string{"h", "o"}, parsedName.Part) {
continue
}

normalizedName := normalizeCPE(parsedName).BindToURI()
if _, ok := seen[normalizedName]; ok {
continue
}
seen[normalizedName] = struct{}{}
cpeItem.Name = normalizedName

parsedCPE, err := wfn.Parse(cpeItem.Cpe23Item.Name)
if err != nil {
log.Printf("unable to parse CPE value %q: %s", cpeItem.Cpe23Item.Name, err)
}

cpeItem.Cpe23Item.Name = normalizeCPE(parsedCPE).BindToFmtString()

processedCpeList.CpeItems = append(processedCpeList.CpeItems, cpeItem)
}

return processedCpeList
}

// normalizeCPE removes the version and update parts of a CPE.
func normalizeCPE(cpe *wfn.Attributes) *wfn.Attributes {
cpeCopy := *cpe

cpeCopy.Version = ""
cpeCopy.Update = ""

return &cpeCopy
}

const (
prefixForNPMPackages = "https://www.npmjs.com/package/"
prefixForRubyGems = "https://rubygems.org/gems/"
prefixForRubyGemsHTTP = "http://rubygems.org/gems/"
prefixForNativeRubyGems = "https://github.com/ruby/"
prefixForPyPIPackages = "https://pypi.org/project/"
prefixForJenkinsPlugins = "https://github.com/jenkinsci/"
prefixForRustCrates = "https://crates.io/crates/"
)

// indexCPEList creates an index of CPEs by ecosystem.
func indexCPEList(list CpeList) *dictionary.Indexed {
wagoodman marked this conversation as resolved.
Show resolved Hide resolved
indexed := &dictionary.Indexed{
EcosystemPackages: make(map[string]dictionary.Packages),
}

for _, cpeItem := range list.CpeItems {
cpeItemName := cpeItem.Cpe23Item.Name

for _, reference := range cpeItem.References {
ref := reference.Reference.Href

switch {
case strings.HasPrefix(ref, prefixForNPMPackages):
addEntryForNPMPackage(indexed, ref, cpeItemName)

case strings.HasPrefix(ref, prefixForRubyGems), strings.HasPrefix(ref, prefixForRubyGemsHTTP):
addEntryForRubyGem(indexed, ref, cpeItemName)

case strings.HasPrefix(ref, prefixForNativeRubyGems):
addEntryForNativeRubyGem(indexed, ref, cpeItemName)

case strings.HasPrefix(ref, prefixForPyPIPackages):
addEntryForPyPIPackage(indexed, ref, cpeItemName)

case strings.HasPrefix(ref, prefixForJenkinsPlugins):
// It _might_ be a jenkins plugin!
addEntryForJenkinsPlugin(indexed, ref, cpeItemName)

case strings.HasPrefix(ref, prefixForRustCrates):
addEntryForRustCrate(indexed, ref, cpeItemName)
}
}
}

return indexed
}

func addEntryForRustCrate(indexed *dictionary.Indexed, ref string, cpeItemName string) {
// Prune off the non-package-name parts of the URL
ref = strings.TrimPrefix(ref, prefixForRustCrates)
ref = strings.Split(ref, "/")[0]

if _, ok := indexed.EcosystemPackages[dictionary.EcosystemRustCrates]; !ok {
indexed.EcosystemPackages[dictionary.EcosystemRustCrates] = make(dictionary.Packages)
}

indexed.EcosystemPackages[dictionary.EcosystemRustCrates][ref] = cpeItemName
}

func addEntryForJenkinsPlugin(indexed *dictionary.Indexed, ref string, cpeItemName string) {
// Prune off the non-package-name parts of the URL
ref = strings.TrimPrefix(ref, prefixForJenkinsPlugins)
ref = strings.Split(ref, "/")[0]

if !strings.HasSuffix(ref, "-plugin") {
// It's not a jenkins plugin!
return
}

ref = strings.TrimSuffix(ref, "-plugin")

if _, ok := indexed.EcosystemPackages[dictionary.EcosystemJenkinsPlugins]; !ok {
indexed.EcosystemPackages[dictionary.EcosystemJenkinsPlugins] = make(dictionary.Packages)
}

indexed.EcosystemPackages[dictionary.EcosystemJenkinsPlugins][ref] = cpeItemName
}

func addEntryForPyPIPackage(indexed *dictionary.Indexed, ref string, cpeItemName string) {
// Prune off the non-package-name parts of the URL
ref = strings.TrimPrefix(ref, prefixForPyPIPackages)
ref = strings.Split(ref, "/")[0]

if _, ok := indexed.EcosystemPackages[dictionary.EcosystemPyPI]; !ok {
indexed.EcosystemPackages[dictionary.EcosystemPyPI] = make(dictionary.Packages)
}

indexed.EcosystemPackages[dictionary.EcosystemPyPI][ref] = cpeItemName
}

func addEntryForNativeRubyGem(indexed *dictionary.Indexed, ref string, cpeItemName string) {
// Prune off the non-package-name parts of the URL
ref = strings.TrimPrefix(ref, prefixForNativeRubyGems)
ref = strings.Split(ref, "/")[0]

if _, ok := indexed.EcosystemPackages[dictionary.EcosystemRubyGems]; !ok {
indexed.EcosystemPackages[dictionary.EcosystemRubyGems] = make(dictionary.Packages)
}

indexed.EcosystemPackages[dictionary.EcosystemRubyGems][ref] = cpeItemName
}

func addEntryForRubyGem(indexed *dictionary.Indexed, ref string, cpeItemName string) {
// Prune off the non-package-name parts of the URL
ref = strings.TrimPrefix(ref, prefixForRubyGems)
ref = strings.TrimPrefix(ref, prefixForRubyGemsHTTP)
ref = strings.Split(ref, "/")[0]

if _, ok := indexed.EcosystemPackages[dictionary.EcosystemRubyGems]; !ok {
indexed.EcosystemPackages[dictionary.EcosystemRubyGems] = make(dictionary.Packages)
}

indexed.EcosystemPackages[dictionary.EcosystemRubyGems][ref] = cpeItemName
}

func addEntryForNPMPackage(indexed *dictionary.Indexed, ref string, cpeItemName string) {
// Prune off the non-package-name parts of the URL
ref = strings.Split(ref, "/v/")[0]
ref = strings.Split(ref, "?")[0]
ref = strings.TrimPrefix(ref, prefixForNPMPackages)

if _, ok := indexed.EcosystemPackages[dictionary.EcosystemNPM]; !ok {
indexed.EcosystemPackages[dictionary.EcosystemNPM] = make(dictionary.Packages)
}

indexed.EcosystemPackages[dictionary.EcosystemNPM][ref] = cpeItemName
}
Loading
Loading