Modify CPE vendor candidate generation approach (#484)

* consider additional vendor candidates for ruby, python, rpm, npm, and java Signed-off-by: Alex Goodman <alex.goodman@anchore.com> * add java pom.xml processing Signed-off-by: Alex Goodman <alex.goodman@anchore.com> * allow for downstream transform control in cpe generation processing Signed-off-by: Alex Goodman <alex.goodman@anchore.com> * migrate CPE generation logic to dedicated package Signed-off-by: Alex Goodman <alex.goodman@anchore.com> * split java manifest groupID extraction into two tiers Signed-off-by: Alex Goodman <alex.goodman@anchore.com> * extract groupID from pom parent project during CPE generation Signed-off-by: Alex Goodman <alex.goodman@anchore.com> * update java groupID processing tests to cover multi-tier approach Signed-off-by: Alex Goodman <alex.goodman@anchore.com> * fix constructor names for cpe.fieldCandidate Signed-off-by: Alex Goodman <alex.goodman@anchore.com> * rename helper function to startsWithTopLevelDomain Signed-off-by: Alex Goodman <alex.goodman@anchore.com> * add nil changes for java manifest sections Signed-off-by: Alex Goodman <alex.goodman@anchore.com> * update comment to reflect parsing maven files Signed-off-by: Alex Goodman <alex.goodman@anchore.com> * split out java description parsing Signed-off-by: Alex Goodman <alex.goodman@anchore.com> * split out pom parent processing Signed-off-by: Alex Goodman <alex.goodman@anchore.com> * simplify vendorsFromGroupIDs and associated tests Signed-off-by: Alex Goodman <alex.goodman@anchore.com> * simplify test type for vendorsFromGroupIDs Signed-off-by: Alex Goodman <alex.goodman@anchore.com> * copy candidate varidations to new instances Signed-off-by: Alex Goodman <alex.goodman@anchore.com> * rename CPE generation string util functions Signed-off-by: Alex Goodman <alex.goodman@anchore.com> * add an explanation around fieldCandidate Signed-off-by: Alex Goodman <alex.goodman@anchore.com> * simplify type for the cpe.fieldCandidateSet Signed-off-by: Alex Goodman <alex.goodman@anchore.com> * make CPE filter function names more readable Signed-off-by: Alex Goodman <alex.goodman@anchore.com> * update groupIDsFromJavaManifest to use a guard clause Signed-off-by: Alex Goodman <alex.goodman@anchore.com> * extract groupID extraction from artifactID fields into a separate function Signed-off-by: Alex Goodman <alex.goodman@anchore.com> * bump goreleaser version to combat failure Signed-off-by: Alex Goodman <alex.goodman@anchore.com>
2026-11-04 16:36:40 +01:00 · 2021-09-03 14:21:25 -04:00
parent 0799fd9d46
commit abbba3fc19
37 changed files with 3208 additions and 1763 deletions
@@ -0,0 +1,59 @@
+package cpe
+
+import (
+	"net/url"
+	"strings"
+)
+
+// candidateProductForGo attempts to find a single product name in a best-effort attempt. This implementation prefers
+// to return no vendor over returning potentially nonsensical results.
+func candidateProductForGo(name string) string {
+	// note: url.Parse requires a scheme for correct processing, which a golang module will not have, so one is provided.
+	u, err := url.Parse("http://" + name)
+	if err != nil {
+		return ""
+	}
+
+	cleanPath := strings.Trim(u.Path, "/")
+	pathElements := strings.Split(cleanPath, "/")
+
+	switch u.Host {
+	case "golang.org", "gopkg.in":
+		return cleanPath
+	case "google.golang.org":
+		return pathElements[0]
+	}
+
+	if len(pathElements) < 2 {
+		return ""
+	}
+
+	return pathElements[1]
+}
+
+// candidateVendorForGo attempts to find a single vendor name in a best-effort attempt. This implementation prefers
+// to return no vendor over returning potentially nonsensical results.
+func candidateVendorForGo(name string) string {
+	// note: url.Parse requires a scheme for correct processing, which a golang module will not have, so one is provided.
+	u, err := url.Parse("http://" + name)
+	if err != nil {
+		return ""
+	}
+
+	cleanPath := strings.Trim(u.Path, "/")
+
+	switch u.Host {
+	case "google.golang.org":
+		return "google"
+	case "golang.org":
+		return "golang"
+	case "gopkg.in":
+		return ""
+	}
+
+	pathElements := strings.Split(cleanPath, "/")
+	if len(pathElements) < 2 {
+		return ""
+	}
+	return pathElements[0]
+}