Add a User field into Document to make it easy to aggregate on github

user level.

api/internal/crawl/doc/doc_test.go

@@ -216,22 +216,26 @@ resources:
 					RepositoryURL: "sigs.k8s.io/kustomize",
 					FilePath:      "some/path/to/base",
 					FileType:      "resource",
+					User:          "sigs.k8s.io",
 				},
 				{
 					RepositoryURL: "sigs.k8s.io/kustomize",
 					FilePath:      "some/path/to/otherbase",
 					FileType:      "resource",
+					User:          "sigs.k8s.io",
 				},
 				{
 					RepositoryURL: "sigs.k8s.io/kustomize",
 					FilePath:      "some/path/to/kdir/file.yaml",
 					FileType:      "resource",
+					User:          "sigs.k8s.io",
 				},
 				{
 					RepositoryURL: "https://github.com/kubernetes-sigs/kustomize",
 					FilePath:      "examples/helloWorld",
 					DefaultBranch: "v3.1.0",
 					FileType:      "resource",
+					User:          "kubernetes-sigs",
 				},
 			},
 		},
@@ -317,11 +321,13 @@ transformers:
 					RepositoryURL: "sigs.k8s.io/kustomize",
 					FilePath:      "some/path/to/kdir/gen.yaml",
 					FileType:      "generator",
+					User:          "sigs.k8s.io",
 				},
 				{
 					RepositoryURL: "sigs.k8s.io/kustomize",
 					FilePath:      "some/path/to/kdir/file.yaml",
 					FileType:      "resource",
+					User:          "sigs.k8s.io",
 				},
 			},
 		},
@@ -352,16 +358,19 @@ transformers:
 					RepositoryURL: "sigs.k8s.io/kustomize",
 					FilePath:      "some/path/to/kdir/tr.yaml",
 					FileType:      "transformer",
+					User:          "sigs.k8s.io",
 				},
 				{
 					RepositoryURL: "sigs.k8s.io/kustomize",
 					FilePath:      "some/path/to/kdir/gen.yaml",
 					FileType:      "generator",
+					User:          "sigs.k8s.io",
 				},
 				{
 					RepositoryURL: "sigs.k8s.io/kustomize",
 					FilePath:      "some/path/to/kdir/file.yaml",
 					FileType:      "resource",
+					User:          "sigs.k8s.io",
 				},
 			},
 		},

api/internal/crawl/doc/docname.go

@@ -11,7 +11,10 @@ import (
 )
 
 type Document struct {
-	RepositoryURL string     `json:"repositoryUrl,omitempty"`
+	RepositoryURL string `json:"repositoryUrl,omitempty"`
+	// User makes it easy to aggregate data in the user level instead
+	// of the repository level
+	User          string     `json:"user,omitempty"`
 	FilePath      string     `json:"filePath,omitempty"`
 	DefaultBranch string     `json:"defaultBranch,omitempty"`
 	DocumentData  string     `json:"document,omitempty"`
@@ -30,6 +33,7 @@ func (doc *Document) GetDocument() *Document {
 func (doc *Document) Copy() *Document {
 	return &Document{
 		RepositoryURL: doc.RepositoryURL,
+		User:          doc.User,
 		FilePath:      doc.FilePath,
 		DefaultBranch: doc.DefaultBranch,
 		DocumentData:  doc.DocumentData,
@@ -56,6 +60,7 @@ func (doc *Document) FromRelativePath(newFile string) (Document, error) {
 			RepositoryURL: repoSpec.Host + path.Clean(repoSpec.OrgRepo),
 			FilePath:      path.Clean(repoSpec.Path),
 			DefaultBranch: repoSpec.Ref,
+			User:          UserName(repoSpec.Host + path.Clean(repoSpec.OrgRepo)),
 		}, nil
 	}
 	// else document is probably relative path.
@@ -63,6 +68,7 @@ func (doc *Document) FromRelativePath(newFile string) (Document, error) {
 	ret := Document{
 		RepositoryURL: doc.RepositoryURL,
 		DefaultBranch: doc.DefaultBranch,
+		User:          UserName(doc.RepositoryURL),
 	}
 	ogDir, _ := path.Split(doc.FilePath)
 
@@ -87,13 +93,7 @@ func (doc *Document) ID() string {
 }
 
 func (doc *Document) RepositoryFullName() string {
-	url := strings.TrimRight(doc.RepositoryURL, "/")
-
-	gitPrefix := "git@github.com:"
-	if strings.HasPrefix(url, gitPrefix) {
-		url = url[len(gitPrefix):]
-	}
-
+	url := TrimUrl(doc.RepositoryURL)
 	sections := strings.Split(url, "/")
 	l := len(sections)
 	if l < 2 {
@@ -101,3 +101,24 @@ func (doc *Document) RepositoryFullName() string {
 	}
 	return path.Join(sections[l-2], sections[l-1])
 }
+
+// TrimUrl removes all the trailing slashes and the "git@github.com:" prefix (if exists).
+func TrimUrl(s string) string {
+	url := strings.TrimRight(s, "/")
+
+	gitPrefix := "git@github.com:"
+	if strings.HasPrefix(url, gitPrefix) {
+		url = url[len(gitPrefix):]
+	}
+	return url
+}
+
+func UserName(repositoryURL string) string {
+	url := TrimUrl(repositoryURL)
+	sections := strings.Split(url, "/")
+	l := len(sections)
+	if l < 2 {
+		return url
+	}
+	return sections[l-2]
+}

api/internal/crawl/doc/docname_test.go

@@ -28,6 +28,7 @@ func TestFromRelativePath(t *testing.T) {
 						RepositoryURL: "example.com/repo",
 						FilePath:      "path/to/other/file/resource.yaml",
 						DefaultBranch: "master",
+						User:          "example.com",
 					},
 				},
 				{
@@ -36,6 +37,7 @@ func TestFromRelativePath(t *testing.T) {
 						RepositoryURL: "example.com/repo",
 						FilePath:      "path/to/other/file/patch.yaml",
 						DefaultBranch: "master",
+						User:          "example.com",
 					},
 				},
 				{
@@ -44,6 +46,7 @@ func TestFromRelativePath(t *testing.T) {
 						RepositoryURL: "example.com/repo",
 						FilePath:      "path/to/file/service.yaml",
 						DefaultBranch: "master",
+						User:          "example.com",
 					},
 				},
 			},
@@ -109,3 +112,39 @@ func TestDocument_RepositoryFullName(t *testing.T) {
 		}
 	}
 }
+
+func TestDocument_UserName(t *testing.T) {
+	testCases := []struct {
+		repositoryURL    string
+		expectedUserName string
+	}{
+		{
+			repositoryURL:    "https://github.com/user/repo",
+			expectedUserName: "user",
+		},
+		{
+			repositoryURL:    "https://github.com//user/repo////",
+			expectedUserName: "user",
+		},
+		{
+			repositoryURL:    "repo/",
+			expectedUserName: "repo",
+		},
+		{
+			repositoryURL:    "",
+			expectedUserName: "",
+		},
+		{
+			repositoryURL:    "git@github.com:user/repo",
+			expectedUserName: "user",
+		},
+	}
+
+	for _, tc := range testCases {
+		returnedUserName := UserName(tc.repositoryURL)
+		if returnedUserName != tc.expectedUserName {
+			t.Errorf("UserName expected %s, got %s",
+				tc.expectedUserName, returnedUserName)
+		}
+	}
+}