EC-651 - Add sub-directory checkout functionality.

This commit adds the ability to specify a sub-directory within a git repo
that should be checked out. This process checks out the repository,
verifies that the sub-directory exists within the repo worktree, copies
the sub-directory to the provided destination and removes the
checked-out repository.

Signed-off-by: Rob Nester <rnester@redhat.com>

common.go

@@ -26,7 +26,7 @@ func (t URIType) String() string {
 	return [...]string{"GitURI", "HTTPURI", "FileURI", "Unknown"}[t]
 }
 
-// expandTilde expands a leading tilde in the file path to the user's home directory
+// ExpandTilde expands a leading tilde in the file path to the user's home directory
 func ExpandTilde(path string) string {
 	if strings.HasPrefix(path, "~/") {
 		homeDir, err := GetHomeDir()
@@ -51,8 +51,12 @@ func ClassifyURI(input string) (URIType, error) {
 		return HTTPURI, nil
 	}
 
+	if strings.HasPrefix(input, "github.com") || strings.HasPrefix(input, "gitlab.com") {
+		return GitURI, nil
+	}
+
 	// Regular expression for Git URIs
-	gitURIPattern := regexp.MustCompile(`^(git@[\w\.\-]+:[\w\.\-]+/[\w\.\-]+(\.git)?|https?://[\w\.\-]+/[\w\.\-]+/[\w\.\-]+(\.git)?|git://[\w\.\-]+/[\w\.\-]+/[\w\.\-]+(\.git)?|[\w\.\-]+/[\w\.\-]+/[\w\.\-]+//.*|file://.*\.git)$`)
+	gitURIPattern := regexp.MustCompile(`^(git@[\w\.\-]+:[\w\.\-]+/[\w\.\-]+(\.git)?|https?://[\w\.\-]+/[\w\.\-]+/[\w\.\-]+(\.git)?|git://[\w\.\-]+/[\w\.\-]+/[\w\.\-]+(\.git)?|[\w\.\-]+/[\w\.\-]+/[\w\.\-]+//.*|file://.*\.git|[\w\.\-]+/[\w\.\-]+(\.git)?)$`)
 	// Regular expression for HTTP URIs (with or without protocol)
 	httpURIPattern := regexp.MustCompile(`^((http://|https://)[\w\-]+(\.[\w\-]+)+.*)$`)
 	// Regular expression for file paths
@@ -91,8 +95,23 @@ func ClassifyURI(input string) (URIType, error) {
 
 	// Check if the input contains a dot but lacks a valid scheme
 	if strings.Contains(input, ".") {
-		return Unknown, fmt.Errorf("HTTP(S) URIs require a scheme (http:// or https://)")
+		return Unknown, fmt.Errorf("got %s. HTTP(S) URIs require a scheme (http:// or https://)", input)
 	}
 
 	return Unknown, nil
 }
+
+// ValidateFileDestination validates the d1estination path for saving files
+func ValidateFileDestination(destination string) error {
+	// Expand the tilde in the file path if it exists
+	destination = ExpandTilde(destination)
+	// Check if the destination file exists.
+	_, err := os.Stat(destination)
+	if err == nil {
+		return fmt.Errorf("destination file already exists: %s", destination)
+	}
+	if os.IsNotExist(err) {
+		return nil
+	}
+	return nil
+}

common_test.go

@@ -1,7 +1,10 @@
 package gogather
 
 import (
+	"errors"
 	"fmt"
+	"os"
+	"path/filepath"
 	"testing"
 )
 
@@ -76,6 +79,8 @@ func TestClassifyURI(t *testing.T) {
 		{input: "/home/user/file.git", expected: GitURI},
 		{input: "https://example.com", expected: HTTPURI},
 		{input: "ftpexamplecom", expected: Unknown},
+		{input: "github.com/user/repo.git", expected: GitURI},
+		{input: "gitlab.com/user/repo.git", expected: GitURI},
 	}
 
 	for _, tc := range testCases {
@@ -106,3 +111,60 @@ func TestClassifyURI_errors(t *testing.T) {
 		}
 	}
 }
+
+// TestValidateFileDestination tests the ValidateFileDestination function.
+func TestValidateFileDestination(t *testing.T) {
+	testCases := []struct {
+		destination string
+	}{
+		{destination: "/path/to/file.txt"},
+		{destination: "/path/to/directory/"},
+	}
+
+	for _, tc := range testCases {
+		err := ValidateFileDestination(tc.destination)
+		if err != nil {
+			t.Errorf("Unexpected error: %v", err)
+		}
+	}
+}
+
+// TestValidateFileDestination_errors tests the ValidateFileDestination function with invalid destinations.
+func TestValidateFileDestination_errors(t *testing.T) {
+	dir, _ := os.MkdirTemp("", "path")
+	err := os.WriteFile(filepath.Join(dir, "file.text"), []byte("test"), 0600)
+	if err != nil {
+		t.Fatalf("Failed to create file: %v", err)
+	}
+	defer os.RemoveAll(dir)
+
+	testCases := []struct {
+		destination string
+		errExpected bool
+		expectedErr error
+	}{
+		{
+			destination: filepath.Join(dir, "file.text"),
+			errExpected: true,
+			expectedErr: errors.New("destination is a file"),
+		},
+		{
+			destination: filepath.Join(dir, "file2.text"),
+			errExpected: false,
+			expectedErr: nil,
+		},
+	}
+
+	for _, tc := range testCases {
+		err := ValidateFileDestination(tc.destination)
+		if tc.errExpected && err == nil {
+			t.Errorf("Expected an error: %s,\n but got nil", tc.expectedErr)
+		}
+		if !tc.errExpected && err != nil {
+			t.Errorf("Expected no error, but got: %s", err)
+		}
+	}
+	t.Cleanup(func() {
+		os.RemoveAll(dir)
+	})
+}

gather/file/file.go

@@ -45,6 +45,7 @@ type FileGatherer struct{}
 func (f *FileGatherer) Gather(ctx context.Context, source, destination string) (metadata.Metadata, error) {
 	// Parse the source URI
 	src, err := url.Parse(source)
+
 	if err != nil {
 		return nil, fmt.Errorf("failed to parse source URI: %w", err)
 	}
@@ -126,7 +127,7 @@ func (f *FileGatherer) copyFile(ctx context.Context, source, destination string)
 func (f *FileGatherer) copyDirectory(ctx context.Context, source, destination string) (metadata.Metadata, error) {
 	src, err := url.Parse(source)
 	if err != nil {
-		return nil, err
+		return nil, fmt.Errorf("failed to parse source URI: %w", err)
 	}
 	dst, err := url.Parse(destination)
 	if err != nil {

gather/file/file_test.go

@@ -269,7 +269,7 @@ func TestFileGatherer_copyDirectory_Source_URIParseError(t *testing.T) {
 	if err == nil {
 		t.Error("expected an error, but got nil")
 	}
-	if err.Error() != "parse \":\": missing protocol scheme" {
+	if err.Error() != "failed to parse source URI: parse \":\": missing protocol scheme" {
 		t.Logf("Expected: %s, Got: %s", "parse :: missing protocol scheme", err.Error())
 		t.Fail()
 	}

gather/file/go.mod

@@ -3,9 +3,9 @@ module github.com/enterprise-contract/go-gather/gather/file
 go 1.21.9
 
 require (
-	github.com/enterprise-contract/go-gather/metadata v0.0.0-20240521080222-0648e998132b
-	github.com/enterprise-contract/go-gather/metadata/file v0.0.0-20240521080222-0648e998132b
-	github.com/enterprise-contract/go-gather/saver v0.0.0-20240521080222-0648e998132b
+	github.com/enterprise-contract/go-gather/metadata v0.0.0-20240523073727-ba2c37023242
+	github.com/enterprise-contract/go-gather/metadata/file v0.0.0-20240523073727-ba2c37023242
+	github.com/enterprise-contract/go-gather/saver v0.0.0-20240523073727-ba2c37023242
 )
 
-require github.com/enterprise-contract/go-gather/saver/file v0.0.0-20240521080222-0648e998132b // indirect
+require github.com/enterprise-contract/go-gather/saver/file v0.0.0-20240523073727-ba2c37023242 // indirect

gather/file/go.sum

@@ -1,8 +1,8 @@
-github.com/enterprise-contract/go-gather/metadata v0.0.0-20240521080222-0648e998132b h1:TeQa9c1daZ/EtMslTDBtsk4IO1hP0rPA3E7PpSk/GGI=
-github.com/enterprise-contract/go-gather/metadata v0.0.0-20240521080222-0648e998132b/go.mod h1:m2HxByQBWZyc99HDs/Lqy7QzU9+XQ2tU0X/mzkCPgPw=
-github.com/enterprise-contract/go-gather/metadata/file v0.0.0-20240521080222-0648e998132b h1:/7DlESAdE4CW1cHOWpnZbAhBEzX1P/U8O5ZnL9Ftm0Y=
-github.com/enterprise-contract/go-gather/metadata/file v0.0.0-20240521080222-0648e998132b/go.mod h1:4PckwLejZstUEBp2QUAdQYQ0O+h5tijrs48j+7OY4OY=
-github.com/enterprise-contract/go-gather/saver v0.0.0-20240521080222-0648e998132b h1:kUSMsZ5q+Bj4OdOphfWwxdcfoMHMKO9t1AUZD9l9WtA=
-github.com/enterprise-contract/go-gather/saver v0.0.0-20240521080222-0648e998132b/go.mod h1:iZsEDpzcJSXdm0xaQqucU0xu/h1hZXsY60z5A0R/4VY=
-github.com/enterprise-contract/go-gather/saver/file v0.0.0-20240521080222-0648e998132b h1:2nWimrcrKVSpdHl9bjTzLKhvVbH7pYr/XFKUkqpdsSE=
-github.com/enterprise-contract/go-gather/saver/file v0.0.0-20240521080222-0648e998132b/go.mod h1:qnNStNDYPJGjJunKANv6jq93ynndcfxmUoeYeBEnZEY=
+github.com/enterprise-contract/go-gather/metadata v0.0.0-20240523073727-ba2c37023242 h1:bRMpqsF+NbPf6R514yzo9fVL+8QqOkFoMpMdYjoPynw=
+github.com/enterprise-contract/go-gather/metadata v0.0.0-20240523073727-ba2c37023242/go.mod h1:m2HxByQBWZyc99HDs/Lqy7QzU9+XQ2tU0X/mzkCPgPw=
+github.com/enterprise-contract/go-gather/metadata/file v0.0.0-20240523073727-ba2c37023242 h1:zA8jD+54i4Yybivs7eI74I2hbrzZzW/ifGoBR+Q4T7U=
+github.com/enterprise-contract/go-gather/metadata/file v0.0.0-20240523073727-ba2c37023242/go.mod h1:4PckwLejZstUEBp2QUAdQYQ0O+h5tijrs48j+7OY4OY=
+github.com/enterprise-contract/go-gather/saver v0.0.0-20240523073727-ba2c37023242 h1:rqsKA4myaeXVRiqfHaL5bL7FLjNhuRWmSGHQZz0djEk=
+github.com/enterprise-contract/go-gather/saver v0.0.0-20240523073727-ba2c37023242/go.mod h1:uOt8X/CztOGi0YC5jERopBQpjXqkU6UPUqPellgBBG8=
+github.com/enterprise-contract/go-gather/saver/file v0.0.0-20240523073727-ba2c37023242 h1:BaaoL22+5U4G51WlxjNR5Vxzla3irRZ9zvglf3JYgrA=
+github.com/enterprise-contract/go-gather/saver/file v0.0.0-20240523073727-ba2c37023242/go.mod h1:qnNStNDYPJGjJunKANv6jq93ynndcfxmUoeYeBEnZEY=

gather/gather.go

@@ -39,16 +39,16 @@ type Gatherer interface {
 // protocolHandlers maps URL schemes to their corresponding Gatherer implementations.
 var protocolHandlers = map[string]Gatherer{
 	"FileURI": &file.FileGatherer{},
-	"GitURI":   &git.GitGatherer{},
-	"HTTPURI":  &http.HTTPGatherer{},
+	"GitURI":  &git.GitGatherer{},
+	"HTTPURI": &http.HTTPGatherer{},
 }
 
 // Gather determines the protocol from the source URI and uses the appropriate Gatherer to perform the operation.
 // It returns the gathered metadata and an error, if any.
 func Gather(ctx context.Context, source, destination string) (metadata.Metadata, error) {
 	srcProtocol, err := gogather.ClassifyURI(source)
 	if err != nil {
-		return nil, err
+		return nil, fmt.Errorf("failed to classify source URI: %w", err)
 	}
 
 	if gatherer, ok := protocolHandlers[srcProtocol.String()]; ok {

gather/gather_test.go

@@ -23,7 +23,7 @@ import (
 	"path/filepath"
 	"testing"
 
-	"github.com/enterprise-contract/go-gather"
+	gogather "github.com/enterprise-contract/go-gather"
 	"github.com/enterprise-contract/go-gather/metadata"
 	"github.com/enterprise-contract/go-gather/metadata/git"
 )
@@ -48,30 +48,31 @@ func TestGather(t *testing.T) {
 	t.Run("UnsupportedProtocol", func(t *testing.T) {
 		source := "ftp://example.com/file.txt"
 		destination := "/path/to/destination"
+		defer os.RemoveAll(destination)
 
 		_, err := Gather(ctx, source, destination)
 		if err == nil {
 			t.Error("expected an error, but got nil")
 		}
 
-		expectedErrorMessage := "unsupported source protocol: ftp"
+		expectedErrorMessage := "failed to classify source URI: unsupported source protocol: ftp"
 		if err.Error() != expectedErrorMessage {
 			t.Errorf("expected error message: %s, but got: %s", expectedErrorMessage, err.Error())
 		}
+		t.Cleanup(func() {
+			os.RemoveAll(destination)
+		})
 	})
 
 	t.Run("SupportedProtocol_git", func(t *testing.T) {
 		source := "git::https://github.com/git-fixtures/basic.git"
 		destination := "/tmp/path/to/destination"
 		defer os.RemoveAll(destination)
 
-		// Add your test logic here
-		// BEGIN: SupportedProtocolTest
 		_, err := Gather(ctx, source, destination)
 		if err != nil {
 			t.Errorf("expected no error, but got: %s", err.Error())
 		}
-		// END: SupportedProtocolTest
 		t.Cleanup(func() {
 			os.RemoveAll(destination)
 		})