Detect invalid characters in file/dir names (#109)

Added logic to validate structure to return validation errors if a file
or directory's name contains characters incompatible with
Archivematica.

internal/activities/validate_structure.go

@@ -7,6 +7,7 @@ import (
 	"os"
 	"path/filepath"
 	"slices"
+	"strings"
 
 	"github.com/artefactual-sdps/preprocessing-sfa/internal/enums"
 	"github.com/artefactual-sdps/preprocessing-sfa/internal/fsutil"
@@ -35,20 +36,24 @@ func (a *ValidateStructure) Execute(
 ) (*ValidateStructureResult, error) {
 	var failures []string
 
-	// Check for empty directories.
+	// Check for empty directories and invalid (Archivematica incompatible) file/directory names.
 	paths := make(map[string]int)
 
 	err := filepath.WalkDir(params.SIP.Path, func(path string, d fs.DirEntry, err error) error {
 		if err != nil {
 			return err
 		}
 
-		if path != params.SIP.Path {
-			relativePath, err := filepath.Rel(params.SIP.Path, path)
-			if err != nil {
-				return err
-			}
+		relativePath, err := filepath.Rel(params.SIP.Path, path)
+		if err != nil {
+			return err
+		}
+
+		if !validateName(d.Name()) {
+			failures = append(failures, fmt.Sprintf("Name %q contains invalid character", relativePath))
+		}
 
+		if path != params.SIP.Path {
 			// Initialize this directory's total number of immediate children.
 			if d.IsDir() {
 				paths[relativePath] = 0
@@ -171,3 +176,16 @@ func extraNodes(sipBase, path string, expected []string, matchDir bool) ([]strin
 
 	return extras, nil
 }
+
+// validateName makes sure only valid characters exist in name.
+func validateName(name string) bool {
+	const validChars = "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789-_.()"
+
+	for i := range len(name) {
+		if !strings.Contains(validChars, string(name[i])) {
+			return false
+		}
+	}
+
+	return true
+}

internal/activities/validate_structure_test.go

@@ -122,6 +122,25 @@ func TestValidateStructure(t *testing.T) {
 	).Path())
 	assert.NilError(t, err)
 
+	badNamingSIP, err := sip.New(fs.NewDir(t, "",
+		fs.WithDir("content",
+			fs.WithDir("d_0000001",
+				fs.WithFile("content.txt", ""),
+			),
+		),
+		fs.WithDir("header",
+			fs.WithFile("content!.txt", ""),
+			fs.WithFile("metadata.xml", ""),
+			fs.WithDir("xsd",
+				fs.WithFile("arelda.xsd", ""),
+			),
+			fs.WithDir("directory$",
+				fs.WithFile("data.xml", ""),
+			),
+		),
+	).Path())
+	assert.NilError(t, err)
+
 	tests := []struct {
 		name    string
 		params  activities.ValidateStructureParams
@@ -180,6 +199,16 @@ func TestValidateStructure(t *testing.T) {
 				Failures: []string{"More than one dossier in the content directory"},
 			},
 		},
+		{
+			name:   "Returns a failure when the name of files and/or directories in a SIP have invalid characters",
+			params: activities.ValidateStructureParams{SIP: badNamingSIP},
+			want: activities.ValidateStructureResult{
+				Failures: []string{
+					"Name \"header/content!.txt\" contains invalid character",
+					"Name \"header/directory$\" contains invalid character",
+				},
+			},
+		},
 	}
 	for _, tt := range tests {
 		t.Run(tt.name, func(t *testing.T) {