feat(pdfengines): add split feature

pkg/gotenberg/fs.go

@@ -3,22 +3,56 @@ package gotenberg
 import (
 	"fmt"
 	"os"
+	"path/filepath"
+	"strings"
 
 	"github.com/google/uuid"
 )
 
+// MkdirAll defines the method signature for create a directory. Implement this
+// interface if you don't want to rely on [os.MkdirAll], notably for testing
+// purpose.
+type MkdirAll interface {
+	// MkdirAll uses the same signature as [os.MkdirAll].
+	MkdirAll(path string, perm os.FileMode) error
+}
+
+// OsMkdirAll implements the [MkdirAll] interface with [os.MkdirAll].
+type OsMkdirAll struct{}
+
+// MkdirAll is a wrapper around [os.MkdirAll].
+func (o *OsMkdirAll) MkdirAll(path string, perm os.FileMode) error { return os.MkdirAll(path, perm) }
+
+// PathRename defines the method signature for renaming files. Implement this
+// interface if you don't want to rely on [os.Rename], notably for testing
+// purpose.
+type PathRename interface {
+	// Rename uses the same signature as [os.Rename].
+	Rename(oldpath, newpath string) error
+}
+
+// OsPathRename implements the [PathRename] interface with [os.Rename].
+type OsPathRename struct{}
+
+// Rename is a wrapper around [os.Rename].
+func (o *OsPathRename) Rename(oldpath, newpath string) error {
+	return os.Rename(oldpath, newpath)
+}
+
 // FileSystem provides utilities for managing temporary directories. It creates
 // unique directory names based on UUIDs to ensure isolation of temporary files
 // for different modules.
 type FileSystem struct {
 	workingDir string
+	mkdirAll   MkdirAll
 }
 
 // NewFileSystem initializes a new [FileSystem] instance with a unique working
 // directory.
-func NewFileSystem() *FileSystem {
+func NewFileSystem(mkdirAll MkdirAll) *FileSystem {
 	return &FileSystem{
 		workingDir: uuid.NewString(),
+		mkdirAll:   mkdirAll,
 	}
 }
 
@@ -44,7 +78,7 @@ func (fs *FileSystem) NewDirPath() string {
 func (fs *FileSystem) MkdirAll() (string, error) {
 	path := fs.NewDirPath()
 
-	err := os.MkdirAll(path, 0o755)
+	err := fs.mkdirAll.MkdirAll(path, 0o755)
 	if err != nil {
 		return "", fmt.Errorf("create directory %s: %w", path, err)
 	}
@@ -52,10 +86,27 @@ func (fs *FileSystem) MkdirAll() (string, error) {
 	return path, nil
 }
 
-// PathRename defines the method signature for renaming files. Implement this
-// interface if you don't want to rely on [os.Rename], notably for testing
-// purpose.
-type PathRename interface {
-	// Rename uses the same signature as [os.Rename].
-	Rename(oldpath, newpath string) error
+// WalkDir walks through the root level of a directory and returns a list of
+// files paths that match the specified file extension.
+func WalkDir(dir, ext string) ([]string, error) {
+	var files []string
+	err := filepath.Walk(dir, func(path string, info os.FileInfo, pathErr error) error {
+		if pathErr != nil {
+			return pathErr
+		}
+		if info.IsDir() {
+			return nil
+		}
+		if strings.EqualFold(filepath.Ext(info.Name()), ext) {
+			files = append(files, path)
+		}
+		return nil
+	})
+	return files, err
 }
+
+// Interface guards.
+var (
+	_ MkdirAll   = (*OsMkdirAll)(nil)
+	_ PathRename = (*OsPathRename)(nil)
+)

pkg/gotenberg/fs_test.go

@@ -1,14 +1,84 @@
 package gotenberg
 
 import (
+	"errors"
 	"fmt"
+	"io"
 	"os"
+	"path/filepath"
+	"reflect"
 	"strings"
 	"testing"
+
+	"github.com/google/uuid"
 )
 
+func TestOsMkdirAll_MkdirAll(t *testing.T) {
+	dirPath, err := NewFileSystem(new(OsMkdirAll)).MkdirAll()
+	if err != nil {
+		t.Fatalf("create working directory: %v", err)
+	}
+
+	err = os.RemoveAll(dirPath)
+	if err != nil {
+		t.Fatalf("remove working directory: %v", err)
+	}
+}
+
+func TestOsPathRename_Rename(t *testing.T) {
+	dirPath, err := NewFileSystem(new(OsMkdirAll)).MkdirAll()
+	if err != nil {
+		t.Fatalf("create working directory: %v", err)
+	}
+
+	path := "/tests/test/testdata/api/sample1.txt"
+	copyPath := filepath.Join(dirPath, fmt.Sprintf("%s.txt", uuid.NewString()))
+
+	in, err := os.Open(path)
+	if err != nil {
+		t.Fatalf("open file: %v", err)
+	}
+
+	defer func() {
+		err := in.Close()
+		if err != nil {
+			t.Fatalf("close file: %v", err)
+		}
+	}()
+
+	out, err := os.Create(copyPath)
+	if err != nil {
+		t.Fatalf("create new file: %v", err)
+	}
+
+	defer func() {
+		err := out.Close()
+		if err != nil {
+			t.Fatalf("close new file: %v", err)
+		}
+	}()
+
+	_, err = io.Copy(out, in)
+	if err != nil {
+		t.Fatalf("copy file to new file: %v", err)
+	}
+
+	rename := new(OsPathRename)
+	newPath := filepath.Join(dirPath, fmt.Sprintf("%s.txt", uuid.NewString()))
+
+	err = rename.Rename(copyPath, newPath)
+	if err != nil {
+		t.Errorf("expected no error but got: %v", err)
+	}
+
+	err = os.RemoveAll(dirPath)
+	if err != nil {
+		t.Fatalf("remove working directory: %v", err)
+	}
+}
+
 func TestFileSystem_WorkingDir(t *testing.T) {
-	fs := NewFileSystem()
+	fs := NewFileSystem(new(MkdirAllMock))
 	dirName := fs.WorkingDir()
 
 	if dirName == "" {
@@ -17,7 +87,7 @@ func TestFileSystem_WorkingDir(t *testing.T) {
 }
 
 func TestFileSystem_WorkingDirPath(t *testing.T) {
-	fs := NewFileSystem()
+	fs := NewFileSystem(new(MkdirAllMock))
 	expectedPath := fmt.Sprintf("%s/%s", os.TempDir(), fs.WorkingDir())
 
 	if fs.WorkingDirPath() != expectedPath {
@@ -26,7 +96,7 @@ func TestFileSystem_WorkingDirPath(t *testing.T) {
 }
 
 func TestFileSystem_NewDirPath(t *testing.T) {
-	fs := NewFileSystem()
+	fs := NewFileSystem(new(MkdirAllMock))
 	newDir := fs.NewDirPath()
 	expectedPrefix := fs.WorkingDirPath()
 
@@ -36,20 +106,117 @@ func TestFileSystem_NewDirPath(t *testing.T) {
 }
 
 func TestFileSystem_MkdirAll(t *testing.T) {
-	fs := NewFileSystem()
+	for _, tc := range []struct {
+		scenario    string
+		mkdirAll    MkdirAll
+		expectError bool
+	}{
+		{
+			scenario: "error",
+			mkdirAll: &MkdirAllMock{
+				MkdirAllMock: func(path string, perm os.FileMode) error {
+					return errors.New("foo")
+				},
+			},
+			expectError: true,
+		},
+		{
+			scenario: "success",
+			mkdirAll: &MkdirAllMock{
+				MkdirAllMock: func(path string, perm os.FileMode) error {
+					return nil
+				},
+			},
+			expectError: false,
+		},
+	} {
+		t.Run(tc.scenario, func(t *testing.T) {
+			fs := NewFileSystem(tc.mkdirAll)
 
-	newPath, err := fs.MkdirAll()
-	if err != nil {
-		t.Fatalf("expected no error but got: %v", err)
-	}
+			_, err := fs.MkdirAll()
 
-	_, err = os.Stat(newPath)
-	if os.IsNotExist(err) {
-		t.Errorf("expected directory '%s' to exist but it doesn't", newPath)
-	}
+			if !tc.expectError && err != nil {
+				t.Fatalf("expected no error but got: %v", err)
+			}
 
-	err = os.RemoveAll(fs.WorkingDirPath())
-	if err != nil {
-		t.Fatalf("expected no error while cleaning up but got: %v", err)
+			if tc.expectError && err == nil {
+				t.Fatal("expected error but got none")
+			}
+		})
+	}
+}
+
+func TestWalkDir(t *testing.T) {
+	for _, tc := range []struct {
+		scenario    string
+		dir         string
+		ext         string
+		expectError bool
+		expectFiles []string
+	}{
+		{
+			scenario:    "directory does not exist",
+			dir:         uuid.NewString(),
+			ext:         ".pdf",
+			expectError: true,
+		},
+		{
+			scenario: "find PDF files",
+			dir: func() string {
+				path := fmt.Sprintf("%s/a_directory", os.TempDir())
+
+				err := os.MkdirAll(path, 0o755)
+				if err != nil {
+					t.Fatalf(fmt.Sprintf("expected no error but got: %v", err))
+				}
+
+				err = os.WriteFile(fmt.Sprintf("%s/a_foo_file.pdf", path), []byte{1}, 0o755)
+				if err != nil {
+					t.Fatalf("expected no error but got: %v", err)
+				}
+
+				err = os.WriteFile(fmt.Sprintf("%s/a_bar_file.PDF", path), []byte{1}, 0o755)
+				if err != nil {
+					t.Fatalf("expected no error but got: %v", err)
+				}
+
+				err = os.WriteFile(fmt.Sprintf("%s/a_baz_file.txt", path), []byte{1}, 0o755)
+				if err != nil {
+					t.Fatalf("expected no error but got: %v", err)
+				}
+
+				return path
+			}(),
+			ext:         ".pdf",
+			expectError: false,
+			expectFiles: []string{"/tmp/a_directory/a_bar_file.PDF", "/tmp/a_directory/a_foo_file.pdf"},
+		},
+	} {
+		t.Run(tc.scenario, func(t *testing.T) {
+			defer func() {
+				err := os.RemoveAll(tc.dir)
+				if err != nil {
+					t.Fatalf("expected no error while cleaning up but got: %v", err)
+				}
+			}()
+
+			files, err := WalkDir(tc.dir, tc.ext)
+
+			if !tc.expectError && err != nil {
+				t.Fatalf("expected no error but got: %v", err)
+			}
+
+			if tc.expectError && err == nil {
+				t.Fatal("expected error but got none")
+			}
+
+			if tc.expectError && err != nil {
+				return
+			}
+
+			if !reflect.DeepEqual(files, tc.expectFiles) {
+				t.Errorf("expected files %+v, but got %+v", tc.expectFiles, files)
+			}
+		})
 	}
 }

pkg/gotenberg/mocks.go

@@ -2,6 +2,7 @@ package gotenberg
 
 import (
 	"context"
+	"os"
 
 	"go.uber.org/zap"
 )
@@ -36,6 +37,7 @@ func (mod *ValidatorMock) Validate() error {
 // PdfEngineMock is a mock for the [PdfEngine] interface.
 type PdfEngineMock struct {
 	MergeMock         func(ctx context.Context, logger *zap.Logger, inputPaths []string, outputPath string) error
+	SplitMock         func(ctx context.Context, logger *zap.Logger, mode SplitMode, inputPath, outputDirPath string) ([]string, error)
 	ConvertMock       func(ctx context.Context, logger *zap.Logger, formats PdfFormats, inputPath, outputPath string) error
 	ReadMetadataMock  func(ctx context.Context, logger *zap.Logger, inputPath string) (map[string]interface{}, error)
 	WriteMetadataMock func(ctx context.Context, logger *zap.Logger, metadata map[string]interface{}, inputPath string) error
@@ -45,6 +47,10 @@ func (engine *PdfEngineMock) Merge(ctx context.Context, logger *zap.Logger, inpu
 	return engine.MergeMock(ctx, logger, inputPaths, outputPath)
 }
 
+func (engine *PdfEngineMock) Split(ctx context.Context, logger *zap.Logger, mode SplitMode, inputPath, outputDirPath string) ([]string, error) {
+	return engine.SplitMock(ctx, logger, mode, inputPath, outputDirPath)
+}
+
 func (engine *PdfEngineMock) Convert(ctx context.Context, logger *zap.Logger, formats PdfFormats, inputPath, outputPath string) error {
 	return engine.ConvertMock(ctx, logger, formats, inputPath, outputPath)
 }
@@ -137,6 +143,15 @@ func (provider *MetricsProviderMock) Metrics() ([]Metric, error) {
 	return provider.MetricsMock()
 }
 
+// MkdirAllMock is a mock for the [MkdirAll] interface.
+type MkdirAllMock struct {
+	MkdirAllMock func(path string, perm os.FileMode) error
+}
+
+func (mkdirAll *MkdirAllMock) MkdirAll(path string, perm os.FileMode) error {
+	return mkdirAll.MkdirAllMock(path, perm)
+}
+
 // PathRenameMock is a mock for the [PathRename] interface.
 type PathRenameMock struct {
 	RenameMock func(oldpath, newpath string) error
@@ -156,4 +171,6 @@ var (
 	_ ProcessSupervisor = (*ProcessSupervisorMock)(nil)
 	_ LoggerProvider    = (*LoggerProviderMock)(nil)
 	_ MetricsProvider   = (*MetricsProviderMock)(nil)
+	_ MkdirAll          = (*MkdirAllMock)(nil)
+	_ PathRename        = (*PathRenameMock)(nil)
 )

pkg/gotenberg/mocks_test.go

@@ -2,6 +2,7 @@ package gotenberg
 
 import (
 	"context"
+	"os"
 	"testing"
 
 	"go.uber.org/zap"
@@ -52,6 +53,9 @@ func TestPDFEngineMock(t *testing.T) {
 		MergeMock: func(ctx context.Context, logger *zap.Logger, inputPaths []string, outputPath string) error {
 			return nil
 		},
+		SplitMock: func(ctx context.Context, logger *zap.Logger, mode SplitMode, inputPath, outputDirPath string) ([]string, error) {
+			return nil, nil
+		},
 		ConvertMock: func(ctx context.Context, logger *zap.Logger, formats PdfFormats, inputPath, outputPath string) error {
 			return nil
 		},
@@ -68,6 +72,11 @@ func TestPDFEngineMock(t *testing.T) {
 		t.Errorf("expected no error from PdfEngineMock.Merge, but got: %v", err)
 	}
 
+	_, err = mock.Split(context.Background(), zap.NewNop(), SplitMode{}, "", "")
+	if err != nil {
+		t.Errorf("expected no error from PdfEngineMock.Split, but got: %v", err)
+	}
+
 	err = mock.Convert(context.Background(), zap.NewNop(), PdfFormats{}, "", "")
 	if err != nil {
 		t.Errorf("expected no error from PdfEngineMock.Convert, but got: %v", err)
@@ -205,6 +214,19 @@ func TestMetricsProviderMock(t *testing.T) {
 	}
 }
 
+func TestMkdirAllMock(t *testing.T) {
+	mock := &MkdirAllMock{
+		MkdirAllMock: func(dir string, perm os.FileMode) error {
+			return nil
+		},
+	}
+
+	err := mock.MkdirAll("/foo", 0o755)
+	if err != nil {
+		t.Errorf("expected no error from MkdirAllMock.MkdirAll, but got: %v", err)
+	}
+}
+
 func TestPathRenameMock(t *testing.T) {
 	mock := &PathRenameMock{
 		RenameMock: func(oldpath, newpath string) error {

pkg/gotenberg/pdfengine.go

@@ -12,6 +12,10 @@ var (
 	// PdfEngine interface is not supported by its current implementation.
 	ErrPdfEngineMethodNotSupported = errors.New("method not supported")
 
+	// ErrPdfSplitModeNotSupported is returned when the Split method of the
+	// PdfEngine interface does not sumport a requested PDF split mode.
+	ErrPdfSplitModeNotSupported = errors.New("split mode not supported")
+
 	// ErrPdfFormatNotSupported is returned when the Convert method of the
 	// PdfEngine interface does not support a requested PDF format conversion.
 	ErrPdfFormatNotSupported = errors.New("PDF format not supported")
@@ -21,6 +25,26 @@ var (
 	ErrPdfEngineMetadataValueNotSupported = errors.New("metadata value not supported")
 )
 
+const (
+	// SplitModeIntervals represents a mode where a PDF is split at specific
+	// intervals.
+	SplitModeIntervals string = "intervals"
+
+	// SplitModePages represents a mode where a PDF is split at specific page
+	// ranges.
+	SplitModePages string = "pages"
+)
+
+// SplitMode gathers the data required to split a PDF into multiple parts.
+type SplitMode struct {
+	// Mode is either "intervals" or "pages".
+	Mode string
+
+	// Span is either the intervals or the page ranges to extract, depending on
+	// the selected mode.
+	Span string
+}
+
 const (
 	// PdfA1a represents the PDF/A-1a format.
 	PdfA1a string = "PDF/A-1a"
@@ -65,6 +89,9 @@ type PdfEngine interface {
 	// is determined by the order of files provided in inputPaths.
 	Merge(ctx context.Context, logger *zap.Logger, inputPaths []string, outputPath string) error
 
+	// Split splits a given PDF file.
+	Split(ctx context.Context, logger *zap.Logger, mode SplitMode, inputPath, outputDirPath string) ([]string, error)
+
 	// Convert transforms a given PDF to the specified formats defined in
 	// PdfFormats. If no format, it does nothing.
 	Convert(ctx context.Context, logger *zap.Logger, formats PdfFormats, inputPath, outputPath string) error