feat(pdfengines): add read metadata route

pkg/modules/api/middlewares.go

@@ -15,9 +15,15 @@ import (
 	"github.com/gotenberg/gotenberg/v8/pkg/gotenberg"
 )
 
-// ErrAsyncProcess happens when a handler or middleware handles a request in an
-// asynchronous fashion.
-var ErrAsyncProcess = errors.New("async process")
+var (
+	// ErrAsyncProcess happens when a handler or middleware handles a request
+	// in an asynchronous fashion.
+	ErrAsyncProcess = errors.New("async process")
+
+	// ErrNoOutputFile happens when a handler or middleware handles a request
+	// without sending any output file.
+	ErrNoOutputFile = errors.New("no output file")
+)
 
 // ParseError parses an error and returns the corresponding HTTP status and
 // HTTP message.
@@ -246,6 +252,13 @@ func contextMiddleware(fs *gotenberg.FileSystem, timeout time.Duration) echo.Mid
 
 			defer cancel()
 
+			if errors.Is(err, ErrNoOutputFile) {
+				// A middleware/handler tells us that it's handling the process
+				// in an asynchronous fashion. Therefore, we must not cancel
+				// the context nor send an output file.
+				return nil
+			}
+
 			if err != nil {
 				return err
 			}

pkg/modules/api/middlewares_test.go

@@ -335,6 +335,15 @@ func TestContextMiddleware(t *testing.T) {
 			}(),
 			expectStatus: http.StatusNoContent,
 		},
+		{
+			request: buildMultipartFormDataRequest(),
+			next: func() echo.HandlerFunc {
+				return func(c echo.Context) error {
+					return ErrNoOutputFile
+				}
+			}(),
+			expectStatus: http.StatusOK,
+		},
 		{
 			request: buildMultipartFormDataRequest(),
 			next: func() echo.HandlerFunc {

pkg/modules/exiftool/exiftool_test.go

@@ -282,14 +282,14 @@ func TestExiftool_WriteMetadata(t *testing.T) {
 				t.Fatal("expected error but got none")
 			}
 
-			if tc.expectError {
-				return
-			}
-
 			if tc.expectedError != nil && !errors.Is(err, tc.expectedError) {
 				t.Fatalf("expected error %v but got: %v", tc.expectedError, err)
 			}
 
+			if tc.expectError {
+				return
+			}
+
 			metadata, err := engine.ReadMetadata(context.Background(), zap.NewNop(), destinationPath)
 			if err != nil {
 				t.Fatalf("expected no error but got: %v", err)

pkg/modules/pdfengines/pdfengines.go

@@ -168,6 +168,7 @@ func (mod *PdfEngines) Routes() ([]api.Route, error) {
 	return []api.Route{
 		mergeRoute(engine),
 		convertRoute(engine),
+		readMetadataRoute(engine),
 	}, nil
 }
 

pkg/modules/pdfengines/pdfengines_test.go

@@ -312,7 +312,7 @@ func TestPdfEngines_Routes(t *testing.T) {
 	}{
 		{
 			scenario:      "routes not disabled",
-			expectRoutes:  2,
+			expectRoutes:  3,
 			disableRoutes: false,
 		},
 		{

pkg/modules/pdfengines/routes.go

@@ -4,6 +4,7 @@ import (
 	"errors"
 	"fmt"
 	"net/http"
+	"path/filepath"
 
 	"github.com/labstack/echo/v4"
 
@@ -78,8 +79,8 @@ func mergeRoute(engine gotenberg.PdfEngine) api.Route {
 	}
 }
 
-// convertRoute returns an [api.Route] which can convert a PDF to a specific
-// PDF format.
+// convertRoute returns an [api.Route] which can convert PDFs to a specific ODF
+// format.
 func convertRoute(engine gotenberg.PdfEngine) api.Route {
 	return api.Route{
 		Method:      http.MethodPost,
@@ -152,3 +153,43 @@ func convertRoute(engine gotenberg.PdfEngine) api.Route {
 		},
 	}
 }
+
+// readMetadataRoute returns an [api.Route] which returns the metadata of PDFs.
+func readMetadataRoute(engine gotenberg.PdfEngine) api.Route {
+	return api.Route{
+		Method:      http.MethodPost,
+		Path:        "/forms/pdfengines/metadata/read",
+		IsMultipart: true,
+		Handler: func(c echo.Context) error {
+			ctx := c.Get("context").(*api.Context)
+
+			// Let's get the data from the form and validate them.
+			var inputPaths []string
+
+			err := ctx.FormData().
+				MandatoryPaths([]string{".pdf"}, &inputPaths).
+				Validate()
+			if err != nil {
+				return fmt.Errorf("validate form data: %w", err)
+			}
+
+			// Alright, let's read the metadata.
+			res := make(map[string]map[string]interface{}, len(inputPaths))
+			for _, inputPath := range inputPaths {
+				metadata, err := engine.ReadMetadata(ctx, ctx.Log(), inputPath)
+				if err != nil {
+					return fmt.Errorf("read metadata: %w", err)
+				}
+
+				res[filepath.Base(inputPath)] = metadata
+			}
+
+			err = c.JSON(http.StatusOK, res)
+			if err != nil {
+				return fmt.Errorf("return JSON response: %w", err)
+			}
+
+			return api.ErrNoOutputFile
+		},
+	}
+}

pkg/modules/pdfengines/routes_test.go

@@ -4,7 +4,9 @@ import (
 	"context"
 	"errors"
 	"net/http"
+	"net/http/httptest"
 	"slices"
+	"strings"
 	"testing"
 
 	"github.com/labstack/echo/v4"
@@ -401,3 +403,107 @@ func TestConvertHandler(t *testing.T) {
 		})
 	}
 }
+
+func TestReadMetadataHandler(t *testing.T) {
+	for _, tc := range []struct {
+		scenario         string
+		ctx              *api.ContextMock
+		engine           gotenberg.PdfEngine
+		expectError      bool
+		expectedError    error
+		expectHttpError  bool
+		expectHttpStatus int
+		expectedJson     string
+	}{
+		{
+			scenario:         "missing at least one mandatory file",
+			ctx:              &api.ContextMock{Context: new(api.Context)},
+			expectError:      true,
+			expectHttpError:  true,
+			expectHttpStatus: http.StatusBadRequest,
+		},
+		{
+			scenario: "error from PDF engine",
+			ctx: func() *api.ContextMock {
+				ctx := &api.ContextMock{Context: new(api.Context)}
+				ctx.SetFiles(map[string]string{
+					"file.pdf": "/file.pdf",
+				})
+				return ctx
+			}(),
+			engine: &gotenberg.PdfEngineMock{
+				ReadMetadataMock: func(ctx context.Context, logger *zap.Logger, inputPath string) (map[string]interface{}, error) {
+					return nil, errors.New("foo")
+				},
+			},
+			expectError:     true,
+			expectHttpError: false,
+		},
+		{
+			scenario: "success",
+			ctx: func() *api.ContextMock {
+				ctx := &api.ContextMock{Context: new(api.Context)}
+				ctx.SetFiles(map[string]string{
+					"file.pdf": "/file.pdf",
+				})
+				return ctx
+			}(),
+			engine: &gotenberg.PdfEngineMock{
+				ReadMetadataMock: func(ctx context.Context, logger *zap.Logger, inputPath string) (map[string]interface{}, error) {
+					return map[string]interface{}{
+						"foo": "bar",
+						"bar": "foo",
+					}, nil
+				},
+			},
+			expectError:     true,
+			expectedError:   api.ErrNoOutputFile,
+			expectHttpError: false,
+			expectedJson:    `{"file.pdf":{"bar":"foo","foo":"bar"}}`,
+		},
+	} {
+		t.Run(tc.scenario, func(t *testing.T) {
+			tc.ctx.SetLogger(zap.NewNop())
+			req := httptest.NewRequest(http.MethodPost, "/forms/pdfengines/metadata/read", nil)
+			rec := httptest.NewRecorder()
+			c := echo.New().NewContext(req, rec)
+			c.Set("context", tc.ctx.Context)
+
+			err := readMetadataRoute(tc.engine).Handler(c)
+
+			if tc.expectError && err == nil {
+				t.Fatal("expected error but got none", err)
+			}
+
+			if !tc.expectError && err != nil {
+				t.Fatalf("expected no error but got: %v", err)
+			}
+
+			var httpErr api.HttpError
+			isHttpError := errors.As(err, &httpErr)
+
+			if tc.expectHttpError && !isHttpError {
+				t.Errorf("expected an HTTP error but got: %v", err)
+			}
+
+			if !tc.expectHttpError && isHttpError {
+				t.Errorf("expected no HTTP error but got one: %v", httpErr)
+			}
+
+			if tc.expectedError != nil && !errors.Is(err, tc.expectedError) {
+				t.Fatalf("expected error %v but got: %v", tc.expectedError, err)
+			}
+
+			if err != nil && tc.expectHttpError && isHttpError {
+				status, _ := httpErr.HttpError()
+				if status != tc.expectHttpStatus {
+					t.Errorf("expected %d as HTTP status code but got %d", tc.expectHttpStatus, status)
+				}
+			}
+
+			if tc.expectedJson != "" && tc.expectedJson != strings.TrimSpace(rec.Body.String()) {
+				t.Errorf("expected '%s' as HTTP response but got '%s'", tc.expectedJson, strings.TrimSpace(rec.Body.String()))
+			}
+		})
+	}
+}