Merge branch 'otiai10:main' into main

.github/workflows/go-ci.yml

@@ -13,15 +13,15 @@ jobs:
   macos:
     strategy:
       matrix:
-        os: [macos-latest, macos-12]
-        go: ['1.17', '1.18', '1.19']
+        os: [macos-latest, macos-13]
+        go: ['1.19', '1.20', '1.21']
     runs-on: ${{ matrix.os }}
     name: macOS Go${{ matrix.go }} on ${{ matrix.os }}
     steps:
       - name: Checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
       - name: Go
-        uses: actions/setup-go@v3
+        uses: actions/setup-go@v4
         with:
           go-version: ${{ matrix.go }}
       - name: Install tesseract
@@ -36,9 +36,9 @@ jobs:
     needs: macos
     steps:
       - name: Checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
       - name: Go
-        uses: actions/setup-go@v3
+        uses: actions/setup-go@v4
         with:
           go-version: '1.18'
       - name: Install packages

.github/workflows/runtime-docker.yml

@@ -27,7 +27,7 @@ jobs:
     name: Docker ${{ matrix.runtime }}
     steps:
       - name: checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
       - name: ${{ matrix.runtime }}
         shell: 'script -q -e -c "bash {0}"'
         run: bash ./test/runtime --driver docker --build --run ${{ matrix.runtime }}
@@ -36,7 +36,7 @@ jobs:
     name: Test Dockerfile on the repo root
     steps:
       - name: checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
       - name: Build
         run: docker build . -t otiai10/gosseract
       - name: Test Run

.github/workflows/runtime-vagrant.yml

@@ -22,6 +22,6 @@ jobs:
     name: Vagrant ${{ matrix.runtime }}
     steps:
       - name: checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
       - name: ${{ matrix.runtime }}
         run: bash ./test/runtime --driver vagrant --build --run ${{ matrix.runtime }}

.github/workflows/windows-ci.yml

@@ -18,9 +18,9 @@ jobs:
     name: Windows Test
     steps:
     - name: Checkout
-      uses: actions/checkout@v3
+      uses: actions/checkout@v4
     - name: Setup Go
-      uses: actions/setup-go@v3
+      uses: actions/setup-go@v4
       with:
         go-version: 1.18
     - name: Install Tesseract

all_test.go

@@ -48,6 +48,12 @@ func TestNewClient(t *testing.T) {
 	Expect(t, client).TypeOf("*gosseract.Client")
 }
 
+func TestDoubleClose(t *testing.T) {
+	client := NewClient()
+	client.Close()
+	client.Close()
+}
+
 func TestClient_SetTessdataPrefix(t *testing.T) {
 	client := NewClient()
 	defer client.Close()

client.go

@@ -9,10 +9,16 @@ import (
 	"image"
 	"os"
 	"path/filepath"
+	"runtime"
 	"strings"
 	"unsafe"
 )
 
+var (
+	// ErrClientNotConstructed is returned when a client is not constructed
+	ErrClientNotConstructed = fmt.Errorf("TessBaseAPI is not constructed, please use `gosseract.NewClient`")
+)
+
 // Version returns the version of Tesseract-OCR
 func Version() string {
 	api := C.Create()
@@ -72,18 +78,29 @@ func NewClient() *Client {
 		shouldInit: true,
 		Languages:  []string{"eng"},
 	}
+	// set a finalizer to close the client when it's unused and not closed by the user
+	runtime.SetFinalizer(client, (*Client).Close)
 	return client
 }
 
 // Close frees allocated API. This MUST be called for ANY client constructed by "NewClient" function.
 func (client *Client) Close() (err error) {
+	// no need for a finalizer anymore
+	runtime.SetFinalizer(client, nil)
+	if client.api == nil {
+		// already closed or not constructed
+		return nil
+	}
 	// defer func() {
 	// 	if e := recover(); e != nil {
 	// 		err = fmt.Errorf("%v", e)
 	// 	}
 	// }()
+
 	C.Clear(client.api)
 	C.Free(client.api)
+	client.api = nil
+
 	if client.pixImage != nil {
 		C.DestroyPixImage(client.pixImage)
 		client.pixImage = nil
@@ -93,6 +110,9 @@ func (client *Client) Close() (err error) {
 
 // Version provides the version of Tesseract used by this client.
 func (client *Client) Version() string {
+	if client.api == nil {
+		return ""
+	}
 	version := C.Version(client.api)
 	return C.GoString(version)
 }
@@ -101,7 +121,7 @@ func (client *Client) Version() string {
 func (client *Client) SetImage(imagepath string) error {
 
 	if client.api == nil {
-		return fmt.Errorf("TessBaseAPI is not constructed, please use `gosseract.NewClient`")
+		return ErrClientNotConstructed
 	}
 	if imagepath == "" {
 		return fmt.Errorf("image path cannot be empty")
@@ -119,6 +139,10 @@ func (client *Client) SetImage(imagepath string) error {
 	defer C.free(unsafe.Pointer(p))
 
 	img := C.CreatePixImageByFilePath(p)
+	if img == nil {
+		return fmt.Errorf("failed to create PixImage from file path: %s", imagepath)
+	}
+
 	client.pixImage = img
 
 	return nil
@@ -128,7 +152,7 @@ func (client *Client) SetImage(imagepath string) error {
 func (client *Client) SetImageFromBytes(data []byte) error {
 
 	if client.api == nil {
-		return fmt.Errorf("TessBaseAPI is not constructed, please use `gosseract.NewClient`")
+		return ErrClientNotConstructed
 	}
 	if len(data) == 0 {
 		return fmt.Errorf("image data cannot be empty")
@@ -140,13 +164,20 @@ func (client *Client) SetImageFromBytes(data []byte) error {
 	}
 
 	img := C.CreatePixImageFromBytes((*C.uchar)(unsafe.Pointer(&data[0])), C.int(len(data)))
+	if img == nil {
+		return fmt.Errorf("failed to create PixImage from bytes: %d", len(data))
+	}
+
 	client.pixImage = img
 
 	return nil
 }
 
 // SetLanguage sets languages to use. English as default.
 func (client *Client) SetLanguage(langs ...string) error {
+	if client.api == nil {
+		return ErrClientNotConstructed
+	}
 	if len(langs) == 0 {
 		return fmt.Errorf("languages cannot be empty")
 	}
@@ -160,55 +191,50 @@ func (client *Client) SetLanguage(langs ...string) error {
 
 // DisableOutput ...
 func (client *Client) DisableOutput() error {
-	err := client.SetVariable(DEBUG_FILE, os.DevNull)
-
-	client.setVariablesToInitializedAPIIfNeeded()
-
-	return err
+	return client.SetVariable(DEBUG_FILE, os.DevNull)
 }
 
 // SetWhitelist sets whitelist chars.
 // See official documentation for whitelist here https://tesseract-ocr.github.io/tessdoc/ImproveQuality#dictionaries-word-lists-and-patterns
 func (client *Client) SetWhitelist(whitelist string) error {
-	err := client.SetVariable(TESSEDIT_CHAR_WHITELIST, whitelist)
-
-	client.setVariablesToInitializedAPIIfNeeded()
-
-	return err
+	return client.SetVariable(TESSEDIT_CHAR_WHITELIST, whitelist)
 }
 
 // SetBlacklist sets blacklist chars.
 // See official documentation for blacklist here https://tesseract-ocr.github.io/tessdoc/ImproveQuality#dictionaries-word-lists-and-patterns
 func (client *Client) SetBlacklist(blacklist string) error {
-	err := client.SetVariable(TESSEDIT_CHAR_BLACKLIST, blacklist)
-
-	client.setVariablesToInitializedAPIIfNeeded()
-
-	return err
+	return client.SetVariable(TESSEDIT_CHAR_BLACKLIST, blacklist)
 }
 
 // SetVariable sets parameters, representing tesseract::TessBaseAPI->SetVariable.
 // See official documentation here https://zdenop.github.io/tesseract-doc/classtesseract_1_1_tess_base_a_p_i.html#a2e09259c558c6d8e0f7e523cbaf5adf5
 // Because `api->SetVariable` must be called after `api->Init`, this method cannot detect unexpected key for variables.
 // Check `client.setVariablesToInitializedAPI` for more information.
 func (client *Client) SetVariable(key SettableVariable, value string) error {
+	if client.api == nil {
+		return ErrClientNotConstructed
+	}
 	client.Variables[key] = value
 
-	client.setVariablesToInitializedAPIIfNeeded()
-
-	return nil
+	return client.setVariablesToInitializedAPIIfNeeded()
 }
 
 // SetPageSegMode sets "Page Segmentation Mode" (PSM) to detect layout of characters.
 // See official documentation for PSM here https://tesseract-ocr.github.io/tessdoc/ImproveQuality#page-segmentation-method
 // See https://github.com/otiai10/gosseract/issues/52 for more information.
 func (client *Client) SetPageSegMode(mode PageSegMode) error {
+	if client.api == nil {
+		return ErrClientNotConstructed
+	}
 	C.SetPageSegMode(client.api, C.int(mode))
 	return nil
 }
 
 // SetConfigFile sets the file path to config file.
 func (client *Client) SetConfigFile(fpath string) error {
+	if client.api == nil {
+		return ErrClientNotConstructed
+	}
 	info, err := os.Stat(fpath)
 	if err != nil {
 		return err
@@ -226,6 +252,9 @@ func (client *Client) SetConfigFile(fpath string) error {
 // SetTessdataPrefix sets path to the models directory.
 // Environment variable TESSDATA_PREFIX is used as default.
 func (client *Client) SetTessdataPrefix(prefix string) error {
+	if client.api == nil {
+		return ErrClientNotConstructed
+	}
 	if prefix == "" {
 		return fmt.Errorf("tessdata prefix could not be empty")
 	}
@@ -349,6 +378,9 @@ func (client *Client) setVariablesToInitializedAPIIfNeeded() error {
 
 // Text finally initialize tesseract::TessBaseAPI, execute OCR and extract text detected as string.
 func (client *Client) Text() (out string, err error) {
+	if client.api == nil {
+		return out, ErrClientNotConstructed
+	}
 	if err = client.init(); err != nil {
 		return
 	}
@@ -362,6 +394,9 @@ func (client *Client) Text() (out string, err error) {
 // HOCRText finally initialize tesseract::TessBaseAPI, execute OCR and returns hOCR text.
 // See https://en.wikipedia.org/wiki/HOCR for more information of hOCR.
 func (client *Client) HOCRText() (out string, err error) {
+	if client.api == nil {
+		return out, ErrClientNotConstructed
+	}
 	if err = client.init(); err != nil {
 		return
 	}
@@ -398,7 +433,7 @@ type BoundingBox struct {
 // GetBoundingBoxes returns bounding boxes for each matched word
 func (client *Client) GetBoundingBoxes(level PageIteratorLevel) (out []BoundingBox, err error) {
 	if client.api == nil {
-		return out, fmt.Errorf("TessBaseAPI is not constructed, please use `gosseract.NewClient`")
+		return out, ErrClientNotConstructed
 	}
 	if err = client.init(); err != nil {
 		return
@@ -439,7 +474,7 @@ func GetAvailableLanguages() ([]string, error) {
 // according to the c++ api that returns a formatted TSV output. Reference: `TessBaseAPI::GetTSVText`.
 func (client *Client) GetBoundingBoxesVerbose() (out []BoundingBox, err error) {
 	if client.api == nil {
-		return out, fmt.Errorf("TessBaseAPI is not constructed, please use `gosseract.NewClient`")
+		return out, ErrClientNotConstructed
 	}
 	if err = client.init(); err != nil {
 		return

go.mod

@@ -2,4 +2,4 @@ module github.com/joelrose/gosseract
 
 go 1.18
 
-require github.com/otiai10/mint v1.4.1
+require github.com/otiai10/mint v1.6.3

go.sum

@@ -1,3 +1,2 @@
-github.com/otiai10/curr v1.0.0 h1:TJIWdbX0B+kpNagQrjgq8bCMrbhiuX73M2XwgtDMoOI=
-github.com/otiai10/mint v1.4.1 h1:HOVBfKP1oXIc0wWo9hZ8JLdZtyCPWqjvmFDuVZ0yv2Y=
-github.com/otiai10/mint v1.4.1/go.mod h1:gifjb2MYOoULtKLqUAEILUG/9KONW6f7YsJ6vQLTlFI=
+github.com/otiai10/mint v1.6.3 h1:87qsV/aw1F5as1eH1zS/yqHY85ANKVMgkDrf9rcxbQs=
+github.com/otiai10/mint v1.6.3/go.mod h1:MJm72SBthJjz8qhefc4z1PYEieWmy8Bku7CjcAqyUSM=

preprocessflags_darwin.go

@@ -0,0 +1,7 @@
+package gosseract
+
+// #cgo CXXFLAGS: -std=c++0x
+// #cgo CPPFLAGS: -I/opt/homebrew/include -I/usr/local/include
+// #cgo CPPFLAGS: -Wno-unused-result
+// #cgo LDFLAGS: -L/opt/homebrew/lib -L/usr/local/lib -lleptonica -ltesseract
+import "C"

preprocessflags_x.go

@@ -1,8 +1,9 @@
+//go:build !darwin
+
 package gosseract
 
 // #cgo CXXFLAGS: -std=c++0x
 // #cgo CPPFLAGS: -I/usr/local/include
 // #cgo CPPFLAGS: -Wno-unused-result
-// #cgo darwin LDFLAGS: -L/usr/local/lib -llept -ltesseract
-// #cgo !darwin LDFLAGS: -L/usr/local/lib -lleptonica -ltesseract
+// #cgo LDFLAGS: -L/usr/local/lib -lleptonica -ltesseract
 import "C"