From 7f69052be6289f2ebe23a8116bbc1f7b6f7514bb Mon Sep 17 00:00:00 2001
From: "lilong.129" <lilong.129@bytedance.com>
Date: Mon, 17 Mar 2025 15:09:04 +0800
Subject: [PATCH] change: FindScreenText return text rect

---
 internal/version/VERSION      | 2 +-
 uixt/ai/cv.go                 | 7 ++-----
 uixt/driver_ext_screenshot.go | 9 ++++-----
 uixt/driver_ext_tap.go        | 3 ++-
 uixt/driver_ext_test.go       | 9 ++++++---
 5 files changed, 15 insertions(+), 15 deletions(-)

diff --git a/internal/version/VERSION b/internal/version/VERSION
index fe8b7add..cbeff748 100644
--- a/internal/version/VERSION
+++ b/internal/version/VERSION
@@ -1 +1 @@
-v5.0.0-beta-2503171436
+v5.0.0-beta-2503171509
diff --git a/uixt/ai/cv.go b/uixt/ai/cv.go
index fa314855..0a6ff51d 100644
--- a/uixt/ai/cv.go
+++ b/uixt/ai/cv.go
@@ -86,13 +86,10 @@ func (t OCRText) Size() types.Size {
 }
 
 func (t OCRText) Center() PointF {
-	return getRectangleCenterPoint(t.Rect)
-}
-
-func getRectangleCenterPoint(rect image.Rectangle) (point PointF) {
+	rect := t.Rect
 	x, y := float64(rect.Min.X), float64(rect.Min.Y)
 	width, height := float64(rect.Dx()), float64(rect.Dy())
-	point = PointF{
+	point := PointF{
 		X: x + width*0.5,
 		Y: y + height*0.5,
 	}
diff --git a/uixt/driver_ext_screenshot.go b/uixt/driver_ext_screenshot.go
index 2ace7882..854baeb6 100644
--- a/uixt/driver_ext_screenshot.go
+++ b/uixt/driver_ext_screenshot.go
@@ -145,7 +145,7 @@ func (dExt *XTDriver) GetScreenTexts(opts ...option.ActionOption) (ocrTexts ai.O
 	return screenResult.Texts, nil
 }
 
-func (dExt *XTDriver) FindScreenText(text string, opts ...option.ActionOption) (point ai.PointF, err error) {
+func (dExt *XTDriver) FindScreenText(text string, opts ...option.ActionOption) (textRect ai.OCRText, err error) {
 	options := option.NewActionOptions(opts...)
 	if options.ScreenShotFileName == "" {
 		opts = append(opts, option.WithScreenShotFileName(fmt.Sprintf("find_screen_text_%s", text)))
@@ -155,16 +155,15 @@ func (dExt *XTDriver) FindScreenText(text string, opts ...option.ActionOption) (
 		return
 	}
 
-	result, err := ocrTexts.FindText(text, opts...)
+	textRect, err = ocrTexts.FindText(text, opts...)
 	if err != nil {
 		log.Warn().Msgf("FindText failed: %s", err.Error())
 		return
 	}
-	point = result.Center()
 
 	log.Info().Str("text", text).
-		Interface("point", point).Msgf("FindScreenText success")
-	return
+		Interface("textRect", textRect).Msgf("FindScreenText success")
+	return textRect, nil
 }
 
 func (dExt *XTDriver) FindUIResult(opts ...option.ActionOption) (point ai.PointF, err error) {
diff --git a/uixt/driver_ext_tap.go b/uixt/driver_ext_tap.go
index b4e8f151..8881fc38 100644
--- a/uixt/driver_ext_tap.go
+++ b/uixt/driver_ext_tap.go
@@ -12,13 +12,14 @@ func (dExt *XTDriver) TapByOCR(text string, opts ...option.ActionOption) error {
 		opts = append(opts, option.WithScreenShotFileName(fmt.Sprintf("tap_by_ocr_%s", text)))
 	}
 
-	point, err := dExt.FindScreenText(text, opts...)
+	textRect, err := dExt.FindScreenText(text, opts...)
 	if err != nil {
 		if actionOptions.IgnoreNotFoundError {
 			return nil
 		}
 		return err
 	}
+	point := textRect.Center()
 
 	return dExt.TapAbsXY(point.X, point.Y, opts...)
 }
diff --git a/uixt/driver_ext_test.go b/uixt/driver_ext_test.go
index a50465ba..dec94fb2 100644
--- a/uixt/driver_ext_test.go
+++ b/uixt/driver_ext_test.go
@@ -47,8 +47,8 @@ func TestDriverExt(t *testing.T) {
 	driverExt.TapByOCR("推荐")
 	texts, _ := driverExt.GetScreenTexts()
 	t.Log(texts)
-	point, _ := driverExt.FindScreenText("hello")
-	t.Log(point)
+	textRect, _ := driverExt.FindScreenText("hello")
+	t.Log(textRect)
 
 	err := driverExt.TapByCV(
 		option.WithScreenShotUITypes("deepseek_send"),
@@ -98,13 +98,14 @@ func TestDriverExt_FindScreenText(t *testing.T) {
 func TestDriverExt_Seek(t *testing.T) {
 	driver := setupDriverExt(t)
 
-	point, err := driver.FindScreenText("首页")
+	textRect, err := driver.FindScreenText("首页")
 	assert.Nil(t, err)
 
 	size, err := driver.WindowSize()
 	assert.Nil(t, err)
 	width := size.Width
 
+	point := textRect.Center()
 	y := point.Y - 40
 	for i := 0; i < 5; i++ {
 		err := driver.Swipe(0.5, 0.8, 0.5, 0.2)
@@ -180,4 +181,6 @@ func TestDriverExt_Action_Risk(t *testing.T) {
 	err = driver.Swipe(0.5, 0.5, 0.5, 0.9,
 		option.WithSwipeOffset(-50, 50, -50, 50))
 	assert.Nil(t, err)
+
+	err = driver.TapByOCR("首页", option.WithTapOffset(-10, 10))
 }