update fetch_screen_size()

Author: garywill

README.md

@@ -4,7 +4,7 @@ Let's do some tries to dump traditional mouse
 
 Use keyboard to click anywhere of screen. OpenCV based screenshot visual recognition. 
 
-> The **vi** in the name here can mean "Vim", "Vision" or whatever
+> The **vi** in the name here can mean "Vim", "Vision-recognition" or whatever
 
 ![screenshot](demo.gif)
 
@@ -26,8 +26,9 @@ This is still very simple and **in alpha preview** stage currently. Anything cou
 
 There're some known issues. Apparently there're many we can do to improve it (open for discussion) :
 
-- Algorithm (filtering, recognizing,  parameters... ) . If you don't have a high contrast GUI, or some icons or buttons don't have clear contour, it may fail to recognize them.
-- It's better to make it faster when using hi-res screens
+- Algorithm (filtering, recognizing,  parameters... ) . Currently, if you don't have a high contrast GUI, or some icons or buttons don't have clear contour, or two objects too close,  it may fail to recognize.
+
+- It's better to make it faster when using hi-res full-screens (in the future will apply grid selecting)
 
 ## Other vi / keybinding projects
 

viscreen.py

@@ -514,14 +514,23 @@ def resetRegions() :
     keypListFiltered = keypList
 
 
-# def fetch_screen_size() :
-#     global screenW, screenH
-#     root = tk.Tk()
-#     w = root.winfo_screenwidth()
-#     h = root.winfo_screenheight()
-#     screenW = w
-#     screenH = h
-#     root.destroy();
+def fetch_screen_size() :
+    global screenW, screenH
+    
+
+    app = QGuiApplication([])
+    screens = QGuiApplication.screens()
+
+    total_geometry = QRect()
+    for s in screens:
+        total_geometry = total_geometry.united(s.geometry())
+
+    print(total_geometry.x(), total_geometry.y(), total_geometry.width(), total_geometry.height())    
+    
+    w = total_geometry.width()
+    h = total_geometry.height()
+    screenW = w
+    screenH = h
 
 
 if __name__ == '__main__':