为 IDL 8 图形窗口自定义鼠标处理程序
15307 为本文评分:
未评分
为 IDL 8 图形窗口自定义鼠标处理程序
匿名用户 2014年5月22日 星期四
这篇博客文章提供了一个示例程序，以解答我们在过去几周技术支持中收到的一些问题。示例代码包含两个例程（"ex_closest_xy_widget" 和 "closest_wid_mouse_down"），并在本文底部展示。
该示例程序解决的主要问题是：
- 给定一个包含不同 X 和 Y 值的数据集，如果给你另一组 X0 和 Y0 坐标，你如何确定哪个数据点最接近这些坐标？
- 如何使用 IDL 8 Graphics 创建一个快速的交互式 GUI 来显示问题 (1) 的信息？
- 如何调整 LEGEND 对象，使其只显示一个符号？
问题 (1) 的一个解决方案是计算数据集中每个数据点的 X 与 X0 以及 Y 与 Y0 之间的差值。然后，计算这些差值的幅值并找到最小值。由于 IDL 是向量化的，这可以用几行代码完成：
d = sqrt((x-(x0))^2+(y-(y0))^2) imin = where(d eq min(d))
IDL 8 Graphics（又称新图形，或称函数图形）是交互式的，你可以自定义图形窗口的事件处理程序。因此，你可以使用 IDL 8 Graphics 生成相当不错的 GUI 应用程序，而无需从头开始构建部件。因此，问题 (2) 的解决方案是：先用原始数据生成一个散点图，然后自定义 MOUSE_DOWN_HANDLER，允许用户在图中选择“目标”坐标，接着你就可以使用问题 (1) 的解决方案来找到最接近的数据点并显示该信息。
最后，如果你尝试使用 LEGEND 函数为带有符号的绘图创建图例，你会发现默认情况下它会显示一个符号三次。如果你想将图例更改为只显示符号一次，可以通过将 SAMPLE_WIDTH 属性设置为 0 来实现（三个符号仍然生成，但彼此重叠在一起）。然而，这可能导致符号出现在图例边框之外。要解决此问题，你可以调整 HORIZONTAL_SPACING 关键字。下面展示了一段演示如何实现此操作的代码示例：
lgnd = legend(target=[p,p0, p1], $ sample_width=0, $ horizontal_spacing=.1, $ /auto_text_color)
示例例程 "closest_wid_mouse_down" 和 "ex_closest_xy_widget" 如下所示。我要感谢我的同事 Jim Uba，他对下面的代码贡献很大，并找出了前面讨论的 LEGEND 技巧。运行此代码后，在窗口中左键单击，它应该会显示一个类似于下图的绘图：
function closest_wid_mouse_down, owin, x0, y0, button, keymods, clicks compile_opt idl2
if (button eq 1 ) then begin ;only perform on left mouse click
;get the state variable from the plot window
state = owin.uvalue
;get the reference to the random data point
p = state['p']
;the x and y points returned by the mouse
;handler are in device coordinates by default
;use the convertcoord method to convert it
;to data coordinates
data_coord = p.convertcoord(x0,y0,/device, /to_data)
x0 = data_coord[0]
y0 = data_coord[1]
state['x0']=x0
state['y0']=y0
;get the reference to the "closest point" scatter plot
p1 = state['p1']
;get the reference to the "selected point" scatter plot
p0 = state['p0']
;get the data
x = state['x']
y = state['y']
; calcuate the distances from the data
; points to the target point
d = sqrt((x-(x0))^2+(y-(y0))^2)
; find the array index of the xy point
; with minimum distance to the target point.
imin = where(d eq min(d))
;change the data of the "selected point" scatter plot
;to the point selected by the user. set hide=0 to make
;sure data displays correctly
p0.setdata, [x0,x0],[y0,y0]
p0.hide=0
;change the data of the "selected point" scatter plot
;to the point selected by the user
p1.setdata, [x[imin],x[imin]], [y[imin],y[imin]]
p1.hide=0
;print the selected point
print, "selected point:"
print, "x: ", x0
print, "y: ", y0
;print the closest point
print, "closest data point:"
print, "x: ", x[imin]
print, "y: ", y[imin]
endif
return, 1 end
pro ex_closest_xy_widget
; create random xy data points x = randomu(seed, 100) y = randomu(seed, 100)
;inital points to get the distance x0 = 0 y0 = 0
;generate scatterplot of random xy points p = scatterplot(x, y, $ sym_color='black', symbol='o', $ sym_size=1, sym_thick=3, $ aspect_ratio=1, $ name='random points', uvalue='p')
;generate scatterplot of the selected point ;this plot isn't very useful at this time because ;the user hasn't selected a point. use hide keyword ;to hide it for now p0 = scatterplot([x0,x0], [y0,y0], $ sym_color='red', symbol='+', $ sym_size=2, sym_thick=3, $ /current, /overplot, $ name='target point', /hide,uvalue='p0')
;generate scatterplot of the selected point ;this plot isn't very useful at this time because ;the user hasn't selected a point. use p1 = scatterplot( [x[0],x[0]], [y[0],y[0]], $ sym_color='green', symbol='+', $ sym_size=2, sym_thick=3, $ /current, /overplot, $ name='closest point',/hide,uvalue='p2')
;generate a state variable to store information needed for the ;handler. note: using a hash is a really convenient way to do ;this because they are easy to generate and use, and the data is ;stored as heap memory state=hash() state['p'] = p state['x'] = x state['y'] = y state['x0'] = x0 state['y0'] = y0 state['p0'] = p0 state['p1'] = p1
;assign the state variable as the uvalue of the ;scatterplot window p.window.uvalue = state
;set the mouse_down_handler of the ;window to "closet_wid_mouse_down" p.window.mouse_down_handler="closest_wid_mouse_down"
;generate a legend. by default this will generate 3 ;symbols. if you make the sample_width=0, and adjust ;the horizontal_spacing, you can make the legend on lgnd = legend(target=[p,p0, p1], $ sample_width=0, $ horizontal_spacing=.1, $ /auto_text_color)
;position the plot and legend in the window p.position = [.54,.45] lgnd.position=[.84,.99]
end