UiPath RPA Academy "Lesson 7 - Image and Text Automation" 之學習心得 PART7

Lesson 7 - Image and Text Automation

暑假即將結束，嶄新學期到來！
督促自己持續努力，並相信自己的進步 :))

Citrix專論

• 雖然Citrix通常都是用在virtual machine，但其實是一個從server投影到user的概念，所以UiPath無法address the interface through the operating system.
• 目前階段需要教導robots

要get info in用

1. ClickImg 
2. ClickTxt 
3. TypeInto

註：

• 有關於Accuracy，如果設定"1"代表所對應圖示需要100%吻合，但事實上設定"0.8"已經是可以有不錯的結果。
• ClickTxt需要OCR來支援，ClickImg基本上很可靠，除了當graphical elements改變。
• 上面動作都是用滑鼠執行，但其實可以利用鍵盤key避免選錯。

要get info out用

1. Select & Copy
2. ScrapeRelative

註：

• Select & Copy是最簡單的方法
• ScrapeRelative藉由GetOCRText來獲得content image(non-selective).

Creating automation in a Citrix environment is challenging because:

You need to interact with the app using only Image Recognition and OCR.
You don’t have direct access to UI elements.
Selectors are hard to create for virtual environments.

Can the robot perform clicks alongside key modifiers (Shift, Ctrl etc) in a Citrix environment?

Yes
No

You can use image/text automation outside of a Citrix enviroment?

False
True

Is Reset Clipping Region mandatory to be executed at the end of a scrape relative sequence?

No, for the next actions we can use other Clipping Regions.
Yes, because Clipping Region is a shared resource.

Consider having an application in Citrix Environment that has a button named ‘Accept’ and also a label that contains the Accept word. How can Click Text be customized in order to access the correct button?

By checking the element’s attributes.
It can’t be done, having to click on a text that is duplicated can’t be controlled.
By using the Occurrence property.

Having an app in a Citrix environment with multiple text-boxes that look the same (size/style), how can you identify one of them to type into?

You can’t identify it if it doesn’t have something unique next to it (text/image).
By using text-box element attributes
By using partial selectors
By clicking relative to an unique text/image next to the textbox

Is it possible to click a button with Click Image Activity if the target is not visible on the screen?

Yes, the robot can click an image even if it's not visible on the screen
No, you could click a button which is not visible only using selectors
Yes, but you have to click the text from the button not the button image

What method would be more reliable when clicking on a specific text label in an application running in a Citrix environment, given the fact that its font size might be easily changed?

Using the Click OCR Text activity.
Using the Click Image activity.(X)
It can’t be done if its size fluctuates.(X)

How can you improve accuracy when scraping with OCR a region that contains only digits?

Make sure the background is dark
Use Get Text for the field in the Citrix Window(X)
Use Google OCR with “Numbers Only”(O)

What does the "Accuracy" property describes in "Click Image" Activity?

“Click Image” does not have such a property
Percent unit (0, 100] of the minimum similarity between the image found and the image you are searching for
Minimum Similarities in [0..1] percentage units for an image to be returned as a match

What activities can be used to interact with applications in a Citrix environment?

Click Text

Click OCR Text

Type Into

Click Image

How can you scrape a field on a Citrix Environment when the value in that field changes each transaction?

Find a static element nearby and use Scrape Relative

It's impossible because you cannot locate the element

Click Image and Click Text are not 100% reliable in Citrix environments. What method can be used instead (when applicable) to have safer actions?

Using full selectors.

Setting focus on a reliable element and then navigating around the app using keyboard (up/down arrows, tab, etc) or using keyboard shortcuts.

Setting the Accuracy property of the Click Image activity to 1.

SHARE AND LIKE IF IT HELPS