sikuli-driver team mailing list archive

[joe fu <question286464@xxxxxxxxxxxxxxxxxxxxx>]

New question #286464 on Sikuli:
https://answers.launchpad.net/sikuli/+question/286464

Hi, I'm currently using Sikuli as UI automation tool to test our product. Since the UI application is onging developed, there is lots of changes in terms of text size and font. I notice Sikuli has a hard time to identify text field image if there is a very minor change in font size or resolution. I also notice it'll make a big difference even only using different screen display method, such as RDP or VNC on the same machine with same resolution on same application. 

So I'm wondering if there is any good way to handle text field efficiently and effectively.

So far I can think of these ways to improve sikuli to look for text field in the screen. Any suggestion or advice or experience sharing is highly appreciated as I know many of you are expects in using Sikuli. Probably we can combine all the good ways to make a good function to handle this kind of situation.

1. OCR, but it's not very accurate and dependable.

2. use region to narrow down the area to search for object, then use low offset for the object to search.

3. use other objects as reference (above, below, left, right) to narrow down the areas to search for object.

4. use a function to define image object with static offset value, default to 0.7, then decrease -0.05 in while loop to find the object, then set the object to use that offset to get the location, then return the object

5. define very low offset, 0.5 for the image, get the list of found objects, get the text from the image object, then compare each object with the expected text to get the correct object.

6. define a list of image objects in array on different resolution or display, use a function to look for each object, then return the object that is found.


-- 
You received this question notification because your team Sikuli Drivers
is an answer contact for Sikuli.