Add sikuli actions associated with ocr

[R-Loic]

Hello, Could you please have a look on the ocr technology associated with sikuli ? The ocr technology will enable to read text on a picture and it would be great if we could for example click on the text. Thank you!

[bcivel]

Hello @R-Loic You have now the possibility to use 2 identifier with sikuli : picture and text. If you set picture= , it will focus on picture If you set text=, it will find the text in the page and then focus on it.

[R-Loic]

Hello @bcivel , The implementation with "text=" doesn't work everytime, for some words sikuli returns the error : Action is Click on text=Monsieur cerberus/Sikuli try to get an image. Could you have a look ? thanks !

[R-Loic]

I got a weird behavior with "text=" :

The control is : VerifyElementPresent returns KO with the following message : Element 'Excel' is present on the page. The execution time is 0 ms. It seems sikuli didn't search for it.

[R-Loic]

When sikuli desn't find the text, it returns : It seems the several words are not well managed.

[bcivel]

Hello @R-Loic In fact, the two different identifier (text and picture) are used by Cerberus to make the difference between action that need to upload and forward picture (picture=) and action that only forward text (text=)

Once forwarded to sikuli, you only have 1 API (for exemple : s.click(element) . And by default, Sikuli try first to load the picture and if it fail, it log the error you get and try to find the text. This is the internal behavior of sikuli.

For the other issues, I'll have a look. But I can already confirm that Sikuli have a lot of difficulty to work with multiple words.