I am trying to read the text of an html element via OCR (website is angular based and content is loaded lazy, why alternatives like html inner text are not working here).
Attended the process is working fine, but runs into timeout when started unattended (Get OCR Text ‘DIV’: Timeout reached).
I isolated and reproduced the problem (see image) by starting the process from studio and minize my remote desktop connection. Seems that OCR is not working, when browser is kind of ‘not visible’.
We played with the WaitForReady-Property as well as Profile, Scale and Invert from OCR Engine Activity (Change of OCR engine is not an option due to cloud service policy, but does not seem to be the problem here as Tesseract is working fine in attended mode).
Hi @tobido,
OCR as name says will not work unattended as it requires screen to be captured during automation. So it requires active session to work. For unattended solutions you need to work with selectors and properties which will help to get text from defined element.
Hi Pawel,
thanks for your answer. OCR stands for optical character recognition, meaning scraping characters instead of reading properties. Not sure what you are referring to here. Many OCR components in many applications work perfectly fine without any screen interaction.
New single page applications are often built with customized html elements, where content is set via data attributes (sth. like < div data-content=“{{ value }}”> < /div >), commonly used in connection with INTL, for instance.
Any idea how I can retrieve this value in unattended mode?
Hi @tobido,
Sorry for wrong shortcut explanation. I was thinking good but wrote something completely wrong… hard day
What I meant is of course screen is not needed, but to capture OCR data the OCR engine need to have access to session so it could grab let’s say proper coordinates of what needs to be captured. And this type of action can’t be 100% unattended. So it can run as background automation.
