Updating the OCR add-on to use the contentRecog framework
Closed this issue · 12 comments
Reported by Brendon22 on 2012-05-13 12:41
Hi, When using The OCR Plugg-in and After doing an OCR of The Screen. And trying to read it NVDA will just admit The NVDA Errar Sound? Using NVDA Version 2012.2beta1.
Attachment NVDA Log.txt added by Brendon22 on 2012-05-16 00:37
Description:
My NVDA Log from 16/05/12
Comment 1 by MHameed on 2012-05-13 12:59
This is not enough information to find the cause of the problem, can you please provide a log.
Thanks.
Comment 2 by jteh on 2012-06-08 06:56
You need to use the NVDA review cursor to read the OCR text, not the system caret. This is by design; the text isn't actually available using the caret.
That said, errors when you try to move the caret still aren't ideal. Ideally, the caret should just work as normal, though I'm not sure how to implement this.
Comment 3 by Brendon22 on 2012-06-08 07:08
Thanks for that, and I agree with you about the errars.
Comment 4 by Brendon22 on 2012-08-21 05:33
Hello,
NVDA is still doing this.
- Open up a window wich I know has a picture in it that has text in it.
- Run OCR with NVDA plus R.
- Try and use your aro keys to read the OCRED text. NVdA gives me the following errar:
NotImplementedError
ERROR - eventHandler.executeEvent (15:26:24):
error executing event: caret on <NVDAObjects.Dynamic_IAccessibleEditWindowNVDAObject object at 0x0568FCD0> with extra args of {}
Traceback (most recent call last):
File "eventHandler.pyc", line 136, in executeEvent
File "eventHandler.pyc", line 84, in init
File "eventHandler.pyc", line 91, in next
File "NVDAObjects\IAccessible__init__.pyc", line 1244, in event_caret
File "NVDAObjects\window\edit.pyc", line 791, in event_caret
File "C:\Users\Brendon Donohue\AppData\Roaming\nvda\addons\ocr\globalPlugins\ocr__init__.py", line 207, in
File "C:\Users\Brendon Donohue\AppData\Roaming\nvda\addons\ocr\globalPlugins\ocr__init__.py", line 96, in init
File "textInfos\offsets.pyc", line 275, in init
File "textInfos\offsets.pyc", line 165, in _getSelectionOffsets
NotImplementedError, Log set to input.
Comment 5 by jteh on 2012-08-21 05:59
As noted in comment:2, you aren't meant to be able to read the OCRed text using the arrow keys.
Comment 6 by Brendon22 on 2012-08-21 06:46
Hello,
Oh Yes, sorry!
Comment 7 by Brendon22 on 2014-05-29 10:52
Hello,
These issues haven't been resolved yet. Any progress on this?
@jcsteh Does this still occur? I don't think I follow the issue correctly, because pressing NVDA+R and then arrowing around doesn't seem to generate errors at my end.
I think so. It will occur if you press NVDA+r while in an editable text field.
This can be fixed by updating the OCR add-on to use the new contentRecog framework.
To summarize:
- While I am not too sure If I am exactly able to reproduce the reported errors myself using NVDA 2017.3 on Windows 8.1, this issue ('s title in particular) needs to be morphed to cover the errors generated by NVDA when a user attempts to review the OCR results using the system caret instead of the review cursor.
- The solution, as noted by @jcsteh in #2331 (comment), is to update the OCR add-on to use the new contentRecog framework.
Please open an issue against the ocr addon itself rather than NVDA. With OCR on Windows 10 this is not occuring.
The OCR addon on github is here:
https://github.com/nvaccess/nvda-ocr
I am closing this for now.