Cleanup keyboard code - Githubissues

dictation-toolbox / dragonfly

Speech recognition framework allowing powerful Python-based scripting and extension of Dragon NaturallySpeaking (DNS), Windows Speech Recognition (WSR), Kaldi and CMU Pocket Sphinx

GNU Lesser General Public License v3.0

383 stars 75 forks source link

Cleanup keyboard code #352

Closed drmfinlay closed 3 years ago

drmfinlay commented 3 years ago

This PR cleans up the keyboard code a little and fixes some bugs.

Summary:

Change the Key and Text action classes to calculate keyboard events at execution time instead of initialization time.
Fix a regression with keyboard input actions on Windows. This restores the list of extended keys removed from sendinput.py in PR 343. If the specified virtual-key is in this list, e.g. the insert key, then the KEYEVENTF_EXTENDEDKEY flag is used.
Fix a bug with Text action default pause values. This change fixes a bug where the PAUSE_DEFAULT setting in the Windows-only settings.cfg file is not used properly by Text actions.
Fix numerous code style issues.
Remove doc build conditionals from keyboard/init.py. These should be removed from other files too. I plan to do that as part of a more general clean up of the platform code.

LexiconCode commented 3 years ago

Let me know when this is ready to test!

drmfinlay commented 3 years ago

Should be ready to test now. I can already see a problem with Key("") that I'll fix shortly.

~I'd like to merge 326 before this; I anticipate a fair few conflicts.~

LexiconCode commented 3 years ago

In my limited testing this does fix https://github.com/dictation-toolbox/dragonfly/issues/350

Is this PR meant to address the difference in speed between DNS text output versus dragonfly?

drmfinlay commented 3 years ago

Yes. It looks like that problem is caused by incorrect pauses in the Typeable event methods. I think it is still not quite right. It would not have been helped by the bug I fixed with aaff1ba.

LexiconCode commented 3 years ago

Yes. It looks like that problem is caused by incorrect pauses in the Typeable event methods. I think it is still not quite right. It would not have been helped by the bug I fixed with aaff1ba.

We are all human! At least I hope :) Let me know if you can hear a high-pitched squeal from one of my fans going bad as you read this text.

drmfinlay commented 3 years ago

Haha :)

Well, I can confirm that the behaviour doesn't occur for me using the hardware or Unicode events. It is the same speed as DNS if I set pause_default to 0.

Incidentally, I plan to remove the settings.cfg file in the future; the settings are either defunct or better handled by setting class/instance members instead.