DF2301Q I2C firmware updates

Hello guys!

First of all DF2301Q is brilliant! And it's almost perfect! 

We would like to include in to one of our's commercial devices, but these few little things regarding firmware must be changed before we can do some mass production. Please hear me out. So… her's the list of features that it desperately needs!

1. Whenever device is powered on it says “You are welcome to use voice recognition assistant”. We need a way to turn this off. Let me give you an example… someone has our room sensor hanged on a wall in a bedroom… during the night electricity went away a few times and came back… Personally I would smash that stupid thing with a hammer if it ever speaks to me something like this when I didn't expected it at all. Best of all, it should not say anything at all, but there could be a command, that we could initiate for this to be said. 

2. LED lights - same thing… we need a way of turning them off. I do not want to vandalize it before I can plug it in to our final product. It needs to have a way of having those lights operational, or completely off.

3. Custom wake word. I have successfully trained it with a custom wake word. I am 40 years old man and I have quiet low voice. My wife and all of my kids are able to wake the device by using this wake word that I trained it on. It looks to me like it learned a phoneme or a grapheme of it, therefore it is able to pick it up from any other person. That's actually incredibly amazing! But, since we are planning to integrate in to our own devices, we would like to give it a custom wake word and if it's in a form of phoneme/grapheme it would be amazing to be able to push that thing directly in to DF2301Q. It would acltualy be a killer feature, because this alone can be used for many other purposes.

4. Learn new command initiation via i2C/UART. We need ability to initiate commands learning procedure directly via API, once the process is started, device should give reasonable outputs on what's happening, therefore when requesting for last command it has performed it should give back some numbers that represents current learning process stages. So for example we initiate custom command learning by playing 201 on to it (currently it only plays the sound and nothing really starts happening), it says, whatever it needs to say and listens, users says command for the first time, device asks to repeat, at this moment calling to get last command id it should print out for example 150, which would mean first step of new command learning phase is completed. User says it again, 151 is given out, eventually if it was able to learn new command it delivers ID of that particular command that was learned. As a developer I would now call 207 to exit learning. 

5. Full control ove