feat: add support for disabling the obstruction pin

[bdraco]

Fall back to reading the obstruction state if the obstruction pin is disabled.

This is a subpar experience currently but better than nothing.

[mariusmuja]

I implemented support for no obstruction pin at one point and the issue I was having was that there was not status message reliably sent when the instruction was cleared, so I had to ask for a status message in that case. Also there are two bits involved in reporting the obstruction, one gets cleared when an obstruction is present (byte1 - bit 6 - maybe a motor enabled bit?) and one gets set (byte2 - bit 2, obstruction detected bit?). Sometimes a status message with the motor enabled bit set is sent, but no obstruction clear bit is sent.

See here the branch with these changes: ESPHome-RATGDO@f0b8380#diff-8ec49aa77fc5c9af707edd16a08e03e3673aa20fc288a014c9d117f0edc5c2a3R159

[bdraco]

I implemented support for no obstruction pin at one point and the issue I was having was that there was not status message reliably sent when the instruction was cleared, so I had to ask for a status message in that case. Also there are two bits involved in reporting the obstruction, one gets cleared when an obstruction is present (byte1 - bit 6 - maybe a motor enabled bit?) and one gets set (byte2 - bit 2, obstruction detected bit?). Sometimes a status message with the motor enabled bit set is sent, but no obstruction clear bit is sent.

I found bit 6 is reliable (its a bit delayed) when the door is open but not when the door is closed

[mariusmuja]

I found bit 6 is reliable (its a bit delayed) when the door is open but not when the door is closed

Nice. I don't remember for sure, but it is quite possible I was testing with the door closed (but with the opener light on, otherwise the obstruction sensors go into sleep mode).