starkillerOG
commented
1 year ago I will not have time myself to work on the "Baichuan" protocol.
A disadvantage will be that this is not a official API and that in the best case it can break when Reolink makes some changes or implements it diffrently in future models, and in worse case Reolink could actively try to protect their "Baichuan" protocol better by changing encryption keys or encryption protocols. We can also not ask Reolink to inform us about these things since this "Bauchuan" protocol is not ment to be used outside of Reolink, they have their HTTP API for that. So they will probably not be willing to help us with it.
That beeing sad, I am open to Pull Requests that would include the "Baichuan" protocol as a backup to the HTTP API. I do want to have it seperated as much as possible, so the 2 approaches that would be acceptable to me: 1) have a seperate reolink_baichuan pypi library (or use the neolink library if it is maintained) that exposes easy commands much like this reolink_aio library. This reolink_aio libary could have the reolink_baichuan library as a dependency and use it as backup for the HTTP API. 2) have a seperate baichuan folder inside reolink_ai folder of this library in which there is a baichuan.py file which has a class that exposes the baichuan protocol as simple commands (much like the "Host" class of this library). That class can then be used as backup in the Host class.
Possible use cases:
_login_try_portswhen both the HTTP port 80 and HTTPS port 443 fail first, use thebaichuanprotocol on port 9000 to get the NetPort settings, if needed enable the HTTPS and other required ports and look up the correct port numbers to use. After that, go back to the HTTP API using the port just enabled/looked up.- Have some decorater for functions in this library (@baichuan_fallback or something) that will make sure that if the HTTP protocol is not supported by a device (seperate floodlight for instance), that it will fall back to the baichuan protocol implementation of the same function that will need to return/set the variables in exactly the same way as the current function, but then using the baichuan protocol implementation.
I will review PRs to this libary and try to help if you @kevin-david have the time to implement something like discribed above.
xannor
kylepyke
I know this would be a lot of work - but mostly wondering if you're open to a contribution if I find some time, and what approach would be acceptable.
As of today, this would require supporting the "Baichuan" protocol which is best understood by looking at this repo: https://github.com/QuantumEntangledAndy/neolink (and the one it's forked from) along with the blog post here: https://www.thirtythreeforty.net/posts/2020/05/hacking-reolink-cameras-for-fun-and-profit/
I took a stab at implementing this with the existing neolink tool here: https://github.com/QuantumEntangledAndy/neolink/pull/14. But it's still a bit of a mess.
More context and previous ideas I had here: https://github.com/fwestenberg/reolink_dev/issues/599#issuecomment-1345356470