Warning
This package will most likely not work with your machine. Any machine, that updated to Gateway v3 in or before September 2023, will need to use https://github.com/zweckj/lmcloud. If you're looking for the Home Assistant integration, please see this one
This repo will be archived.
This is a prototype library for interacting with the local network API of a La Marzocco espresso machine. It's still in the beginning stages, but currently it's able to retrieve configuration and status information from the machine and turn it on/off. I've reverse-engineered this from decoding the network traffic between the mobile app and the machine, so it's incomplete, imperfect, and may have significant bugs. It works on my La Marzocco GS/3, but I can't guarantee that you'll have a similar experience with yours.
You take any and all risks from using this on your machine!
Using this library is an advanced exercise. You'll need to do the following:
- Find the
client_idandclient_secretfor your machine by following these instructions. - You'll need the username & password that you used to register with La Marzocco when you set up remote access, and the username is most likely the email address that you used for registration.
Once you have the client ID, client secret, username, and passowrd, construct a file called config.json with these contents and put it in the directory along with test.py. filename and key are only required if you run parse.py.
{
"filename": "dose_tea.json",
"key": "long-key-string",
"host": "ip_addr",
"port": "1774",
"client_id": "long_string",
"client_secret": "another_long_string",
"username": "email_address",
"password": "password"
}
Now, run python test.py and you should get a prompt that looks like this:
1=Power <on/off>, 2=Status, 3=Coffee Temp <temp>, 4=Steam Temp <temp>, 5=PB <on/off>, 6=Auto on/off <0=global or day> <on/off>, 7=Dose <sec>, 8=Tea Dose <sec>, 8=PB times <on off>:
The app requests all status & config information every 20 seconds, so you should see the values change when the state of the machine changes.
The machine can only accept a single connection at a time, but the library keeps the connection open long enough only to send commands and receive the responses. Both the mobile app and this library will attempt to reconnect if the port is in use, but you may need to wait a bit while using the mobile app for it to try again.
The package's external API can be found in __init__.py. request_status() will automatically connect to the machine, retrieve lots of status and configuration information, and build a dict that can be retrieved by calling the current_status() API. Several properties are available for direct access and there's a set of services that allow the user to change machine settings. Users can register for a callback when new data is received.
The raw API is comprised of "read" messages that start with "R", "write" messages that start with "W", and "streaming" messages that start with "Z". Following the initial letter, all messages have a 16-bit address and 16-bit length followed by data to write or that was read. In essence, the API is just a peek/poke API into the memory space of the machine, and the machine updates the contents when changes are made on the machine and reacts to writes that occur.
The entire message is sent in ASCII-encoded hex, encrypted, base-64 encoded, and framed between "@" and "%" characters.
You can find an extensive set of interesting memory regions and their maps in msgs.py here: https://github.com/rccoleman/lmdirect/tree/master/lmdirect.