python-onecta

python class to connect to Daikin Onecta cloud interface

See also https://github.com/jwillemsen/daikin_onecta which is a Home Assistant integration, and https://github.com/Apollon77/daikin-controller-cloud which is similar but in javascript or typescript or some-such. (I wasn't really sure what to do with that.)

I don't currently use Home Assistant, so prefer CLI-based tools, and trying to learn python at the same time.

See those other projects for more of the background on getting yourself setup, and of course the daikin api docs themselves at https://developer.cloud.daikineurope.com/ (requires a daikin login).

Assuming you've already got the Onecta app running on a phone, you should already have login details.

The python script requires third-party library requests from pip, but is otherwise just standard python3. Only tested on linux, though. The paths for the configuration and key file might need to be changed for windows. (See top of the class in daikin.py)

Authentication and the redirect url

There are three steps to the initial authorisation process:

1. use browser to connect to a specially-encoded url
2. click on the agree button which generates a code
3. turn the code into an access token.

The script can generate the correct url for step 1, and then take care of step 3.

A slightly confusing aspect of the authenication process is the need for a redirection url. When you authenticate your app, your browser will be directed to this url with a generated code. I assume the idea is that the server that hosts this url will store the code away somewhere. (There's another variable you can also supply which I guess stores some additional context, but I'm not using that.)

The URL has to be https. I'm not quite sure if it has to be accessible from the internet (ie if you need port-forwarding on your router to run it on home network) - I did get that set up, but not sure it's needed.

But I now have a much simpler solution: on a free-hosting site I have for other stuff, I've set up a trivial php script - all it does is echo the ?code=... from the url into the text. You're welcome to use that as your redirect url. (I usually remember to update the certificate, but it may work even after they've expired if you force your browser...)

I don't record any codes sent, though they might show up in the access logs that I don't have access to. But the code is useless without the app id and secret, which I won't have access to anyway.

The url is https://ibmx20.infinityfreeapp.com/daikin.php

If you use something different, you'll have to modify the redir near the top of the source file. (The app needs to send the url from time to time, and it needs to match the one stored on the "app".)

Create an app with redirect url

The first thing you need to do is create an app on the website. Well, they call it an app, but it's little more than a name to identify some tokens.

In the drop-down menu at top right, choose "My Apps", and create one. You have to give it the url redirect as above. (Note that the app has to send this url from time to time, so it has to match.)

So once you've put that into the web page, you can "proceed" to create the app. It will print an id and a secret that you will need to copy.

Create local config file

The script loads a config file during startup to get the app details. Currently expects it to be in ~/.daikin_app.json

That should be json-formatted, like

{
  "id": "XXX",
  "secret": "YYYYY"
}

If you want to be able to make changes using the scripts, you can add a third field "device" - you can find the value you need once you've got the basic functionality working.

Authenticate the app to use your daikin system.

Invoke the python script with a single parameter code. That will print out the url you need to open to authenticate the app. That should bring up a page telling you about the app, and invite you to "agree". If you do, it will forward the browser to the redirect script which, assuming you're using my php script, will just show the code on the page.

Copy that code, then invoke the python script again with code followed by this (very long) code. The script will then connect to the daikin id system and generate an access token, which it will store in a file.

And that should be it done. The token only lasts for an hour, but the script should automatically refresh it as required.

The authentication lasts a year - you'll have to do this manual step again then.

Or if you lose the file containing the key, you'll have to start again.

Test

Invoke the script with get info to get basic information, or get gateway-devices to see all your devices.

Invoke with sensors to show a snapshot of the few temperatures it makes available.

Or 'debug' shows the config, including remaining lifetime of the access token.

Making changes with the API

As mentioned above, you'll have to add a "device" to the config file to make changes. You can get this using get sites which will output

[
    {
        "_id": "XXX-YYYY-ZZZZ",
        "id": "XXX-YYYY-ZZZZ",
        "gatewayDevices": [
          "AAAA-BBB-CCC"
        ],
        ...
    }
]

Or you can also get the gatewayDevice id from get gateway-devices.

These scripts assume only one device

The scripts could possibly issue a get sites automatically if you've not conifugured a site.

Problems

There is a potential problem if you have multiple scripts sharing the key - if they decide to refresh at the same time, they will clash and may well corrupt the key/refresh token, which means you'll have to start again. The script does use file locking which is intended to reduce this risk, but I've not tested it aggressively, so I can't be sure it actually works...

daikin-monitor.py

This is a script that prints out the sensor temperatures every 10 minutes.

Output looks like

2024-10-25--00-05: outdoor=13 room=20.3 / 20.5 hw=38 lwt=20 (offs=0)
2024-10-25--00-15: outdoor=13 room=20.4 / 20.5 hw=38 lwt=23 (offs=0)
2024-10-25--00-25: outdoor=12 room=20.4 / 20.5 hw=34 lwt=37 (offs=0)
...

where second number after room is the target temperature, and offs is the 'leaving water offset' which can be set via app or API (to tweak the lwt).

I actually run a slightly modified version which also displays the the power consumption as measured by a CT clamp monitored by my Zappi charger.

daikin-consumption.py

Fetches and outputs recent consumption figures. Assumption is that you want complete period, so it shows yesterday or last week, rather than in-progress today or this week.

I invoke it from cron every morning to record daily results, and every monday morning to record weekly results.