Decoded - EvoHome API access to control remotely.

[victorp1612]

Hi Freeranger,
Thanks for your evologger tool , working fine but I'm having an issue with uploading to Plotly and was hoping you can help .
Evologger is picking up data no problem but I then get a 404 error back from Plotly

INFO:evohome-logger:::2016-06-20 06:29:05: Attic (20.0 A, 5.0 T) Bathroom (20.5 A, 16.0 T) Caoimhe (19.5 A, 16.0 T) Ensuite (21.0 A, 17.0 T) Hall (19.5 A, 16.0 T) Kitchen (19.0 A, 15.0 T) Living room (20.0 A, 16.0 T) Master Bed (20.5 A, 16.0 T) Saoirse (19.5 A, 17.5 T) Hot Water (55.0 A, 55.0 T)
ERRORlotly-plugin::
Plot.ly API error - aborting
Server responded with status code: 404
and message: streamtoken not registered or valid.
INFO:evohome-logger:::Going to sleep for 5 minutes

I then realized that I hadn't initialized the API streams as per your instructions so ran python plugins/plotly/__init__.py but I get the following :

Python 2.7.11 (v2.7.11:6d1b6a68f775, Dec 5 2015, 20:40:30) [MSC v.1500 64 bit (AMD64)] on win32
Type "copyright", "credits" or "license()" for more information.
>>>
========= RESTART: C:\Python27\evologger\plugins\plotly\__init__.py =========

Traceback (most recent call last):
File "C:\Python27\evologger\plugins\plotly\__init__.py" , line 17, in <module>
plotly_username = config.get("Plotly", "username")
File "C:\Python27\lib\ConfigParser.py", line 607, in get
raise NoSectionError(section)
NoSectionError: No section: 'Plotly'
>>>

Any hints please ?
Thanks
Vic.

[freeranger]

Traceback (most recent call last):
File "C:\Python27\evologger\plugins\plotly\__init__.py" , line 17, in <module>
plotly_username = config.get("Plotly", "username")
File "C:\Python27\lib\ConfigParser.py", line 607, in get
raise NoSectionError(section)
NoSectionError: No section: 'Plotly'
>>>

Any hints please ?
Thanks
Vic.

Hi Vic,

Glad at least part of it is working for you

Did you run "python plugins/plotly/__init__.py" from the evologger folder, not from the plotly plugin folder?
Does the config.ini file contain:

[Plotly]
username=<Your Plot.ly username>
apiKey=<You PLot.ly API key

If both the above are true, temporarily copy the config.ini file into the plotly plugin folder and see what happens....

[victorp1612]

All sorted now Freeranger , thanks for the tip : I was indeed running the script from the plotly plugin folder. Running it from the evologger folder resolved the issue and I now see data flowing in Plotly :-)
Thanks again for your help !

[freeranger]

All sorted now Freeranger , thanks for the tip : I was indeed running the script from the plotly plugin folder. Running it from the evologger folder resolved the issue and I now see data flowing in Plotly :-)
Thanks again for your help !

Phew
BTW if you can figure out what would be needed to map actual vs expected temp using streams then I can update the plugin accordingly.

[Conseils]

Hi,
03 Aug 2016 06:00 (UK) this morning the https://tccna.honeywell.com/WebApi/ seemed off line for two hours.

After, at 08:00 03 Aug 2016 06:00 (UK) I seemed to start receiving temperature data from my valves with increased resolution.

i.e. from ±0.5 °C
to ±0.01°C

It's unlikely its accurate to 0.01°C though the 0.1°C gives trending, which is nice.

I haven't upgraded the Raspberry PI so think at the moment think it's the Honeywell WebApi change. ( A picture of the impact is shown, though its trivial to rectify )

Anyone care to confirm the observation?

evohome-00.jpg

[Caesium]

Yep, same for me. My graphs suddenly got a lot smoother!

[DBMandrake]

I too saw an outage from approximately the same time however I do not see any increase in resolution.

I use both munin-evohome, and have recently set up domoticz as well - both are still showing 0.5 degree increments. Both use the version 2 API so on a hunch I did some manual tests with my munin-evohome plugin, which can be switched between V1 and V2 API's. The first is the V2 API the second is the older V1 API:

pi@pi1monitor:~ $ ./evohome_test
{'temp': 22.0, 'setpoint': 22.0, 'thermostat': 'EMEA_ZONE', 'name': u'Bathroom', 'id': u'1377389'}
{'temp': 21.0, 'setpoint': 21.0, 'thermostat': 'EMEA_ZONE', 'name': u'Living Room', 'id': u'1378782'}
{'temp': 18.5, 'setpoint': 18.0, 'thermostat': 'EMEA_ZONE', 'name': u'Main Bedroom', 'id': u'1438045'}
{'temp': 21.5, 'setpoint': 20.0, 'thermostat': 'EMEA_ZONE', 'name': u'Kitchen', 'id': u'1643296'}
{'temp': 20.5, 'setpoint': 20.0, 'thermostat': 'EMEA_ZONE', 'name': u'Hall', 'id': u'1643302'}
pi@pi1monitor:~ $ nano evohome_test
pi@pi1monitor:~ $ ./evohome_test
{'temp': 22.11, 'setpoint': 22.0, 'thermostat': u'EMEA_ZONE', 'name': u'Bathroom', 'id': 1377389}
{'temp': 20.77, 'setpoint': 20.0, 'thermostat': u'EMEA_ZONE', 'name': u'Hall', 'id': 1643302}
{'temp': 21.86, 'setpoint': 20.0, 'thermostat': u'EMEA_ZONE', 'name': u'Kitchen', 'id': 1643296}
{'temp': 21.01, 'setpoint': 21.0, 'thermostat': u'EMEA_ZONE', 'name': u'Living Room', 'id': 1378782}
{'temp': 18.98, 'setpoint': 18.0, 'thermostat': u'EMEA_ZONE', 'name': u'Main Bedroom', 'id': 1438045}

So apparently this change only applies to the (deprecated ?) V1 API at the moment!

I'll switch my munin-evohome graph plugin back to the V1 API for now, its only a one line tweak in the script to do so. (I think in the evohome-munin plugin Git repo, the V1 API is already the default, but I had changed it to V2 recently on my local copy) The domoticz plugin I don't think can be made to use the older API - anyone know ? It would be nice to get the full un-biased resolution on my Domoticz dashboard.

BTW as well as not being rounded to 0.5 degrees it appears the figures do not have the set point bias either. Check out the Main Bedroom temperature which says 18.5 on the rounded output with an 18 degree setpoint even though the "raw" figure is 18.98 - clearly biased towards the setpoint on the rounded figures...

I checked my controller and it is still running the same firmware, so it appears the evohome has been sending raw un-modified temperatures to the servers all along... which means all Evohome devices (evotouch controller, HR92, DTS92, honeywell servers) deal with the raw figures internally and then independently apply a setpoint bias and rounding before local display. (Or in the sever case, before returning figures via the API)

This probably explains why for example a DTS92 might consistently disagree by 0.5 degrees with the Evotouch, as their biasing/rounding algorithm might be slightly different and give a different result for the same raw value which straddles two rounded figures. I've sometimes noticed this on my DTS92.

As the phone apps use the V2 API they are currently still displaying setpoint biased rounded figures. It will be interesting to see whether increased resolution is extended to the V2 API in future.

[DBMandrake]

Had a quick look at the domoticz evo-update.sh script which I'm using from here:

https://www.domoticz.com/wiki/Evohom...ller_or_RFG100

Unfortunately it's not just a simple matter of flipping the python import to the old API like I can do in my munin script, because it is also retrieving information other than just temperatures which is not available in the V1 API.

So it would be necessary to rewrite the script a bit to use both API's, to retrieve temperatures via the V1 API and everything else via the V2 API. A little bit of a pain but not too difficult, so if nobody else steps forward I'll have a go at hacking this in to the script.

Does anyone else here actually use Domoticz with the Honeywell API instead of an HGI80?

[victorp1612]

Does anyone else here actually use Domoticz with the Honeywell API instead of an HGI80?

Yep, I Do.
I did a quick check this morning and noticed a strange one : only 1 out of my 9 rooms is now showing more granular info.
I'll check again when i get home tonight and report back.

[DBMandrake]

Are you sure you're seeing more granular info on that one zone ? The existing domoticz script polls the V2 API which still only gives 0.5 degree resolution, so I don't really see how that one zone could have additional resolution.

Is that from looking at the exact temperatures on the temperature dashboard or are you judging by graphs ? The graphs will have interpolation which may make it appear there is more detail than there really is.

I'll see if I get a chance to try adapting the script tonight or maybe tomorrow.