EvoControl - custom Alexa skill for Evohome

[evoboff]

When you have more than one location, a new button is added to the bottom right of the dashboard which allows location-selection by touch inputs.

When you only have a single location, for privacy purposes the location address is not retrieved and you are only told you have 'a single location' when you list.

[philchillbill]

I've changed the code to just use the name. If you're already using the skill, say "Update my Configuration" to pick up the changes. Any new users will get the change automatically.

EDIT: if both systems are at the same address, would it be useful to display all zones from both systems simultaneously? If zone names are unique then this should be doable.

[philchillbill]

V1.2 is live with redesigned buttons & new schedule editor



V1.2 of the skill is now live with several improvements and new features. No need to update anything as you always run the latest code when you invoke the skill. Summary:

• The most striking change is visual and comes from scrapping all the AlexaButtons and moving to AlexaIconButtons. The rather wide minimum width of the AlexaButton means I could place at most 4 buttons side-by-side on smaller screens while the skill needs up to 7 buttons in places. The disadvantage of AlexaIconButtons is that they are unlabelled. I've now created my own captioning for them to overcome this limitation. Being able to have more buttons on the screen at the same time means that in some parts of the UI, the nesting level has changed for the better. For many schedule-related screens, colour-use and opacity-use are more consistent and the fontSize adapts not only to screen size of the Echo but to the number of switchpoints to be displayed.
• The panel icon has been replaced by a little pictogram of the Evotouch.
• The bulk switchpoint editor shown in post #109 is now live. You can use it to e.g. lower or raise all Fetch TCC button in the schedule view. If you want to save a schedule that you freshly modified outside the skill (either on your Evotouch or in the TCC app) to an ABC preset, this button will grab the changes from TCC in case you already had the skill open with stale data. If you opened the skill after making the changes then the skill would of course already have the newest live schedule.
• The Scene Editor now launches differently depending on whether you already have defined one or more scenes, or your list is empty. If empty, a blank editor template is shown with a minimum of buttons (much more intuitive). If you have already defined any scenes then the viewer is launched without having to select view-mode specifically. There's then a '+' button in that view to add a new scene.
• The opacity of the buttons in the dashboard footer has changed to make them stand out more. They did look a trifle disabled/greyed-out before.
• The APL code for all visuals was tweaked so that the various UI-screens look better on non-standard Alexa-enabled devices. For example, the Fire TV stick, the big picture-frame 15" Echo, the small round Echo Spot. I don't have all of these devices myself so it was tweaked in the Amazon APL simulator environment. I hope it all looks correct in reality (there are known limitations of the simulator environment). Let me know otherwise.
• The Groups screen got a visual makeover and just like with scenes it launches differently depending on whether you already have defined any groups or not.
• Both Scenes and Groups can now be renamed.
• A few minor bugs were solved pertaining to the undo command(s) together with DHW in the mix.

Enjoy!

[evoboff]

If both systems are at the same address, would it be useful to display all zones from both systems simultaneously? If zone names are unique then this should be doable.

Yes, that would be great. They have unique zone names.

[nealroberts20]

[philchillbill]

Smart heat won’t open on my Portal today - start to load with hourglass - then purple screen with ‘tcc Token refresh declared a 400 invalid grant failure, there may be a temp service issue ….’
Was same this morning and this eve..

Honeywell seems to be revoking the validity of oauth tokens before they are supposed to expire. Their stated lifetime is 1799 seconds and I was waiting for that to pass before renewing because I assumed I could trust the stated validity. I've now changed the code to try and refresh if there's an early expiration but something seems to have goten out of lock-step. Can you login to your Dashboard and press the Re-Auth button for TCC so that you get a new token to start with?

[nealroberts20]

Done reauthoring - got it working again.

[philchillbill]

setpoints for each zone as you will normally want to tweak the setpoints based on the current setpoints. However, if you press 'Actuals' then the editor is populated instead with the current actuals for each zone. This could be useful if you are happy with current temperatures and want to maintain them, perhaps applying a small offset in this editor.

[philchillbill]

The UI for multiple locations has changed. You can now directly select 1 or 2 via dedicated buttons on the dashboard, while the currently-selected location number has a white border (just like the ABC buttons in the Schedule View). When you tap one of these buttons it will pulsate while the data for that location is loaded from TCC. Note that if you only have a single location then these buttons are hidden.

[philchillbill]

[Stevedh]

Is there an easy way to determine if this skill is using the Domoticz HGI80, preferably without looking at a show. I have noticed that my temperatures are now being reported at only 0.5 degree accuracy and asking about demand doesn't give a percentage.
I have checked the linking etc and it all seems fine.
I did lose my internet connection for the https link it uses for a couple of days.
so basically.
Is there an easy way to determine of it is using my fake HGI80.
and if not an easy way to fix it again.

[philchillbill]

Is there an easy way to determine if this skill is using the Domoticz HGI80, preferably without looking at a show. I have noticed that my temperatures are now being reported at only 0.5 degree accuracy and asking about demand doesn't give a percentage.
I have checked the linking etc and it all seems fine.
I did lose my internet connection for the https link it uses for a couple of days.
so basically.
Is there an easy way to determine of it is using my fake HGI80.
and if not an easy way to fix it again.

Take a look in the settings page and verify that the option 'Use HGI-80/Domoticz for control when possible' is enabled. Its auto-disables when you have Domoticz connectivity issues. The only way to check without looking at a screen is to say "Update my configuration" and listen for if the spoken response mentions Domoticz being successfully queried.

[Stevedh]

Take a look in the settings page and verify that the option 'Use HGI-80/Domoticz for control when possible' is enabled. Its auto-disables when you have Domoticz connectivity issues. The only way to check without looking at a screen is to say "Update my configuration" and listen for if the spoken response mentions Domoticz being successfully queried.

I tried that, just said had 9 zones like last time and dodgy zones were ignored or something like that.
So if it doesn't mention Domoticz... I should assume its disconnected. I did try Modifying and saving the HGI-80 setting on the EvoControl Dashboard, but that didn't change anything.

Ok I made the effort to go downstairs and have a look at my show and 'Use HGI-80/Domoticz for control when possible' was not enabled, enabled that and all working fine now.
Did an "Update my configuration" to see what it should say and heard, 'HIG80 device successfully queried' so I now know what to listen for.

Is there any way to do 'Use HGI-80/Domoticz for control when possible' without the show.

What would be really nice... although I expect not possible and way beyond the brief for this, would be to have a web page that could emulate the show for this.

[philchillbill]

I tried that, just said had 9 zones like last time and dodgy zones were ignored or something like that.
So if it doesn't mention Domoticz... I should assume its disconnected. I did try Modifying and saving the HGI-80 setting on the EvoControl Dashboard, but that didn't change anything.

Ok I made the effort to go downstairs and have a look at my show and 'Use HGI-80/Domoticz for control when possible' was not enabled, enabled that and all working fine now.
Did an "Update my configuration" to see what it should say and heard, 'HIG80 device successfully queried' so I now know what to listen for.

Is there any way to do 'Use HGI-80/Domoticz for control when possible' without the show.

What would be really nice... although I expect not possible and way beyond the brief for this, would be to have a web page that could emulate the show for this.

I can take a look at having Domoticz connectivity auto-retry every time the skill is launched after a failure and not immediately unselect the slider.

A WebApp for this would be a crazy amount of work. All the visuals are done in Alexa Presentation Language which is JSON based, so there's zero HTML or CSS to use as a basis.

[philchillbill]

Some users with Domoticz connectivity issues (ngrok URL changed on restart) are unlinking and then re-linking the skill to fix. That's complete overkill and will cause you to lose any saved schedules, activities, groups and other preferences because the database is cleared when you unlink! All that's needed is to login to your smartskills EvoControl dashboard and press the MODIFY button to the right of 'HGI-80 is connected via Domoticz' tickmark. Enter the new URL for ngrok and Verify and all will work again.

It may also help people to be aware of how a skill like this works. Its code runs in the cloud in an AWS Lambda container which is elastic and only allocated to this skill for milliseconds at a time on an as-needed basis. It's all event driven too. When you start the skill on an Echo Show, the Show keeps the display 'alive' for you even though the skill in the cloud has processed the previous event and long since 'evaporated' (the container is simply re-allocated to another skill for somebody else, etc). When you issue a new spoken or touch command, that new event is sent off to the cloud and the skill momentarily comes to life to handle it before evaporating again. Any response-time or latency differences you notice between successive commands is down to how long it takes for the Lambda container to come to life. That includes having to retrieve your session-state from short-term persistence, as Lambdas are actually stateless. If your Echo itself is busy doing something under the hood (like refreshing a calendar or whatever), response can be (unpredictably) slower for that reason even though you are unaware as to why. Perhaps worth knowing