MMM-Hoymiles-Wifi

schris88

@sdetweil said in MMM-Hoymiles-Wifi:

i bet the config is already passed to the node helper which execs the python script.

Actually the python script is being executed via a shell script (in order to make it running using pm2).
I do not see anything in the node_helper.js file.

This is mainly due to the fact that the developer of the module used MMM-HTMLSnippet as basis and concentrated more on the data that he wanted to display.

The challenge is now to have everything running inside the module, without this ‘external’ stuff.
As all modules tend to do.

This is true, it´s a hacked proof of concept, nothing to be proud of at the moment :D

Not sure how much time I have for it till end of year, I can grant write access to anyone who´s willing to help.

evroom

@schris88 said in MMM-Hoymiles-Wifi:

This is true, it´s a hacked proof of concept, nothing to be proud of at the moment :D

Not sure how much time I have for it till end of year, I can grant write access to anyone who´s willing to help.

Not to worry. You sparked an idea.

I have some modifications ready (I forked it), but fighting with github to get it synced (github wants to delete all my modifications when I try).
Perhaps granting me access (evroom) wil make it easier.
Although it is not really the way to use github, I guess.

E.J.

sdetweil

@evroom

you fork
clone fork //git clone
make new branch // git checkout -b nane
make code changes
// this part multiple times
add to files to commit // git add filename
git commit
git push // to add to github fork
//

on github go to original repo source
create pull request // for origin to pull changes from your fork

if subsequent push to same branch, PR will be updated automatically

evroom

@sdetweil said in MMM-Hoymiles-Wifi:

git commit
git push // to add to github fork

I am stuck here.
Will open a new topic for it …

evroom

Hi there,

I made some major changes to the code and hopefully did all the GitHub stuff correct in order to allow persons interested to give it a (new) try.

https://github.com/schris88/MMM-Hoymiles-Wifi

There is still a lot to do (see ‘TODO’ in the README file), but I think the basis provided by Christian (@schris88) has been strengthened.

Best regards,

E.J.

evroom

@schris88

Christian,

This is on the TODO list:
change pv_data to dtu date for combined output of inverter

But I do not really understand what you mean by it.
Do you mean combined output of 2 panels iso just 1 ?

Please elaborate.

E.J.

schris88

@evroom Hi there, maybe you can confirm, if the inverter has multiple inputs there seems to be an array of pv_data objects in the json (as seen in the test.json you pushed)

so the meter should combine all pv data power to have the total, right.
But we dont need to do that by hand because the combined value should be in dtu_power already, you can get rid of all the try except.

I just didnt change it in the beginning because for inverters with only one input it makes no difference because its pv_data power and dtu_power is always the same value.

evroom

@schris88

First of all, those are details I need to have a look at now that the basis is there.
Perhaps there are other values and queries that make sense and that I can use. Like dtu_power and dtu_daily_energy.
Also saving some data to not have counters showing 0 when the invertor is inactive needs to be implemented.
During testing I found out that the pv_data for port 1 and 2 were different sometimes and that is why I needed to add those checks.
And they can never hurt. Better to be safe than sorry.
Furthermore, when you have only 1 panel connected, it is needed anyway.

evroom

@schris88

Just FYI, I declare(d) more parameters than really necessary and actually used, in order to allow for more counters to display.
For example a gauge for each panel; more details on the energy production of each panel; etc.
I will make a remark about this in the code.

schris88

@evroom That sounds good, but we should loop through the pv_data objects and display each as a graph. In the end, there could be a config setting like detailed: true to activate the display of all pv_data objects as individual graphs. If set to false, we could display a single graph with the total power.

There are Hoymiles inverters with 6 inputs, and adding 6 × 6 to 8 variables can quickly add up. This might not be an issue for us, but it’s worth considering if the module should be as generic as possible.

Sorry I´m a IT-Consultant who, sometimes it breaks through :D