Bug fix IKEA VINDSTYRKA: Translate VOC index to tVOC level.

[ebaauw]

The IKEA VINDSTYRKA doesn't report tVOC level, but some VOC index instead, see #7246.

This PR translates the VOC index to tVOC level, using the translation table from:
https://learn.kaiterra.com/hs-fs/hubfs/Email%20Series/IAQ%20Series/AQI%20tvoc.png?width=1932&name=AQI%20tvoc.png:

[SwoopX]

@ebaauw Maybe I can convince you on the Develco approach, as it is a bit leaner?

{
          "name": "state/airqualityppb",
          "awake": true,
          "parse": {
            "at": "0x0000",
            "cl": "0xfc03",
            "ep": 38,
            "fn": "zcl:attr",
            "mf": "0x1015",
            "eval": "if (Attr.val != 65535) { Item.val = Attr.val; }"
          }
        },
        {
          "name": "state/airquality",
          "awake": true,
          "parse": {
            "fn": "numtostr",
            "srcitem": "state/airqualityppb",
            "op": "le", "to": [65, "excellent", 220, "good", 660, "moderate", 5000, "unhealthy", 65535, "out of scale"]
          }
        },

[ebaauw]

Maybe I can convince you on the Develco approach, as it is a bit leaner?

Sorry, I don't understand what you mean?

• The VINDSTYRKA doesn't report tVOC density in ppb, but some index value instead, which needs to be translated to the tVOC density, following the listed table.
• I thought we decided we wanted to move towards measured_value and corresponding capabilities?

[LeoeLeoeL]

Please check this:

The "excellent/good/moderate etc." seems to be updated non correctly.

[ebaauw]

Please check this:

I have no idea what "this" is. Note that "this" uses six colours, where the deCONZ API only reports five values, so my guess would be the graphs attach their own labels, based on airqualityppb, instead of using airquality.

Please list the ZHAAirQuality API resource, after reading the VINDSTYRKA's 0xFC7E cluster attributes in the GUI. And please double-check that you're using the DDF from this PR, having installed both the DDF and the JS script file to /usr/share/deCONZ/devices/ikea.

Note that measured_value, airqualityppb and airquality are updated when reading or receiving a report from 0xFC7E/0x0000. Also note that PM2.5 has a different scale, which is used for airquality in the ZHAParticulateMatter resource.

The ZHAAiqQuality should report the same airquality values as the other ZHAAirQuality sensors, since I simply copied the numtostr logic from those:

From	To	AirQuality
0	65	excellent
66	220	good
221	660	moderate
661	5000	unhealthy
5001	65535	out of scale

I don't know the source of this mapping; in particular, I don't know where the 65 threshold comes from, nor why unhealthy doesn't go up to 5500. Last time I checked, this is a jungle, with different organisations / government bodies from different countries recommending/prescribing different thresholds, labels or even number of labels used.

[LeoeLeoeL]

Don't look at the colours. This colour code is for PM 2.5. I only used that for TVOC.

But look those values:
Column 1: 83 - Excellent
Column 2: 150 - Good
Column 3: 206 - Excellent
How can Column 3 be excellent with a value higher than in column 2 that is good?
I verified; The "excellent/good/moderate etc." seems to not be updated when they need.
I came back to 2.23.2 and everything works ok again.

P.S.:
The mapping of PM25 in the DDF (2.32.2) is:
10 excellent,
20 good,
25 moderate,
50 poor,
75 unhealthy,
65535 out of scale
but shoul be:
0 Good
13 Moderate
36 Unhealthy for sensitive groups
56 Unhealthy
151 Very unhealthy
251 Hazardous

[ebaauw]

@LeoeLeoeL, please open new issues for the PM2.5 mapping, including the source for this mapping, and the airquality not being updated issue. This PR is about the tVOC index for the Vindstyrka.