Closed mawinkler closed 2 years ago
Hi Markus,
The update you pushed to better handle the missing forecast seems to have had an unintended effect of generating way more errors (999 occurrences):
This error originated from a custom integration.
Logger: homeassistant
Source: custom_components/astroweather/entity.py:33
Integration: AstroWeather (documentation, issues)
First occurred: 00:33:00 (999 occurrences)
Last logged: 08:52:04
Error doing job: Task exception was never retrieved
Traceback (most recent call last):
File "/usr/src/homeassistant/homeassistant/helpers/entity_platform.py", line 735, in _update_entity_states
await asyncio.gather(*tasks)
File "/usr/src/homeassistant/homeassistant/helpers/entity.py", line 540, in async_update_ha_state
self._async_write_ha_state()
File "/usr/src/homeassistant/homeassistant/helpers/entity.py", line 590, in _async_write_ha_state
state = self._stringify_state()
File "/usr/src/homeassistant/homeassistant/helpers/entity.py", line 559, in _stringify_state
if (state := self.state) is None:
File "/usr/src/homeassistant/homeassistant/components/weather/__init__.py", line 306, in state
return self.condition
File "/config/custom_components/astroweather/weather.py", line 158, in condition
return self._current.condition_plain
File "/config/custom_components/astroweather/entity.py", line 33, in _current
return self.coordinator.data[0]
IndexError: list index out of range
Hi Tom, this bug was revealed when there was no data available at all. While fixing it the backend got alive again as well about 30mins ago. Issue 7timer
For now, it should be sufficient to just reload the astroweather integration and all should turn colorful on your side again. I will push the change in my entity.py with the next release to prevent the error you noted to occur again (if 7timer goes down again).
Switch back to 0.20.7fixed that Problem for me.
Backend problems were resolved and service seems to be stable again.
Is it not working again? Been a few days and still showing everything is unavailable. tried reloading the integration as well as rebooting the server and it didn't help like it did a couple weeks ago.
Hm, I'm not seeing any errors and it works flawlessly on my side. Do you have any logs to provide?
Where can I get the logs?
in home-assistant.log or via the frontend Configuration --> Server Controls --> Logs
Hi, can it be, that at the time the error occurred you had very bad conditions? If so, I just released a new card which fixes a bug in the detection of the weather entity. Give it a try.
Actually everything went unavailable and it added new devices but it didn't add all of them. I look at the forcast today & tomorrow the most since it is a simple gauge but noticed on the card today that it's completely wrong. It's raining all day and it shows 0% clouds. Not sure how to add a screenshot but here is a link to one I just took. https://photos.app.goo.gl/YqydWQ4Hbc1EUAJN7
I downloaded & added the card yesterday afternoon (cst), is a newer one available?
It shows 0% cloudless, so fully clouded. There should be a new version of the card available after your next update run of hacs
ok, so then the chart in the middle is actually CloudLESS not Clouds as it says..correct?
Any way to get the forcast today & tomorrow working again?
As I can see the forecast today and tomorrow seem to work on your screenshot?? It looks pretty the same as on my screenshot here card. What are you missing?
Question: How did you build your allsky camera? Is it self-made or a commercial one?
Self made. Running on a Pi using an ASI120mc inside an old CCTV dome camera housing. Attaches to a pole that goes to a box where the Pi, & relays are located. Has 2 5w ceramic resisters mounted in the dome for dew/heat. Also have DHT22 sensors in the dome that will trigger the relays to turn on/off. I need to do some changes inside the dome to get rid of the red paint reflection from the camera. you should be able to see it at http://allsky.duckdns.org
I manually set the status but after reboot those 2 went back to unavailable.
Guess I will end up using the deepsky forcast. The higher the percentage the better the viewing correct?
The cloudless is confusing. Shows 12% but then Total Cloud Cover less then 20% doesn't make sense. I read cloudless 12% meaning there is 82% cloud cover. Getting hung up on the words I think more then anything and not understanding them.
Think there is a TOD issue. You can see from my last screenshot that tomorrow looked better then today, but tonight at 10pm CST it says today is at 76% and tomorrow is bad. Happened to notice this when going out to cover the scope up for the day.
So, the logic is as follows: I'm calculating the condition based on transparency, seeing, and cloud cover. When you think about the influence of these measurements it obviously means the lower the measurement is, the better for viewing conditions. This also means if I want to calculate an easy-to-read overall condition value, which at its max has the highest value, I need to invert the measurements. In the top right of your screenshot you see the values in bolt Condition, Cloudless, Seeing, and Transparency. Condition is calculated based on the other three. In the table below, you have actually the same values, the current and the forecast.
Flip
Guess I will end up using the deepsky forcast. The higher the percentage the better the viewing correct?
The cloudless is confusing. Shows 12% but then Total Cloud Cover less then 20% doesn't make sense. I read cloudless 12% meaning there is 82% cloud cover. Getting hung up on the words I think more then anything and not understanding them.
Oh, the top part of the card displays the current situation. The second shows the forecast for this evening and tomorrow evening in plain text. The lower part is the forecast in a three hour interval.
The evening forecasts for Today and Tomorrow are calculated on the forecast based using your local time and a timeframe of 9hs. From a comment In my code:
# between 9pm to 3am.
# Possible timestamps within the data:
# 15 18 (21 00 03) 06 09 12
# 16 (19 22 01) 04 07 10 13
# 17 (20 23 02) 05 08 11 14
# Relevant ones in brackets
Since I'm only having data based on three hours intervals the time interval for the calculation might start at 7pm, 8pm, or 9pm depending on your timezone. If it's now for example 10pm your local time, I miss the start for today and flip over to tomorrow and the day after. I know that this is not perfect but for the current night you don't need a forecast anymore. I think the best would be to attach a date instead today
or tomorrow
.
What do you think about this?
Think it looks good. Some feedback Cloudless - Still doesn't make sense. In the US we just dont refer to things in this way. We look at things as how much cloud cover there is, not now much there isn't Chart - Without hovering over the points there is no way to know what each of those lines are. Think the colors of the lines should match the color of text above it or have labels on the chart.
Would still like a simple gauge like you had before as an entity for the forecast. The deep sky forecast will work but for simple quick look the gauge is what I looked at 99% of the time.
Here is another project I did. Something cheaper then Bluetti or Jackery. I have the LCD also sending data out via MQTT as well now and eventually looking at building a plugin for Nina to show battery status and input into the advance sequence. https://diysolarforum.com/threads/astrophography-build.36487/
I totally agree, cloud coverage is much more common, even here in Europe. Using cloud coverage would then mean, that the line in the chart would be in the top area if there are many clouds, but this would be bad for the viewing conditions of course. I was thinking more of a way of creating a graphical display that directly shows the viewing conditions' quality. Same for the distinct values. If cloudless is 100%, this means from the perspective of the clouds my condition will be good. Compared to that, if I'd use cloud cover, it would mean I'm going to stay inside today. If that makes kind of sense.
Line colors agreed. I had that in the beginning, but ppl are using different skins and some combinations look very weird. So I just stuck to the default and highlight color for now.
For the gauge, you can still do that. Just use the sensors:
with this yaml code:
- type: entities
entities:
- positions:
minmax: inside
entity_row: true
min: 0
max: 100
target: 80
name: Today
type: custom:bar-card
entities:
- entity: sensor.astroweather_deepsky_forecast_today
- positions:
minmax: inside
entity_row: true
min: 0
max: 100
target: 80
name: Tomorrow
type: custom:bar-card
entities:
- entity: sensor.astroweather_deepsky_forecast_tomorrow
Dont get me wrong, I like all the extra data..and the more data you have the better. Think it looks great, just takes some understanding is all and getting used to it. I do mainly AP and not so much visual and right now my AP is all over the place from solar, lunar, deep sky & comets. Working through multiple Astro League programs. Normally I just do deep sky & some planetary.
I like the horizontal guage better, going to move over to that.
Figured the colors for the chart would be skin dependent and difficult to code for, that's why I thought a legend might be better.
I'll get used to it no doubt. Will be using it more due to Spring & rain showers.
All good :-) I really like any kind of constructive feedback, no doubt. I just pushed some changes to the repos. I didn't do new releases until now, but if you redownload with hacs from the main branch you'll have the latest version of the card and the integration itself. So you need to reload both from main. It's still using cloudless, but your other proposals are in.
Your level of AP sounds impressive! I just started with that hobby about 15 months ago but I'm now totally addicted to it. I wish we had more clear skies here in Bavaria, though. Are you sharing your photographs (e.g. on Astrobin) or having a web page?
I only started last summer myself. Right now just have some stuff in my google photos. Here is a link to some. That red thing is my arduino based led light panel, 3d printed. The Bodes Galaxy is one of the best images I have done so far with my QHY camera. The first few images were taken back when I was just using my Nikon cameras & lenses. Moved up to an 80mm tripplet now and have an 8" sct for planetary (for the moment)
https://photos.app.goo.gl/P7eHuZckqBTNTenx8
On Wed, Apr 6, 2022 at 1:38 PM Markus Winkler @.***> wrote:
All good :-) I really like any kind of constructive feedback, no doubt. I just pushed some changes to the repos. I didn't do new releases until now, but if you redownload with hacs from the main branch you'll have the latest version of the card and the integration itself. So you need to reload both from main. It's still using cloudless, but your other proposals are in.
Your level of AP sounds impressive! I just started with that hobby about 15 months ago but I'm now totally addicted to it. I wish we had more clear skies here in Bavaria, though. Are you sharing your photographs (e.g. on Astrobin) or having a web page?
— Reply to this email directly, view it on GitHub https://github.com/mawinkler/astroweather/issues/11#issuecomment-1090603147, or unsubscribe https://github.com/notifications/unsubscribe-auth/AUN2CFLY433TQTW27BACTXLVDXK25ANCNFSM5QZMCZKA . You are receiving this because you commented.Message ID: @.***>
Nice Here are some of mine. I'm owning a 10" Newton with a QHY268c and do have captured the M81 & M82 as well ;-) Astrobin
Wow...far better then mine
On Wed, Apr 6, 2022 at 2:02 PM Markus Winkler @.***> wrote:
Nice Here are some of mine. I'm owning a 10" Newton with a QHY268c and do have captured the M81 & M82 as well ;-) Astrobin https://www.astrobin.com/users/maralex/
— Reply to this email directly, view it on GitHub https://github.com/mawinkler/astroweather/issues/11#issuecomment-1090644093, or unsubscribe https://github.com/notifications/unsubscribe-auth/AUN2CFJ3KD32DKDB3IAYTGDVDXNUPANCNFSM5QZMCZKA . You are receiving this because you commented.Message ID: @.***>
There seems to be a problem on the 7timer side accessing the API endpoints of NOAA. It is currently under investigation. I do not have any information on when this will be fixed, but hopefully soon.