Cumulus Support

HansR wrote: ↑Mon 28 Oct 2019 6:15 am I see you're hacking your way around the line thickness

Lol,

Yes and I require TWO script lines as you have an extra space on line 2348 before "color".
Wasting a precious byte...

But Yes, I find it instantly more comprehendible to my ailing brain.

Cheers.

Congratulations Hans, the graphs module is amazing.

One issue I have on my pages is the degree symbol showings as Temp. (°C)
in both www.komokaweather.ca/wxFWI.php and www.komokaweather.ca/wxGraphs.php

I am using the same Saratoga sample script as for www.komokaweather.ca/wxTop_10.php where it appears correct.
I presume this is caused by a charset=isoxxxx but have not been able to locate where that might be. Any suggestion where to look?

Enjoy,
Paul

Hi Paul,

Thnx for the compliment and good to see you're still here. Had not seen you for some time.
The °C must have to do with some character encoding thing, I agree. It surprises me in the sense that I did not notice when you gave me the link some time ago when you had the FWI page first time. (and btw: I now understand why and what beteljuice communicated with me )

The character encoding is not my strong point and I would rather not start encoding 16 bit characters as betel suggested. But I'll look into it.
Currently busy with something else, so hopefully you have some patience.

Btw, you use the same template as I do and we both have the same <meta http-equiv="Content-Type" content="text/html; charset=ISO-8859-1" /> meta in the source. So it must be something deep. You're the only one who has this issue, so what is your deviation

Best, Hans

@Phil and @MeteoSangonera:

You can view the first live try of pwsFWI with a 5 day prediction on my site.
Release may take some time because of testing, checking and errorhandling to add. I will also add some indication that the viewer actually sees a prediction. And apart from that some text needs to be added, conditionally because the user can choose not to use prediction.

Comment is always welcome.

Cheers, Hans

Great job.

I think it would be more interesting to separate the prediction of the measured data into two tables. it would be more obvious for those who consult the data.

The API is quite simple to use, isn't it?

The most interesting thing now is to check if the prediction fits more or less to the data that will be obtained

HansR wrote: ↑Mon 28 Oct 2019 9:04 pm You can view the first live try of pwsFWI with a 5 day prediction on my site.

Looking forward to seeing it live.

Release may take some time because of testing, checking and errorhandling to add.

Always willing to be a test & fail user here.

I will also add some indication that the viewer actually sees a prediction.

That was going to be my initial ccomment before I read the hole post.
Was thinking the background for the prediction rows would be a different colour.


Presuming this is using the DarkSky API?

Phil.

PaulMy wrote: ↑Mon 28 Oct 2019 2:14 pm One issue I have on my pages is the degree symbol showings as Temp. (°C)

OK, will be solved in the next release.

Version 1.8.0 is online

Please note this version contains a prediction addition to the pwsFWI module.
That means a significant complexity has been added, not the least in usage.
Read the part of pwsFWI in the ReadMe, it has significantly changed and pay attention to the Known Issues.

I am interested in the results i.e. the calculated value against the predicted value. To not make it too complex, I would like to ask you (if you will be using the prediction feature) to note the predictive value two days ahead. So, today, the 30th/10 you would note the value of 1/11 and when that day has past, you will note the actual value with it. Do it until you're fed up with it and then you send it to me

I am really curious how this will behave. Any problem, let me know.

Cheers, Hans

Hi Hans,
I have updated to v1.8.
I registered for yourweather API and got the Activate your account email, and clicked to Activate but can't log in yet, and have sent a message to them through the Contact.

Will let you know on Prediction when I have access to their API page.

Enjoy,
Paul

@PaulMy: That is very strange Paul, with me it was 'instantaneous'. Try going to the main page and then click on the API link again, that should give a fresh login page. Otherwise I really wouldn't know. You did not accidently mistype a letter? Anyway,good luck!

Well, that is nice: my almost neighbour, @hvd51, at a bird fly distance of 16 km North of my station has implemented the pwsFWI prediction first (Thanks Harm!). This immediately brings to attention a side effect which can easily be seen through the tabular display of the predictions (which may change btw every time you query the system):

1) the differences between local meteorology can be very big (most notably in this case wind speed: 100 vs 30 and 82 vs 22 km/hr for 2 and 3 November)

OR

2) the differences between predictions which may vary a lot, even between queries (if I query 4 hours later I may get a very different prediction)

Interesting exercise if it were only for these observations!

The prediction is operational on my website. http://www.meteosangonera.es/fwi/.

I think there's a bug in methodology.

For the calculated data, you use the minimum humidity of the day. (In my case for 29/10/2019 = 35%), but for the calculated data the table shows the predicted average humidity value (for example in my case for 30/10/2019: Daily API humidity value=67%. This value is the average of the day, having a maximum of 88% and a minimum of 39%).

@meteosangonera:

meteosangonera wrote: ↑Wed 30 Oct 2019 6:51 pm The prediction is operational on my website. http://www.meteosangonera.es/fwi/.

Thanks, looks great.

meteosangonera wrote: ↑Wed 30 Oct 2019 6:51 pm I think there's a bug in methodology.

That seems quite a big word. My method of calculation does not change because of an API.

meteosangonera wrote: ↑Wed 30 Oct 2019 6:51 pm For the calculated data, you use the minimum humidity of the day. (In my case for 29/10/2019 = 35%), but for the calculated data the table shows the predicted average humidity value (for example in my case for 30/10/2019: Daily API humidity value=67%. This value is the average of the day, having a maximum of 88% and a minimum of 39%).

I guess I took the simple route and just got the first value. Sorry for that, too much focus on the algorithms.
I get your point, so to get the minimum value I have to search all hour blocks to get the minimum humidity value. Same holds for the average wind I guess.

I take that with me as an improvement, not as a bug in methodology.

I'm sorry if it bothered you. The translator plays bad tricks on me. You do a great job.

No, does not bother me, just to get things straight. Non-English myself, get lost in translation too often