Display the output of multiple temperature stations on one page.
For instance, you can set up several Raspberry Pis around the house, each with a temperature/humidity sensor, and display the output from all of them on one page, along with reports from your official weather station or any other weather station that you can fetch via an API. See the client directory for types of weather reports supported so far.
Reports are also logged, and there is a very limited facility for reporting historic data. I hope to improve that eventually and add graphs.
This is not limited to weather data: you could use the framework for any quantities you want to check, report and log. It would be better to have a more general name, and if I think of a good name I'll rename the repository.
To test the watchweather server, first create the needed configuration files:
mkdir ~/.config/watchweather/
cat >~/.config/watchweather/fields <<EOF
temperature
humidity
rain
EOF
Then run the server:
export FLASK_APP=server/watchserver.py
flask run
If you want to disable debug mode so you can access the server from another machine:
flask run --host=0.0.0.0
For more debugging messages, try
export FLASK_DEBUG=1
though this is deceptive since debug mode is already on.
The clients report data to the server using the program client/stationreport.py.
There are several client modules in the client/ directory that run on Raspberry Pi using sensors like Si7021 or HTU21D.
There's also a client module called observerscraper that scrapes the web page from an Ambient Weather Observer. The Observer can have more than one sensor (for instance, Outdoor and Indoor) so the observerscraper client includes sub-stations which will make separate reports.
Run a single report by specifying which module you want to use, e.g.:
client/stationreport.py "Sensor Location" servername Si7021
If your sensors are reporting correctly, you can make reports in a loop, like this:
nohup client/stationreport.py -l Observer moon Si7021 &
or while debugging, you can do a simpler test, like:
while true; do
client/stationreport.py Location servername Si7021
sleep 30
done
You can also use -p port, if you need a port other than 5000, and -v for verbose if you want to keep track of your reports.
For writing your own sensor, see the client/README.md.
The base page gives a simple list of options:
[http://localhost:5000/](http://localhost:5000/)
This is an overview that shows all stations in a pretty, large-text format. It's good for keeping an eye on a few variables, like temperature and pressure, for lots of stations. It's currently not so good for stations that report a lot of variables.
[http://localhost:5000/stations](http://localhost:5000/stations)
[http://localhost:5000/details/all](http://localhost:5000/details/all)
This shows a table of all parameters known about a station. It's better for stations like the observerscraper that report a long list of values.
You can use "all" as the station name to see a table of all stations that have reported.
[http://localhost:5000/details/all](http://localhost:5000/details/all)
A table of selected data reported by the station in the last 7 days. Useful for keeping track of quantities like daily rainfall that get reset each day when you didn't stay glued to watchweather until midnight.
This is the page the clients use to make their reports.
Some pages, like /stations, auto-refresh. Some versions of Firefox disable this by default. To fix that, go to about:config, search for refresh and if accessibility.blockautorefresh is true, double-click on that line to change it to false. Supposedly the default is false but something is changing it to true for me and a lot of people on the web.
If you don't have any sensors yet and just want to test the server, you can add some dummy station data using 'test' as the sensor name:
client/stationreport.py Nearby localhost testclient
client/stationreport.py 'Far away' localhost testclient
If you have a running instance on a server, with stored data, and you want to test a local instance against that data, just copy the data over. For instance, if you're running watchweather on a production server and caching data in /var/www/watchweather/.cache, and locally you're using the default of ~/.cache/watchweather, then:
rsync -av servername:/var/www/watchweather/.cache/watchserver/