rpi-weather

Python 3.9 application to get local weather forecast and display results via icons on LED 8x8 matrices.

Hardware

Overview

Here's a general overview about the hardware setup to give you some orientation: https://caternuson.github.io/rpi-weather/. The wodden case has roughly these dimensions: 19cm (length) x 13cm (width) x 8cm (height). The jumper cables should have a length of 16-18 cm.

This program should work with any Raspberry Pi although I have only tested it with an original Model A (yes, an A). Also, any 4 Adafruit 8x8 LED Matrices with I2C Backpacks should work.

Soldering the address jumpers

Be sure to solder the address jumpers to set unique addresses for each. Expected range is 0x70-0x73. Make sure you solder the pins of the matrices in the right direction for this project (in other words, do not follow the soldering instructions as presented in this tutorial, the pins have to be the other way around )

Software

A brief description of the various software components.

weather.py - gets and displays forecast from NOAA (US ONLY)
weather_metoffice.py - gets and displays forecast from metoffice.gov.uk (UK ONLY)
weather_forecastio.py - gets and displays forecast from forecast.io
weather_openweather.py - gets and displays forecast from openweathermap.org
rpi_weather.py - defines a class for interfacing with the hardware
led8x8icons.py - contains a dictionary of icons
clock.py - displays the time, for use as a clock

Dependencies

Optional: Before installing any software, it is convenient to set up your raspberry pi using the rpi imager (so you can ssh into it, or even use a remote desktop client to get a graphical user interface). If you are using Windows on your hostmachine and you want to use the Remote Desktop Client make sure that xrdp is installed on the pi by running sudo apt install xrdp.

You need to then install Python & the Adafruit_CircuitPython_HT16K33 Library so that your Rasperry Pi can communicate with the Adafruit LED matrices. For that you can follow this tutorial.

Install

Simply clone this repo and run:

$ git clone https://github.com/caternuson/rpi-weather.git
$ cd rpi-weather
$ sudo python weather.py

Configure (NOAA)

The forecast location is specified with a zipcode. A default zipcode can be set in the code:

ZIPCODE = 98109

A zipcode can also be passed in from the command line, which will override the default:

$ sudo python weather.py 98109

Configure (metoffice.gov.uk)

You will need an API key to use these services, available here. You will also need to determine your location ID. To get that you need to hit this url:

http://datapoint.metoffice.gov.uk/public/data/val/wxfcs/all/json/sitelist?key=[YOUR_API_KEY_HERE]

and search for your nearest location. If you can't find the location you are after you can hit the main site and use their front-end search to find the closest location they have, then get the code for that from the api response above.

Once you have this info, create a file called weather.cfg with the following contents:

[config]
API_KEY: your_api_key
LOCATION_ID: your_location_id

replacing the your_* info as needed.

Configure (forecast.io and openweathermap.org)

You will need an API key to use these services. Each website has instructions for how to do this. You will also need the latitude and longitude for your location. Once you have this info, create a file called weather.cfg with the following contents:

[config]
APIKEY: your_api_key
LAT: your_latitude
LON: your_longitude

replacing the your_* info as needed. NOTE: west longitudes are negative, use decimal values for both.

Automation

The easiest way to have the program run on a daily basis is to use cron. Use crontab -e to add the following entry, which will run the program every morning at 4AM:

0 4 * * * sudo -E PYTHONPATH=$PYTHONPATH python /home/pi/rpi-weather/weather.py

NOTE: If you installed the program in a different location, change the path accordingly.

NOAA REST

The forecast is determined using the NOAA REST web service. Specifically, the Summarized Data for One or More Zipcodes. A typical request looks like:

http://graphical.weather.gov/xml/sample_products/browser_interface/ndfdBrowserClientByDay.php?zipCodeList=98109&format=12+hourly&numDays=4

The key bits being:

zipCodeList - zipcode(s) for forecast
format - choose either 12 or 24 hour period
numDays - number of days in forecast

The request returns XML data. The icons are set by a simple text search in the weather-summary attribute of the weather-conditions tag.

forecast.io and openweathermap.org

An ICON_MAP is defined to map forecast results to a specific LED 8x8 icon.

Icons

Icon	Weather
	SUNNY
	RAIN
	CLOUD
	SHOWERS
	STORM
	SNOW
	none of the above

JohannesWiesner/rpi-weather