Data from the Web

The only thing missing after the wallpaper text is the Internet connection, so you acquire the weather data. The decisive commands are shown in line 11 of Listing 3, which imports the SSID and the WiFi password from the secrets.py file. The code in lines 14-21 then establishes the connection with the ESP32 chip.

01 import board
02 import busio
03 from digitalio import DigitalInOut
04 import time
05 import neopixel
06 import displayio
07 from adafruit_esp32spi import adafruit_esp32spi, adafruit_esp32spi_wifimanager
08
09 # ---------------------------------------------
10
11 from secrets import secrets  # file secrets.py
12
13 def get_wifi(secrets):
14   esp32_ready = DigitalInOut(board.ESP_BUSY)
15   esp32_gpio0 = DigitalInOut(board.ESP_GPIO0)
16   esp32_reset = DigitalInOut(board.ESP_RESET)
17   esp32_cs    = DigitalInOut(board.ESP_CS)
18   spi         = busio.SPI(board.SCK, board.MOSI, board.MISO)
19   esp         = adafruit_esp32spi.ESP_SPIcontrol(spi, esp32_cs, esp32_ready, esp32_reset, esp32_gpio0)
20   status_rgb  = neopixel.NeoPixel(board.NEOPIXEL, 1, brightness=0.2)
21   return adafruit_esp32spi_wifimanager.ESPSPI_WiFiManager(esp, secrets, status_rgb)
22
23 # --- main-loop   ------------------------------
24
25 connection = get_wifi(secrets)
26 try:
27   response = connection.get("https://wttr.in/München?AT0&lang=en")
28   text = response.text
29 except:
30   text = "No data received"
31 group.append(get_label(text, FONT, COLOR))
32 display.show(group)
33
34 while True:
35   time.sleep(60)

Once the connection is up, you can send standard HTTP requests like GET and POST and evaluate the server response – typically by HTML, but with JSON to query data servers. In contrast, wttr.in returns plain text (Listing 3, lines 28 and 30), which the program outputs directly after converting it to a label (Figure 5). I adapted the URL in line 27 to suit the small PyPortal, so the output only shows the current weather data instead of a detailed multiple-day forecast. Change the city name before the question mark in the URL for your location, and use lang=en for English output.

Figure 5: The weather forecast for Munich, as requested from wttr.in, shows light fog. A picture of clouds was adapted for the background.

With fewer than 50 lines of code, the display shows the current weather data, but the program still lacks the ability to query the weather report regularly. To do this, you need to add a get() command to the infinite loop at the end of the program. Also, you might want to display the current time and the indoor temperature from the built-in sensor in the PyPortal.

Unfortunately, the temperature sensor in the PyPortal turns out to be a flop. The display backlighting heats it up, so it delivers useless values. As a consequence, Adafruit got rid of this sensor in the PyPortal Pynt and Titano variants. A better alternative would be to attach an alternative device (e.g., a BME280 or LM75) with a short cable to the I2C interface.

An extended version of this sample program is on GitHub [5].

Better than the Raspberry Pi?

The sample project has shown at least some of the PyPortal's capabilities. The question arises as to how the intelligent Adafruit display compares with a Pi Zero plus a miniature display.

The biggest benefit the PyPortal offers is its compact design. The processors, components, and connections on the rear are unlikely to wear out. Conversely, the Raspberry Pi is attached in a fairly precarious way to the socket array or HDMI port on a typical miniature screen. In the former case, the screen usually blocks pins that it doesn't need, so connecting additional sensors then means purchasing a multiplexer board.

Unlike the Raspberry Pi, the PyPortal comes with a tiny speaker for sound output. It is more suitable for warnings and signal tones than for listening to music, but the Pi Zero lacks this feature completely.

Regardless of the hardware, the intelligent display is far easier to set up than the Pi Zero and a small screen. Tinkering with the /boot/config.txt file with special overlays or HDMI parameters is not necessary. To switch it off, all you have to do is pull the plug, because there is no operating system that needs to be shut down cleanly first.

The PyPortal's current draw is about 200mA when the display is on and about 70mA otherwise. Therefore, the display is not suitable for battery operation. Even in continuous operation, though, it is unlikely to increase your electricity bill that much. A display of exactly the same size for the Raspberry Pi was not available for comparison measurements, but with a 4-inch screen from Waveshare, the combination clocked up to 260mA. The difference of 60mA between the PyPortal and Pi Zero including display can be measured, but in practice it hardly plays a role.

Of course, PyPortal does not always give you the better solution because many of its advantages turn into disadvantages in certain applications. If you need considerably more screen surface or better resolution, there is no escaping the Raspberry Pi, which also offers a faster CPU, more RAM, and multitasking. These more extensive resources also give you scope for more demanding and complex programs. When designing graphical user interfaces, especially, toolkits with layout management offer more possibilities for the developer than the PyPortal's displayio library.

Conclusions

As the lively community that has grown up around the PyPortal shows, Adafruit has its finger on the users' pulse with this display. Thanks to the fixed form factor, everyone programs the same hardware, and open source promotes the exchange of ideas and solutions. If you want to discover more details about the PyPortal or are looking for suggestions, you will find the guides [6] to be a comprehensive resource. In many projects, teaming the Raspberry Pi and the PyPortal also makes sense. The Pi handles the computationally intensive and memory-intensive work, and the PyPortal presents the results.