35. Final: Integration#
At this point the components are nearly done!
Time to connect them 😈
35.1. Debugging Tips#
Here are some known “gotchas.” You should really read them all.
If you have one to add, please edit and submit a pull request!
Docker#
Can’t connect to DFEC AI server#
The DFEC AI server’s public URL is only IPv6! By default Docker only gives your container an IPv4 address.
Edit
/etc/docker/daemon.json
(with Vim or Nano)After the first curly brace, make a new line and add:
"ipv6": true,
(with the comma because it must be valid JSON)Save and exit
Restart the docker daemon:
sudo systemctl restart docker
35.2. Grading#
This table shows how many points you get for each segment of the integration. Each segment is equally weighted.
For the ______ segment |
If you… |
You get this percentage for that segment |
---|---|---|
Trigger voice recording |
Trigger with button |
85% total |
Trigger voice recording |
Trigger with Arduino KWS |
100% total |
Voice transcription |
Cold start voice-to-text (model not loaded until trigger) |
20% total |
Voice transcription |
Warm start voice-to-text (model loaded before trigger) |
100% total |
Networking |
Send transcription to LLM; LLM extracts and returns the location |
+50% |
Networking |
GET weather for location from https://wttr.in and print to screen |
+50% |