bmpotter / self-amd64

watson-intu Self build on amd64 / x86

self-amd64

Watson-Intu Self Docker container build for x86 (amd64) Requires sound card, mic, speaker (or headphones), and camera (webcam preferred)

+ +

Build

Dev version (leaves build artifacts in place, retains utilities like vim) # 5.9GB image

docker build --squash --build-arg "mode=dev" -f Dockerfile.x86 -t openhorizon/amd64-x86-intu-self:edge .

(omit "--squash" if you're not using docker experimental features)

Prod version (cleans up build artifacts, removes vim and unncessary deps, cleans apt-cache) # 1.9GB image

docker build --squash --build-arg "mode=prod" -f Dockerfile.x86 -t openhorizon/amd64-x86-intu-self .

Pull Docker image from openhorizon public repo

docker pull openhorizon/amd64-x86-intu-self:edge     # approx 1.9GB
docker pull openhorizon/amd64-x86-intu-self-dev:edge     # approx 5.9GB

Run

0. Clone this repo, and cd to the ./config directory. Use bootstrap-example.json as a starting point

1. Set up an IBM account with credentials for Watson services, according to this guide

2. Copy your Watson services credential pwds/URLs into bootstrap-example.json, save as bootstrap.json.

3. On your x86 system, run aplay -l and arecord -l, identify the integer number of your sound card. Edit the alsa.conf file with these values. (see below for an example, if your sound card number is '2'.) Save the file.

   defaults.ctl.card 2 # approx line 79 defaults.pcm.card 2

4. Run Self in a container, linking your current config files:

   docker run -it --rm --privileged -p 9443:9443 -v $PWD:/config openhorizon/amd64-x86-intu-self:edge /bin/bash -c "ln -s -f /config/bootstrap.json bin/linux/etc/shared/; ln -s -f /config/alsa.conf /usr/share/alsa/alsa.conf; ./bin/linux/run_self.sh"

5. When Self starts, it will verbally greet you. You can browse to http://localhost:9443/www/dashboard#/ to see the Self dashboard. (substitute your TX2's IP address in place of "localhost" to visit the UI from another computer on LAN)