phas

README

Personal Home Automation Service

Installation

You can use PHAS on Raspberry Pi, but also on your personal computer.

Below you'll find installation instructions for all supported platforms:

• Linux - Raspberry Pi and Desktop
• Windows
• macOS

Linux

These instructions apply both to Raspberry Pi and Desktop distributions of Linux (tested under Debian and Ubuntu).

First, make sure the system is up to date by using the following commands:

sudo apt-get update
sudo apt-get upgrade

Then, install Go 1.13.3 or newer installed from the binary distribution. While Go 1.11 is still supported, it's better to use a newer version of Go and benefit from all the improvements in Go Modules, compilation speed and debugging experience.

For Raspberry Pi, use the below Go distribution. If you are on a Linux Desktop, then skip to the Linux Desktop instructions below this section.

cd /tmp
wget https://dl.google.com/go/go1.13.3.linux-armv6l.tar.gz
tar -zxvf go1.13.3.linux-armv6l.tar.gz
sudo mv go /usr/local/
rm go1.13.3.linux-armv6l.tar.gz

For Linux Desktop, you can skip this if you are on a Raspberry Pi.

cd /tmp
wget https://dl.google.com/go/go1.13.3.linux-amd64.tar.gz
tar -zxvf go1.13.3.linux-amd64.tar.gz
sudo mv go /usr/local/
rm go1.13.3.linux-amd64.tar.gz

Manually add Go's binary files, and your $GOPATH/bin binaries to your $PATH. E.g., place export PATH="${PATH}:/usr/local/go/bin:/home/pi/go/bin" at the end of ~/.bash_profile such as:

# Include the Go SDK and GOPATH/bin in PATH
export PATH="${PATH}:/usr/local/go/bin:/home/pi/go/bin"

If the file does not exist, then create it.

Now, let's install PHAS:

# clone this repository
git clone https://github.com/dlsniper/phas.git /home/pi/phas
cd /home/pi/phas

# Copy the required system libraries in the correct paths.
# You can skip these if you will always use the build commands with the correct CGO environment variables configured, see below.
sudo cp ./lib/picovoice.h /usr/local/include/
sudo cp ./lib/pv_porcupine.h /usr/local/include/

# If you are on Linux Desktop, then replace raspberrypi_arm11 with linux
sudo cp ./lib/raspberrypi_arm11/libpv_porcupine.* /usr/local/lib/

# Enable the vendoring mode for Go Modules so that no additional downloads
# are needed to compile the application
export GOFLAGS=-mod=vendor

# This is the target on which PHAS runs. The following values are useful:
# windows     - for Windows OS
# macos       - for macOS
# raspberrypi - for Raspberry PI 3 Linux OS
# linux       - for Linux OS
export PHAS_OS=raspberrypi

# Build the binary
go build -o phas

Linux Destkop

You need to install the following dependencies:

apt-get install pkg-config portaudio19-dev

Windows

For a Windows user, the compilation will be a bit more tricky.

First, you need to have msys2 installed.

Add C:\msys64\mingw64\bin to your Windows PATH environment variable

Then, open the MSYS2 MinGW 64-bit command prompt and run the following commands:

# Ensure msys2 is updated. You might need to run this command a couple of times
pacman -Syu

# Install build tools dependencies
pacman -S mingw-w64-x86_64-gcc mingw-w64-x86_64-pkg-config

# Install the dependency for recording audio from the microphone
pacman -S mingw-w64-x86_64-portaudio

Building on Windows

Building PHAS on Windows needs to be performed like this:

• set the following environment variables:

export CGO_CFLAGS=-I/Users/florin/phas/lib
export CGO_LDFLAGS=-L/Users/florin/phas/lib/macos

Also, include C:\msys64\mingw64\bin in your PATH on Windows.

For PHAS to work, you need to configure the following environment variable PHAS_OS=raspberrypi.

• then run go build -o phas

macOS

For macOS, you need to run the following commands:

brew install portaudio pkg-config

During the compilation step, you also need to set the following environment variable:

export CGO_CFLAGS=-I/Users/florin/phas/lib
export CGO_LDFLAGS=-L/Users/florin/phas/lib/macos
export DYLD_LIBRARY_PATH="${DYLD_LIBRARY_PATH}:/Users/florin/phas/lib/macos/"

Getting GCP credentials for the voice APIs

Before we can run the application, we need to get valid GCP credentials to make use of the voice APIs.

If you don't yet have a Google Cloud Platform account yet, then first head over to this link and create one.

Then, read the link here about obtaining and providing service account credentials manually. Select Service User | API Keys Viewer from the Role and the json format.

After this, go to APIs & Services and select +Enable APIs and Services button. Here, you'll need to enable the following services:

• Cloud Speech-to-Text API, Cloud Speech-to-Text API Length Standard option should be enough
• Cloud Text-to-Speech API, WaveNet option should be enough

Finally, place the downloaded file in a location such as /home/pi/phas-gcp-key.json. Then, add the following environment variables to your .bash_profile:

# These are needed to tell PHAS what GCP project to use and what credentials
export GCP_PROJECT_ID=<your GCP project ID>
export GOOGLE_APPLICATION_CREDENTIALS="/home/pi/phas-gcp-key.json"

If you are on Windows, then place the file under C:\Users\<username>\phas-gcp-key.json and then set the environment variables accordingly:

GCP_PROJECT_ID=<your GCP project ID>
GOOGLE_APPLICATION_CREDENTIALS="/home/pi/phas-gcp-key.json"

Running the application

Since we installed all dependencies and everything is up to date, let's run the application. Execute .\phas.exe if you are Windows, and ./phas if you are on all other supported platforms.

License

This repository and all code from it is licensed under the Apache 2 license.

The code under lib directory belongs to the Picovoice Porcupine library, also available under Apache 2 license.

The code under the vendor directory belongs to their original creators and is available under their respective license. It is included here to enable the smooth operation of the workshop.