-
Notifications
You must be signed in to change notification settings - Fork 3
Getting started
This will help you get up and running with Jenova.
The following packages need to be installed via pip
-
gunicorn
: lightweight WSGI HTTP Server -
falcon
: super fast RESTful framework -
httplib2
: HTTP client
For the bot, these packages are required:
-
pygame
: for audio playback -
pyttsx
,gTTS
: TTS engines
For the server, these packages are required:
-
sklearn
,nltk
: for machine learning algorithms and utilities -
numpy
,scipy
: libraries used bysklearn
-
tinysegmenter
: Japanese tokenizer
Python3 and ev3dev-lang-python are also required to run the application. It's ok to run the application without ev3dev
, but you won't have the features related to Lego EV3 like motor controlling.
- Install the dependencies
It depends on the operating system you are using, here are the example for Ubuntu 17.04
sudo apt-get install python3 python3-pip python3-ev3dev
sudo pip3 install gunicorn falcon httplib2 pygame pyttsx gTTS numpy scipy sklearn nltk tinysegmenter
If you want to use mary-tts
as TTS engine, it must be installed separately:
https://github.com/marytts/marytts/wiki/Local-MaryTTS-Server-Installation
(Note that there are some problem with installing and running pyttsx
with Python3. Checkout this fork of pyttsx instead)
- Checkout the source code and install the framework
git clone --recursive https://github.com/dungba88/jenova.git
cd jenova
sudo pip3 install orion/
- Run the bot
gunicorn main --chdir bot/ -b 0.0.0.0:8081 --reload
gunicorn main --chdir server/ -b 0.0.0.0:8080 --reload
(You may need to run with nohup
command)
Now the bot can be accessed from http://localhost:8081 and the server can be accessed from http://localhost:8080
- Start the UI
For this, you need to install a web server which supports static files, like Apache
or nginx
, and make the ui/ folder accessible by HTTP. Setup will depend on which web server you choose.
Now you are ready to interact with bot!