About Jarvis - An Intelligent AI Consciousness 🧠

Jarvis is a voice commanding assistant service in Python 3.5+ It can recognize human speech, talk to user and execute basic commands.

Assistant Skills

• Opens a web page (e.g 'Jarvis open youtube')
• Play music in Youtube (e.g 'Jarvis play mozart')
• Increase/decrease the speakers master volume (also can set max/mute speakers volume) ** (e.g 'Jarvis volume up!')
• Opens libreoffice suite applications (calc, writer, impress) (e.g 'Jarvis open calc')
• Tells about something, by searching on the internet (e.g 'Jarvis tells me about oranges')
• Tells the weather for a place (e.g 'Jarvis tell_the_skills me the weather in London')
• Tells the current time and/or date (e.g 'Jarvis tell me time or date')
• Set an alarm (e.g 'Jarvis create a new alarm')
• Tells the internet speed (ping, uplink and downling) (e.g 'Jarvis tell_the_skills me the internet speed')
• Tells the internet availability (e.g 'Jarvis is the internet connection ok?')
• Tells the daily news (e.g 'Jarvis tell me today news')
• Spells a word (e.g 'Jarvis spell me the word animal')
• Creates a reminder (e.g 'Jarvis create a 10 minutes reminder')
• Opens linux applications (e.g 'Jarvis open bash/firefox')
• Tells everything it can do (e.g 'Jarvis tell me your skills or tell me what can you do')
• Tells the current location (e.g 'Jarvis tell me your current location')
• Tells how much memory consumes (e.g 'Jarvis tell me your memory consumption)
• Tells users commands history (e.g 'Jarvis tell me my history')
• Write/tell 'remember' and enable learning mode and add new responses on demand! (e.g 'Jarvis remember')
• Clear bash console (e.g 'Jarvis clear console')
• Has help command, which prints all the skills with their descriptions (e.g 'Jarvis help')
• Do basic calculations (e.g 'Jarvis (5 + 6) * 8' or 'Jarvis one plus one')
• Change settings on runtime (e.g 'Jarvis change settings')

Assistant Features

• Asynchronous command execution & speech recognition and interpretation
• Supports two different user input modes (text or speech), user can write or speek in the mic.
• Answers in general questions (via call Wolfram API), e.g ('Jarvis tell me the highest building')
• Change input mode on run time, triggered by a phrase e.g 'Jarvis change settings')
• Easy voice-command customization
• Configurable assistant name (e.g 'Jarvis', 'Sofia', 'John' etc.) (change on run time supported)
• Log preview in console
• Vocal or/and text response
• Keeps commands history and learned skills in MongoDB.'

Getting Started

Create KEYs for third party APIs

Jarvis assistant uses third party APIs for speech recognition,web information search, weather forecasting etc. All the following APIs have free no-commercial API calls. Subscribe to the following APIs in order to take FREE access KEYs.

• OpenWeatherMap: API for weather forecast.
• WolframAlpha: API for answer questions.
• IPSTACK: API for current location.

Setup Jarvis in Ubuntu/Debian system

• Download the Jarvis repo localy:

git clone https://github.com/ggeop/Jarvis.git

• Setup Jarvis and system dependencies:

bash setup.sh

• Put the Keys in settings

NOTE: For better exprerience, before you start the application you can put the free KEYs in the settings.py

nano Jarvis/src/jarvis/jarvis/setting.py

Start voice commanding assistant

• Start the assistant service:

bash run_jarvis.sh

How to add a new Skill to assistant

You can easily add a new skill in two steps.

• Create a new configurationin SKILLS in skills/registry.py

{ 
  'enable': True,
  'func': Skills.new_skill,
  'tags': 'tag1, tag2',
  'description': 'skill description..'
}               

• Create a new skill module in skills/collection

Desicion Model

Extract skill

The skill extraction implement in a matrix of TF-IDF features for each skill. In the following example he have a dimensional space with three skills. The user input analyzed in this space and by using a similarity metric (e.g cosine) we find the most similar skill.

Contributing

• PRs are welcome ☺️
• Try to follow PEP 8 guidelines and add comments!