The goal of this project is to create an interactive, text-based chatbot which is capable of completing 30 turns of conversation in a realistic manner. This must be accomplished using proper communication, structure and planning in order to accurately represent the industry standards.
Our project is simulating a blind first date scenario with the mysterious ‘Mackenzie’. It will cover topics such as hobbies, interests and goals. The scenario has multiple endings, both positive and negative, based on user responses. The majority of the responses will contain humour as the primary personality trait.
The project is being implemented using python and the library wit.ai which was chosen because it handles all of the natural language processing for us which turns the text into intents. Python was chosen due to it working well with wit.ai and we were all familiar with it.
This was created and test on python3.7
If python doesn’t work replace it with py
- Install Pycharm or any alternative python compiler
- Open command prompt and ensure you have the latest pip and wheel:
- Check if you have wheel installed
python -m pip help wheel - If you don’t have wheel installed run this command:
python -m pip install wheel --user - If you do have wheel installed run this command:
python -m pip install --upgrade pip wheel setuptools
- Install the dependencies
python -m pip install docutils pygments pypiwin32 kivy.deps.sdl2 kivy.deps.glew python -m pip install kivy.deps.gstreamer python -m pip install googletrans - Install kivy and Wit.ai:
python -m pip install kivy python -m pip install wit - Run MackGui.py on python compiler
python MackGui.py
- Install Pycharm or any alternative python compiler
- Open terminal and ensure you have the latest pip and wheel:
- Check if you have wheel installed
python3 -m pip help wheel - If you don’t have wheel installed run this command:
python3 -m pip install wheel --user - If you do have wheel installed run this command:
python3 -m pip install --upgrade pip wheel setuptools
- Install the dependencies
python3 -m pip install docutils pygments pypiwin32 kivy.deps.sdl2 kivy.deps.glew python3 -m pip install kivy.deps.gstreamer python3 -m pip install googletrans - Install kivy and Wit.ai:
python3 -m pip install kivy python3 -m pip install wit - Run MackGui.py on python compiler
python3 MackGui.py
All classes are camelCase, contain only lowercase letters with capital letters signifying the start of a new word. Our classes are organized using the following folder structure. note gui will be included in Mack
- Source
- Data structure
- Wit
- Mack
- Personality
- Docs
- Gantt Chart
- WBS
- Project Proposal
- Project Report
- Toggl Summary
Additions:
- Added 6 new topics (1 per team member - 2 points)
- Colour Yellow
- Gin
- Rum
- Amaretto
- Tequila
- Jager
Mack: HI! My name is Mack and I am a chatbot!
User: do you like yellow
Mack: I love yellow! It's such a happy colour :))
User: what do you know about the colour yellow
Mack: yellow is a primary colour!
User: do you like gin?
Mack: I haven't tried gin before, but you can mix it with many other things.
User: what do you know about gin?
Mack: Gin derives its predominant flavour from juniper berries!
User: do you like tequila?
Mack: I haven't tried tequila before, however I have found many people like Corona.
User: what do you know about tequila?
Mack: Tequila is a distilled beverage.
User: Do you like amaretto
Mack: I haven't tried amaretto before, however I think I would like it.
User: what do you know about amaretto?
Mack: Amaretto is a sweet italian liqueur that originated in Sarononna.
User: do you like rum?
Mack: I haven't tried rum before, however I've heard pina coladas are good.
User: what do you know about rum?
Mack: Rum is made from sugarcane byproducts, such as molasses!
User: do you like jager?
Mack: I haven't tried jager before, but the logo is cool.
User: what do you know about jager?
Mack: Jager is a digestiff made with 56 herbs
- Added 5 Different Off Topic Responses (3 points)
- I'm sorry I don't understand, wanna talk about sports?
- I do not understand you at all :|
- I am very confused right now. Could we talk about something else?
- I don't understand, wanna talk about lemonsqures?
- I'm sorry I don't know about that.
User: is it okay for me to eat yellow snow
Mack: I'm sorry I don't understand, wanna talk about sports?
User: is it okay for me to eat yellow snow
Mack: I do not understand you at all :|
User: is it okay for me to eat yellow snow
Mack: I am very confused right now. Could we talk about something else?
- Mack can now communicate with more than 1 language (both input and output)
User: hallo, hoe gaat het met jou?
Mack: Het gaat goed, dank je.
User: dat is heel leuk!
Mack: Het spijt me, ik heb die emotie niet.
User: wat vind je van liemoen koekjes?
Mack: Ik geniet echt van binary, hoe zit het met jou?
- Mack can now communicate with itself through our socket feature (15 points)
Mack: Soccer
Mack 2: What about soccer?
Mack: France won the last world cup.
Mack 2: I'm sorry, I don't know about that.
Mack: I'm not sure, sorry.
Mack 2: Interesting, go on.
Mack: Interesting, go on.
Mack 2: Interesting, go on.
Mack: Interesting, go on.
Unchanged Features from A2:
- GUI was already implemented during A2 (5 points)
- wit.ai was used to create Mack.ai which is a language toolkit (5 points)
- wit.ai handles spelling mistakes (5 points)
Description of how these features improved Mack.ai:
1. By adding more language support Mack is now usable by a wider range of people thus
increasing its usability.
2. By increasing the number of potential topics Mack is able to talk about a wider range
of topics, making it a more interesting conversation.
3. By implementing the socket feature, Mack can now talk to itself and have a normal
flow conversation.
In order to run the Mack to Mack feature you must run the server.py file then run the
mack.py file while the server.py file is still running.
January 31st, 2019 11:59 PM
February 15th, 2019 11:59 PM
April 5th, 2019 11:59 PM