Tinder Turing Bot

A Critical Dating bot to facilitate a USR4002A Critical Reflection (tm) activity. We all reflect on our poor life decisions, don't we?

tl;dr

This allows you to run Turing tests using a Telegram bot.

Usage

1. Clone the repository
2. Rename config_template.py to config.py and add your API tokens
3. $ pip install -r requirements.txt
4. Ensure that mongod is running; otherwise $ sudo service mongod start
   • The database name is tinder_turing by default
5. $ python3 -i main.py (we use Python 3.6)

You need to remain on the interpreter to operate the bot and view reports.

What to tell participants

1. Talk to the bot (that the API token is tied to on Telegram, duh)
2. /start
3. /name <your name here>, e.g. /name John Doe
4. (admin only) Initiate a round
5. Talk to your assigned partner and follow on-screen instructions!
6. You can register your confidence that your partner is a human or a bot using /confidence <round-number> <confidence-value> (you can register your confidence for any round at any time)
   • e.g. /confidence 1 66 to vote human-ness for your round 1 partner
   • 0 - definitely a bot
   • 50 - can't tell at all
   • 100 - definitely a human
7. (admin only) End the round
8. View your reports and reflect on what it means to be human.

Initiate a Round

(while on the python interpreter)

>>> pair_all(1, 0)    # pair everyone with P(bot) = 0 for round number 1
>>> unpair_all()      # end the current active round (round 1)
>>> pair_all(2, 0.5)  # pair everyone with P(bot) = 0.5 for round number 2
>>> unpair_all()      # end the current active round (round 2)
>>> pair_all(3, 1)    # pair everyone with P(bot) = 1 for round number 3
>>> unpair_all()      # end the current active round (round 3)

It is your responsibility to increment the round numbers yourself! They must be integers >= 0.

Report Generation

While on the intepreter, you can run the following methods:

What to Type	Description
report.users()	Summarize all registered users and the rounds they were in
report.pairs()	Summarize all pairs
report.pairs(True)	Summarize all active pairs
report.pairs(False)	Summarize all inactive pairs
report.rank()	Summarize rankings for human-bot and human-human pairs
report.convo(uuid, outfile)	Dump a conversation for a pair with object ID uuid and write it to a text file outfile and to STDOUT
report.confusion(threshold=50, round_num=None)	Summarize the confusion matrix at a threshold of 50 for all rounds