ChatbotQuery

Framework to easily build functional simple chatbots with purpose of maintain basic conversations for querying databases.

Installation

It could be installed using shell

git clone https://github.com/tgquintela/chatbot_query
.\install

Main dependences:

• numpy
• scipy
• scikit-learn
• pandas
• jellyfish
• textblob
• stemming

Testing

You need to ensure that all the package requirements are installed. ChatbootQuery provide a testing module which could be called by importing the module and applying test function. It can be tested from the shell

>> ./test_script.sh

and also for developers you can test from the source using nosetests of nose package.

>> python3 -m "nose" --with-coverage --cover-package=chatbotQuery

Examples

There are different examples. That scripts can be runned using python

>> python example_script.py

The examples are:

• example_discourse.py: A simple bot who only speaks.
• example_basic_conversation.py: A simple bot who creates a profile of the user, if it don’t have that profile in its database.
• example_dbquery.py: Example of querying a database. There is a process of interaction as a substitution of a simple query. It is an accumulative query using the accepted queries done before.

Next steps

• Parse XML and AIML format file for chooser.
• Support for AIML sentence buildings.
• Refactorize DBAPI to Support other DBs and different queries communication.
• Build network of possible states in conversation state machines.