Optimized for python 3.6 - Language changes subjected
This project aims in creating a chatbot for 'Museu da Pessoa'. The idea is use Telegram/WhatsApp/Facebook APIs to get all incoming messages, process it with NLP and intent recognition techniques (in market - IBM Watson or in house made API) and replies it as a person from the Museum's collection.
- Bruno Paes
- Fernando Sinigaglia
- Guilherme Heitzmann
- Leonado Briotto
- Leonardo Messias
- André Insardi
- Carlos Rafael
- Flávio Marques
For developers, python requirements could be find in the project's root. For installing the requirements, in your venv or anaconda env, just run the following command:
pip install -r requirements.txt
.
└── labs-chickenbot
├── data
│ └── credentials.json
├── docs
│ ├── reference_articles
│ ├── ...
│ └── credits.txt
├── src
│ ├── __init__.py
│ └── bot.py
├── tests
│ └── unittests
│ ├── data
│ └── __init__.py
├── .gitignore
├── LICENSE
├── README.md
└── requirements.txt
- data: The data dir. Group of non-script support files.
- docs: The documentation dir.
- src: The scripts & source code dir.
- tests: The unittests dir.
For running it, on the ~/src
directory just run the follow command:
python bot.py
- Instead of python, Node JS may be used.
- Chickenbot is just a provisional name
This project is under development.
It needed to:
-
Create the bot end - responsible for getting and replying messages.
- Getting messages - almost done
- Replying messages - almost done
-
Create the NLP end:
- Use in house api?
- Techniques:
- Word Embeddings
- Word2Vec
- Word Histogram - word bag
- Techniques:
- Use in market api?
- IBM Watson API
- Use in house api?