1. Create virtualenv Please look this site - https://virtualenv.pypa.io/en/stable/userguide/

2. if you are using Ubuntu pip install -r requirement.txt

3. If you are using windows machine pip install -r requirement.txt The above command will fail for few packages like mitie. So you have to install the following packages manually pip install numpy,scipy,sklearn,spacy

4. You will have below parts in this projects

   1. Brain

   This part takes care of Natural language understanding. It will recall all information which are stored on brain and response back to human. As of now, it can able to get person's information, their siblings and electronic device handling like switch ON/OFF

   2. Web(Human Interface)

   This is a flask application which use to interact human with DUTCH

5. As pre-requists, you need to train this dutch to do that you need two json files config and training_data. config - to train the model using either mitie or spacy training_data - contains data about what the dutch can able to do for you

6. To train your model you can use either spacy or mitie algorithm. For my application I am using mitie.

   1. If you are using mitie you should download a .dat file from https://github.com/mit-nlp/MITIE/releases/download/v0.4/MITIE-models-v0.2.tar.bz2

   2. configuration file format https://nlu.rasa.com/pipeline.html#mitie-sklearn

   python -m rasa_nlu.train -c config_mitie.json

   This step will take times based on your machine configuration

7. Once the training is done you will have a models folder will be created with a timestamp and given name

8. Now you are ready to configure this model with rasa-nlu

   python -m rasa_nlu.server -c config_mitie.json --server_model_dirs=./model_YYYYMMDD-HHMMSS

9. To setup the knowledge base for the DUTCH,

   1. Create sqlite db using these files creation_script.sql and insert.sql. Use this link for sqlite DB studio http://sqlitebrowser.org/
   2. Export the created sqlite db and keep that file inside brain folder
   3. Use config.py file to define all required configurations

10. train dutch data

    python -m rasa_nlu.train -d "/brain/data/dutch-data.json" -c "/brain/config_mitie.yml" --project "dutch" -o "/models/nlu"

11. run rasa_nlu server

    python -m rasa_nlu.server --path "/models/nlu" -c "/brain/config_mitie.yml"

    Test the rasa_nlu server http://localhost:5000/parse?q=who are you&project=dutch&model=model_20180706-222616

12. Done! :). Now you are ready to get conversation with your dutch

    python boot.py

Cheers!!

• The electronic device handling part is in-progress. It will connect with a raspi hardware and will do based on sending command to the hardware like switch ON/OFF computer, lights, TV, etc.,

Keep watch...