used starter Pack for Rasa Stack

For more information on the Rasa Stack, please visit the docs here:

• Rasa Core
• Rasa NLU

Setup

pip install rasa_core

RASA-NLU with spyCy ***************************************

pip install rasa_nlu[spacy]
python -m spacy download en_core_web_md
python -m spacy link en_core_web_md en

== for tensorflow *****************************************

pip install rasa_nlu[tensorflow]

To install the necessary requirements, run:

pip install -r requirements.txt

install all dependencies ever needed

pip install -r alt_requirements/requirements_full.txt

---

Usage

Run duckling docker container if you use ner_duckling_http in pipeline (duckling used to recognize mostly etc. dates https://duckling.wit.ai/) http://rasa.com/docs/nlu/master/components/#id2

docker run -p 8000:8000 rasa/duckling

Before train nlu data take a look at nlu_data.md = need to type in proper data for training. Specially Synonyms part, as for now it is made as example. remember Synonyms maps synonym to entity, but you need to make a proper training data with synonyms... Rememeber! when entity name ans slot name are equal slot set automaticaly when intent received. To train the NLU model, run make train-nlu

To train the Core model, run make train-core

To run the bot on the command line run make cmdline To run the core-server on the command line run make core-server

test nlp url:

http://localhost:5005/conversations/default/parse?query=hello

Interactive learning

make core-learn then proceed conversation like here https://rasa.com/docs/core/interactive_learning/ to execute ations during training make sure actions.py RUN methods do what needs to be done.. e.g if bot needs to execute action_check_room_available , the CheckRoomAvailable.run must set slot is_room_available TRUE/FALSE, deppending if room available. Used fake data from learning_data.json

the second possible way to execute actions is here https://rasa.com/docs/core/customactions/ Custom Actions.