/rasa-demo-pydata18

Demo application for the PyData Conference in Amsterdam 2018

Primary LanguageJupyter NotebookMIT LicenseMIT

Conversational AI with Rasa NLU and Core

Demo application for the PyData Conference in Amsterdam 2018

You can find the slides for the presentation at Slideshare.

Installation

The notebooks contain the installation instructions as their first step. Basically, you'll need a virtualenv with rasa_core, rasa_nlu[spacy] and spacy installed.

Also make sure to install a spacy language model for en.

Notebooks

  • rasa-moodbot-demo-starter this is what we are going to start with
  • rasa-moodbot-demo-complete this is how our notebook should look like once we are done
  • rasa-moodbot-demo-complete-executed if hell breaks loose, this is a backup notebook with all the output cells filled

The additional files (nlu.md, stories.md, domain.yml, config.yml, story_eval.pdf, story_graph.png) will be explained and created in the notebooks.

Once you ran the notebooks, there should be a models/ folder containing your trained Rasa model.

Additional Material

Running duckling locally

  • run the duckling docker container

     docker run -p 8000:8000 rasa/duckling
  • add the duckling component to your pipeline:

     - name: "ner_duckling_http"
       url: localhost:8000

Slides

TBC - I'll add the slides once the talk is done