Getting Started with Rasa NLU¶

In this tutorial you will create your first Rasa NLU bot. You can run all of the code snippets in here directly, or you can install Rasa NLU and run the examples on your own machine.

As an example we’ll start a new project to help people search for restaurants. We’ll start with an extremely simple model of those conversations. You can build up from there.

Let’s assume that anything our users say can be categorized into one of the following intents:

greet
restaurant_search
thankyou

Of course there are many ways our users might greet our bot:

Hi!
Hey there!
Hello again :)

And even more ways to say that you want to look for restaurants:

Do you know any good pizza places?
I’m in the North of town and I want chinese food
I’m hungry

The first job of Rasa NLU is to assign any given sentence to one of the intent categories: greet, restaurant_search, or thankyou.

The second job is to label words like “chinese” and “North” as cuisine and location entities, respectively. In this tutorial we’ll build a model which does that.

1. Prepare your NLU Training Data¶

Training data is essential for developing chatbots and voice apps. The data is just a list of messages that you expect to receive, annotated with the intent and entities Rasa NLU should learn to extract.

The best way to get training data is from real users, and a good way to get it is to pretend to be the bot yourself. But to help get you started, we have some demo data here. See Training Data Format for details of the data format.

If you are running this in the docs, it may take a few seconds to start up. If you are running locally, copy the text between the triple quotes (""") and save it in a file called nlu.md.

nlu_md = """
## intent:greet
- hey
- hello
- hi
- good morning
- good evening
- hey there

## intent:restaurant_search
- i'm looking for a place to eat
- I want to grab lunch
- I am searching for a dinner spot
- i'm looking for a place in the [north](location) of town
- show me [chinese](cuisine) restaurants
- show me a [mexican](cuisine) place in the [centre](location)
- i am looking for an [indian](cuisine) spot
- search for restaurants
- anywhere in the [west](location)
- anywhere near [18328](location)
- I am looking for [asian fusion](cuisine) food
- I am looking a restaurant in [29432](location)

## intent:thankyou
- thanks!
- thank you
- thx
- thanks very much
"""
%store nlu_md > nlu.md

2. Define your Machine Learning Model¶

Rasa NLU has a number of different components, which together make a pipeline. Create a markdown file with the pipeline you want to use. In this case, we’re using the pre-defined tensorflow_embedding pipeline. If you are running this locally instead of here in the docs, copy the text between the (""") and save it in a file called nlu_config.yml.

nlu_config = """
language: en
pipeline: tensorflow_embedding
"""
%store nlu_config > nlu_config.yml

To choose which pipeline is best for you read Choosing a Rasa NLU Pipeline.

3. Train your Machine Learning NLU model.¶

To train a model, start the rasa_nlu.train command, and tell it where to find your configuration and your training data:

If you are running this in your computer, leave out the ! at the start.

!python -m rasa_nlu.train -c nlu_config.yml --data nlu.md -o models --fixed_model_name nlu --project current --verbose

We are also passing the --project current and --fixed_model_name nlu parameters, this means the model will be saved at ./models/current/nlu relative to your working directory.

4. Try it out!¶

There are two ways you can use your model, directly from python, or by starting a http server. Details of running the Rasa NLU HTTP server are in config.

To use your new model in python, create an Interpreter object and pass a message to its parse() method:

This will not work if you haven’t run the cells above!

from rasa_nlu.model import Interpreter
import json
interpreter = Interpreter.load("./models/current/nlu")
message = "let's see some italian restaurants"
result = interpreter.parse(message)
print(json.dumps(result, indent=2))

Spend some time playing around with this, for example try sending some different test messages to Rasa NLU. Also try adding some new words and phrases to your data at the top of the page, then re-train your model to teach Rasa NLU. Remember that this is just a toy example, with just a little bit of training data. To build a really great NLU system you’ll want to collect some real user messages!

How was the tutorial? Click to Vote

Note

For windows users the windows command line interface doesn’t like single quotes. Use doublequotes and escape where necessary. curl -X POST "localhost:5000/parse" -d "{/"q/":/"I am looking for Mexican food/"}" | python -m json.tool

Have questions or feedback?¶

We have a very active support community on Rasa Community Forum that is happy to help you with your questions. If you have any feedback for us or a specific suggestion for improving the docs, feel free to share it by creating an issue on Rasa NLU GitHub repository.