DataCamp Building Chatbots in Python
BUILDING CHATBOTS IN PYTHON
Virtual assistants and
accessing data
Alan Nichol
Co-founder and CTO, Rasa
DataCamp Building Chatbots in Python
Virtual assistants
Common chatbot use cases:
Scheduling a meeting
Booking a flight
Searching for a restaurant
Require information about the outside world
Need to interact with databases or APIs
DataCamp Building Chatbots in Python
Basic SQL
name pricerange area rating
Bill's Burgers hi east 3
Moe's Plaice low north 3
Sushi Corner mid center 3
SELECT * from restaurants;
SELECT name, rating from restaurants;
SELECT name from restaurants WHERE area = 'center' AND pricerange = 'hi';
DataCamp Building Chatbots in Python
SQLite with Python
In [1]: import sqlite3
In [2]: conn = sqlite3.connect('hotels.db')
In [3]: c = conn.cursor()
In [4]: c.execute("SELECT * FROM hotels WHERE area='south' and pricerange='hi'")
Out[4]: <sqlite3.Cursor at 0x10cd5a960>
In [5]: c.fetchall()
Out[5]: [('Grand Hotel', 'hi', 'south', 5)]
DataCamp Building Chatbots in Python
SQL injection
# Bad Idea
query = "SELECT name from restaurant where area='{}'".format(area)
c.execute(query)
# Better
t = (area,price)
c.execute('SELECT * FROM hotels WHERE area=? and price=?', t)
DataCamp Building Chatbots in Python
BUILDING CHATBOTS IN PYTHON
Let's practice!
DataCamp Building Chatbots in Python
BUILDING CHATBOTS IN PYTHON
Exploring a DB with natural
language
Alan Nichol
Co-founder and CTO, Rasa
DataCamp Building Chatbots in Python
Example messages
"Show me a great hotel"
"I'm looking for a cheap hotel in the south of town"
"Anywhere so long as it's central"
DataCamp Building Chatbots in Python
Parameters from text
In [1]: message = "a cheap hotel in the north"
In [2]: data = interpreter.parse(message)
In [3]: data
Out[3]:
{'entities': [{'end': '7', 'entity': 'price', 'start': 2, 'value': 'lo'},
{'end': 26, 'entity': 'location', 'start': 21, 'value': 'north'}],
'intent': {'confidence': 0.9, 'name': 'hotel_search'}}
In [4]: params = {}
In [5]: for ent in data["entities"]:
...: params[ent["entity"]] = ent["value"]
In [6]: params
Out[6]: {'location': 'north', 'price': 'lo'}
DataCamp Building Chatbots in Python
Creating a query from parameters
In [7]: query = "select name FROM hotels"
In [8]: filters = ["{}=?".format(k) for k in params.keys()]
In [9]: filters
Out[9]: ['price=?', 'location=?']
In [10]: conditions = " and ".join(filters)
In [11]: conditions
Out[11]: 'price=? and location=?'
In [12]: final_q = " WHERE ".join([query, conditions])
In [13]: final_q
Out[13]: 'SELECT name FROM hotels WHERE price=? and location=?'
DataCamp Building Chatbots in Python
Responses
In [1]: responses = [
"I'm sorry :( I couldn't find anything like that",
"what about {}?",
"{} is one option, but I know others too :)"
]
In [2]: results = c.fetchall()
In [3]: len(results)
Out[3]: 4
In [4]: index = min(len(results), len(responses)-1)
In [5]: responses[index]
Out[5]: '{} is one option, but I know others too :)'
DataCamp Building Chatbots in Python
BUILDING CHATBOTS IN PYTHON
Let's practice!
DataCamp Building Chatbots in Python
BUILDING CHATBOTS IN PYTHON
Incremental slot filling and
negation
Alan Nichol
Co-founder and CTO, Rasa
DataCamp Building Chatbots in Python
Incremental filters
DataCamp Building Chatbots in Python
Basic Memory
In [1]: def respond(message, params):
...: # update params with entities in message
...: # run query
...: # pick response
...: return response, params
# initialise params
In [2]: params = {}
# message comes in
In [3]: response, params = respond(message, params)
DataCamp Building Chatbots in Python
Negation
"where should I go for dinner?" "what about Sally's Sushi Place?"
"no I don't like sushi" "ok, what about Joe's Steakhouse?"
DataCamp Building Chatbots in Python
Negated entities
assume that "not" or "n't" just before an entity means user wants to exclude this
normal entities in green, negated entities in purple
DataCamp Building Chatbots in Python
Catching negations
In [1]: doc = nlp('not sushi, maybe pizza?')
In [2]: indices = [1, 4]
In [3]: ents, negated_ents = [], []
In [4]: start = 0
...: for i in indices:
...: phrase = "{}".format(doc[start:i])
...: if "not" in phrase or "n't" in phrase:
...: negated_ents.append(doc[i])
...: else:
...: ents.append(doc[i])
...: start = i
DataCamp Building Chatbots in Python
BUILDING CHATBOTS IN PYTHON
Let's practice!
