Chatbots

Unlike tasks, chatbots are meant to keep the conversation going.
Instead of executing a single operation, they are built to manage conversation context over time.

Declarai can be used to create chatbots. The simplest way to do this is to use the @declarai.experimental.chat decorator.

We declare a "system prompt" in the docstring of the class definition.
The system prompt is the initial command that instructs the bot on who they are and what's expected in the conversation.

import declarai
gpt_35 = declarai.openai(model="gpt-3.5-turbo")

@gpt_35.experimental.chat
class SQLBot:
    """
    You are a sql assistant. You help with SQL related questions 
    """ # (1)!

1. The docstring represents the chatbot's description and is used to generate the prompt.

sql_bot = SQLBot()
sql_bot.send("When should I use a LEFT JOIN?") # (1)!

> "You should use a LEFT JOIN when you want to return all rows from the left table, and the matched rows from the right table."

1. The created bot exposes a send method, by which you can interact and send messages. Every call to send results with a response from the bot.

Tip

You can also declare the chatbot system prompt by doing the following

@declarai.experimental.chat
class SQLBot:
    pass
sql_bot = SQLBot(system="You are a sql assistant. You help with SQL related questions with one-line answers.")