1. Basics

The following tutorial shows basic usage of pipeline module as an extension to dff.script.core.

Here, __call__ (same as run) method is used to execute pipeline once.

# installing dependencies
%pip install -q dff

Note: you may need to restart the kernel to use updated packages.

from dff.script import Context, Message

from dff.pipeline import Pipeline

from dff.utils.testing import (
    check_happy_path,
    is_interactive_mode,
    HAPPY_PATH,
    TOY_SCRIPT,
    TOY_SCRIPT_ARGS,
)

Pipeline is an object, that automates script execution and context management. from_script method can be used to create a pipeline of the most basic structure: “preprocessors -> actor -> postprocessors” as well as to define context_storage and messenger_interface. Actor is a component of :py:class:.Pipeline, that contains the :py:class:.Script and handles it. It is responsible for processing user input and determining the appropriate response based on the current state of the conversation and the script. These parameters usage will be shown in tutorials 2, 3 and 6.

Here only required parameters are provided to pipeline. context_storage will default to simple Python dict and messenger_interface will never be used. pre- and postprocessors lists are empty. Pipeline object can be called with user input as first argument and dialog id (any immutable object). This call will return Context, its last_response property will be actors response.

pipeline = Pipeline.from_script(
    TOY_SCRIPT,
    # Pipeline script object, defined in `dff.utils.testing.toy_script`
    start_label=("greeting_flow", "start_node"),
    fallback_label=("greeting_flow", "fallback_node"),
)

For the sake of brevity, other tutorials might use TOY_SCRIPT_ARGS to initialize pipeline:

assert TOY_SCRIPT_ARGS == (
    TOY_SCRIPT,
    ("greeting_flow", "start_node"),
    ("greeting_flow", "fallback_node"),
)

if __name__ == "__main__":
    check_happy_path(pipeline, HAPPY_PATH)
    # a function for automatic tutorial running (testing) with HAPPY_PATH

    # This runs tutorial in interactive mode if not in IPython env
    # and if `DISABLE_INTERACTIVE_MODE` is not set
    if is_interactive_mode():
        ctx_id = 0  # 0 will be current dialog (context) identification.
        while True:
            message = Message(input("Send request: "))
            ctx: Context = pipeline(message, ctx_id)
            print(ctx.last_response)

(user) >>> text='Hi'
 (bot) <<< text='Hi, how are you?'
(user) >>> text='i'm fine, how are you?'
 (bot) <<< text='Good. What do you want to talk about?'
(user) >>> text='Let's talk about music.'
 (bot) <<< text='Sorry, I can not talk about music now.'
(user) >>> text='Ok, goodbye.'
 (bot) <<< text='bye'
(user) >>> text='Hi'
 (bot) <<< text='Hi, how are you?'