colorfulscoop / msgflow

msgFlow is a simple chatbot framework written in Python for simple configuration, customization and connection with several services

msgflow

msgFlow is a simple chatbot framework written in Python for simple configuration, customization and connection with several services.

Installation

Prepare Python >= 3.9. Then install msgFlow from PyPI.

$ pip install msgflow

msgFlow separas dependent libraries for each service. To make all the services available, install msgFlow with the following options.

$ pip install msgflow[webapi,slack,cron]

Quick Start

msgFlow requires a configuration file and an App script to run. To create those files, create your working directory first.

$ mkdir work
$ cd work/

msgFlow provides init command to generate the configuratio file and App script for you.

$ python -m msgflow init
$ ls
app.py  config.yml

Configuration file tells msgFlow which service is used to get inputs and return the response. and which App class is used. The App script defines the App class which defines how to generate a response from the input.

As a default, Config file uses stdin to get the input and show the output in the stdout.

Let us run the msgFlow with the Config setting. You can utilize the run command with --config_file option which specify the default config file.

$ python -m msgflow run --config_file config.yml
INFO:/work/msgflow/main.py:{"level": "info", "message": {"service": "CliService", "post_service": "CliService"}, "time": "2020-12-26 11:10:43.886375"}

The prompt you> waits for your input. Input your utterance there.

you> Hello
App got message: Message(text="Hello", dialog_id="you")
bot> Thank you for your message "Hello"!
you> World
App got message: Message(text="World", dialog_id="you")
bot> Thank you for your message "World"!

The default App class shows the message that what kind of message was reached, and then return the response with your input utterance.

To finish the conversation, input /exit.

you> /exit
Bye!

Service List

Check available services in docs.