I wanted to run SlackBot inside Pepper.

Until now, Slack and Pepper have been linked using Slack's Incoming Web Hook. I was posting using the request module from the work app, I wanted to be able to receive notifications from the Slack side, so I thought about installing SlackBot in Pepper.

I want to use it across all work apps.

What I wanted to do with SlackBot.

--Create a channel that collects Pepper bots. --Send various events to Pepper with a message addressed to the bot. --Send various events to all Pepper with a message addressed to @Channel. --Collect various sensor information of Pepper in Slack.

Also, I wanted to use it across the board regardless of the work application, so I implemented it as a service by referring to this article. http://qiita.com/yacchin1205/items/11301d79380d08d2dbf6

Development environment

Mac OS Sierra Python 2.7.10

Enable bots in Slack

キャプチャ.PNG

Environment

Install QiBuild

pip install qibuild

Install Python SDK

Please drop the following version instead of the latest version. The latest version (2.5.5 at the time of writing) seems to have many changes, and there was little information to be helpful, so I downloaded 2.4.3.

pynaoqi-python2.7-2.4.3.28-mac64

Leave it through path. Also, add it to .bash_profile.

$ export PYTHONPATH=${PYTHONPATH}:/path/to/pynaoqi-python2.7-2.4.3.28-mac64
$ export DYLD_LIBRARY_PATH=${DYLD_LIBRARY_PATH}:/path/to/pynaoqi-python2.7-2.4.3.28-mac64

The following commands can be used if the path is correct.

$ python

>> import qi
>> import naoqi

Work tree preparation

The final required environment is as follows.

SlackBotService
　qiproject.xml
　manifest.xml
　slackbotservice.py
　slackbotservice.pml
　bot_settings.py
　lib
　　backports
　　slackbot
　　slacker
　　websocket
　　six.py
　　plugins
　　　__init__.py
　　　mention.py
　　　proxy.py

Execute the following command in the same hierarchy as the project directory. Various builds and packaging will be possible in the lower hierarchy.

$ pwd
/path/to/SlackBotService
$ qibuild init
$ qipy bootstrap

Install slackbot module

$ pip install slackbot

The required modules will be installed together. Copy the following installed modules under SlackBotService / lib.

$ pwd
/path/to/SlackBotService/lib
$ ls
backports/
slackbot/
slacker/
websocket/
six.py

Copy websocket / cacert.pem under SlackBotService / lib.

$ pwd
/path/to/SlackBotService/lib
$ ls
backports/
slackbot/
slacker/
websocket/
six.py
cacert.pem

Implement SlackBot

Describe the settings

Describe the settings of SlackBot.

SlackBotService/bot_settings.py

# -*- coding: utf-8 -*-

import os
from slackbot import settings

settings.API_TOKEN = "XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX"
settings.DEFAULT_REPLY = u"It is not well understood."
settings.PLUGINS = [
    'plugins',
]

os.environ['WEBSOCKET_CLIENT_CA_BUNDLE'] = os.path.join('./bin', 'cacert.pem')

The following part loads the above key file which is not bundled at the time of packaging. Of course, I get a warning, but I didn't want to pollute the environment of Pepper itself, so I did this once.

os.environ['WEBSOCKET_CLIENT_CA_BUNDLE'] = os.path.join('./bin', 'cacert.pem')

Implement the main script that is started as a service

Run SlackBot when running the script.

SlackBotService/slackbotservice.py

# -*- coding: utf-8 -*-

from slackbot.bot import Bot
from bot_settings import *


class SlackBotService:
    def __init__(self):
        self.bot = Bot()
        self.bot.run()


def main():
    slackbotservice = SlackBotService()


if __name__ == "__main__":
    main()

Implemented SlackBot plugin

If something is spoken, try to make Pepper speak.

SlackBotService/lib/plugins/mention.py

# -*- coding: utf-8 -*-

from slackbot.bot import respond_to, listen_to
from slacker import Slacker
from plugins import proxy
from bot_settings import settings

slack = Slacker(settings.API_TOKEN)


@respond_to(u'')
def respond_any_word(message):
    """
Receive any character.
    :param message:Received message object
    """
    _text = message.body.get('text', '')
    _ts = message.body.get('ts', '')
    _user = message.body.get('user', '')
    _team = message.body.get('team', '')
    _type = message.body.get('type', '')
    _channel = message.body.get('channel', '')
    proxy.animated_speech(_text)
    message.reply("done")

Implemented a proxy to make Pepper speak

SlackBotService/lib/plugins/proxy.py

# -*- coding: utf-8 -*-

from naoqi import ALProxy

animatedSpeechProxy = ALProxy("ALAnimatedSpeech", "127.0.0.1", 9559)


def animated_speech(text):
    _text = text if isinstance(text, str) else text.encode("utf-8")
    configuration = {"bodyLanguageMode": "contextual"}
    animatedSpeechProxy.say(_text, configuration)

Package your app

qiproject.xml Define the module to bundle. All necessary modules such as libraries and scripts are defined.

<project version="3">
    <qipython name="slackbotservice">
        <package name="backports" src="lib" />
        <package name="slackbot" src="lib" />
        <package name="slacker" src="lib" />
        <package name="websocket" src="lib" />
        <package name="plugins" src="lib" />
        <script src="lib/six.py" />
        <script src="lib/cacert.pem" />
        <script src="slackbotservice.py" />
        <script src="bot_settings.py" />
    </qipython>
</project>

manifest.xml Define the service. The \ tag defines information as a service recognized within Pepper.

name --Define the physical name of the service. You can call the service from various qicli commands and apps with the name defined here.
autorun --Specify true so that it starts automatically.
execStart --Define a script to run as a service. Make sure to use python bundled with your application.

<package uuid="slackbotservice" version="0.1.0">
    <services>
        <service name="SlackBotService" autorun="true"
            execStart="/bin/bash ./python bin/slackbotservice.py" />
        <executableFiles>
            <file path="python" />
        </executableFiles>
    </services>
    <requirements>
        <naoqiRequirement minVersion="2.3"/>
        <robotRequirement model="JULIETTE"/>
    </requirements>
</package>

.pml file

Define the package.

<?xml version="1.0" encoding="UTF-8" ?>
<Package name="slackbotservice" format_version="4">
    <Manifest src="manifest.xml" />
    <qipython name="slackbotservice" />
</Package>

Project packaging

$ qipy install slackbotservice
$ qipkg make-package slackbotservice.pml

Deploy package

$ qipkg deploy-package slackbotservice-0.1.0.pkg --url nao@<Pepper IP>

Check the operation