Introduction

The 3rd laboratory hackathon was held. Other members Article 1, Article 2 This time, I decided to make a Raspberry Pi that speaks tweets. As for the motive for creating it, the article that I referred to when I made the initial settings of Raspberry Pi for the first time described Raspberry Pi that speaks loudly tweets. .. When I read it, I felt that I would like to make a Raspberry Pi that speaks the tweets of ** Tsui Abolition ** in the laboratory someday. I referred to this time ** Article **

What you prepared

--Raspberry Pi 2 + WiFi dongle --MicroSD with Raspbian installed --Power adapter --Ordinary earphones (for speakers) --Keyboard, mouse and display

Speaker check

Check the sound of the speaker

$ speaker-test -t wav

You can stop the sound with Ctr + C. At this time, sound came out from the display connected by HDMI, so in order to make sound from the earphone (virtual speaker)

$ amixer cset numid=3 1

You can convert from HDMI output to analog output by inputting, and you can play sound from the speaker.

Volume control

$ alsamixer

Installation of OpenJtalk (speech synthesis software)

What is OpenJtalk

--Open source speech synthesis engine --The following is required to make OpenJtalk speak --Text with the content to speak --Dictionary folder --Audio file --OpenJtalk standard voice --nitech_jp_atr503_m001.jhtsvoice (male voice)

Install OpenJtalk / dictionary voice data

$ sudo apt-get update
$ sudo apt-get install open-jtalk 
$ sudo apt-get install open-jtalk-mecab-naist-jdic 
$ sudo apt-get install htsengine libhtsengine-dev 
$ sudo apt-get install hts-voice-nitech-jp-atr503-m001

If the message "Do you want to continue [Y / n]?" Is displayed on the way, type "Y" and press the "Enter" key to continue the installation.

Creating jtalk.sh

Next, we will create a script to run OpenJtalk.

`jtalk.sh`


#!/bin/sh
tmpfile=/tmp/jtalk.wav
htsvoice="/usr/share/hts-voice/\
nitech-jp-atr503-m001/\
nitech_jp_atr503_m001.htsvoice"
echo "$1" | open_jtalk \
-x /var/lib/mecab/dic/open-jtalk/naist-jdic \
-m $htsvoice \
-ow $tmpfile && \
aplay --quiet $tmpfile
rm $tmpfile

Give execute permission to the created script.

$ chmod 755 jtalk.sh

Test the following on the command line

$ ./jtalk.sh Hello

If you hear from the speaker as "Hello" OK!

Get Twitter API Access Token

Click "create an Apps"
Enter the required items (appropriate and okay)
Click "create" and then click "Details" of the created APP.
Select the "Keys and tokens" tab and click the "Generate" button for Access token & access token secret.
Make a note of the displayed "API key", "API secret key", "Access token", and "Access token secret".

Creating a Python program

python settings

#Creating a virtual environment
$ sudo apt-get install python3-dev python3-venv
$ python3 -m venv env
$ cd env
$ source bin/activate

Install tweepy

(env) $ python -m pip install tweepy

Create teitter.py

#-*- coding:utf-8 -*-
import tweepy
import os
#Set various keys
CONSUMER_KEY = 'xxxxxxxxxxxxxxx'
CONSUMER_SECRET = 'xxxxxxxxxxxxxxxxxxxxxxx'
auth = tweepy.OAuthHandler(CONSUMER_KEY, CONSUMER_SECRET)
ACCESS_TOKEN = 'xxxxxxxxxxxxxxxxxxxx'
ACCESS_SECRET = 'xxxxxxxxxxxxxxxxxxxxxxxxxx'
auth.set_access_token(ACCESS_TOKEN, ACCESS_SECRET)
#Create an API instance
api = tweepy.API(auth)
print('Done!') #Ready

#Get tweets
for status in tweepy.Cursor(api.user_timeline,screen_name = "ID of the acquired user",exclude_replies = True).items():
    tw_text = status.text
    tw_text = ''.join(tw_text.splitlines())
    tw_text = tw_text.replace('(',' ').replace(')',' ').replace(' ','')
    print(tw_text)
    os.system('/home/pi/jtalk.sh ' + tw_text)
    break #For the time being, only the latest tweets

Actually move!

Type the following code

(env) $ python twitter.py

The latest tweet of the specified user was spoken in a male voice! (Even if it's a key, he talked if I followed him.)

in conclusion

We have succeeded in making Raspberry Pi speak the latest tweets of the user specified in this hackathon, so I would like to improve it so that the specified user can speak it when tweeting!

Thank you for reading!

How to make a Raspberry Pi that speaks the tweets of the specified user