[PYTHON] I made a Discord chat reading bot by imitating the appearance

Introduction

The chat reading bot used in Discord has temporarily become unusable, so I made it on a temporary basis. I made it by rewriting a part of the code referring to Creating a bot that reads out Discord chat.

Advance preparation

I did the bot creation and environment construction while looking at the following.

-[Discord bot creation tutorial](https://ikayome.hateblo.jp/entry/2019/07/03/Discord_bot%E4%BD%9C%E6%88%90%E3%83%81%E3%83% A5% E3% 83% BC% E3% 83% 88% E3% 83% AA% E3% 82% A2% E3% 83% AB) -Speech synthesis on Windows Open JTalk -Install ffmpeg on windows

ffmpeg puts the unzipped files in C: \ open_jtalk \ bin. Don't forget the Path setting ...

bot source code

Please rewrite the token of each bot in the part of the bottom line, which is the token copied when the bot was created.

read_bot.py

import discord
from discord.ext import commands
import asyncio
import os
import subprocess
import ffmpeg
from voice_generator import creat_WAV

client = commands.Bot(command_prefix='.')
voice_client = None


@client.event
async def on_ready():
    print('Logged in as')
    print(client.user.name)
    print(client.user.id)
    print('------')


@client.command()
async def join(ctx):
    print('#Get voicechannel')
    vc = ctx.author.voice.channel
    print('#Connect to voice channel')
    await vc.connect()

@client.command()
async def bye(ctx):
    print('#Disconnect')
    await ctx.voice_client.disconnect()

@client.event
async def on_message(message):
    msgclient = message.guild.voice_client
    if message.content.startswith('.'):
        pass

    else:
        if message.guild.voice_client:
            print(message.content)
            creat_WAV(message.content)
            source = discord.FFmpegPCMAudio("output.wav")
            message.guild.voice_client.play(source)
        else:
            pass
    await client.process_commands(message)


client.run("Token copied when creating the bot")

Audio file source code

There are various comments out in the voice file path, but if you download the Source code of Sample Script in MMD Agent You can change it to a female voice. (Since there is a htsvoice file in the Voice / mei directory of the unzipped file, place the mei directory in the same hierarchy as the source code.)

voice_generator.py

import subprocess
import re

# remove_custom_emoji
#Emoji ID is not read aloud
def remove_custom_emoji(text):
    pattern = r'<:[a-zA-Z0-9_]+:[0-9]+>'    #Custom emoji pattern
    return re.sub(pattern,'',text)   #Replacement process

# urlAbb
#Omitted for URL
def urlAbb(text):
    pattern = "https?://[\w/:%#\$&\?\(\)~\.=\+\-]+"
    return re.sub(pattern,'The URL is omitted!',text)   #Replacement process

# creat_WAV
# message.Write content to a text file
def creat_WAV(inputText):
        # message.Write content to a text file

    inputText = remove_custom_emoji(inputText)   #Emoji ID is not read aloud
    inputText = urlAbb(inputText)   #Omitted for URL
    input_file = 'input.txt'

    with open(input_file,'w',encoding='shift_jis') as file:
        file.write(inputText)

    command = 'C:/open_jtalk/bin/open_jtalk -x {x} -m {m} -r {r} -ow {ow} {input_file}'

    #Dictionary Path
    x = 'C:/open_jtalk/bin/dic'

    #Voice file Path
    m = 'C:/open_jtalk/bin/nitech_jp_atr503_m001.htsvoice'
    #m = 'C:/open_jtalk/bin/mei/mei_sad.htsvoice'
    #m = 'C:/open_jtalk/bin/mei/mei_angry.htsvoice'
    #m = 'C:/open_jtalk/bin/mei/mei_bashful.htsvoice'
    #m = 'C:/open_jtalk/bin/mei/mei_happy.htsvoice'
    #m = 'C:/open_jtalk/bin/mei/mei_normal.htsvoice'

    #Vocalization speed
    r = '1.0'

    #Output file name and Path
    ow = 'output.wav'

    args= {'x':x, 'm':m, 'r':r, 'ow':ow, 'input_file':input_file}

    cmd= command.format(**args)
    print(cmd)

    subprocess.run(cmd)
    return True

if __name__ == '__main__':
    creat_WAV('test')

Directory structure

The composition is like this.

How to use

There was a description up to the point of inviting the bot, but it was not written about how to use it, so I will write it appropriately.

Command prompt start

Drag and drop "read_bot.py" from C: \ open_jtalk \ bin to the command prompt and Enter

Then the bot will log in

Call to voice channel

First of all, I enter the voicing. (This bot basically comes in at the person who called the voicing.) The command to call is .join

Let me talk

If you chat appropriately, it will be read. It doesn't read emojis or custom emojis ... Also, I try to omit the URL.

Evacuate

The command to leave the bot from the voice channel is .bye

Clean up

At the command prompt, use Ctrl + C or close the window.

end

I feel that even simple English can be read in alphabets, so I'd like to add a dictionary registration function.

Well it will be done soon. maybe