etwas/transcribe_bot
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Update dependencies, openai to v1

Browse source
This commit is contained in:
EinEtw4s 2024-05-09 15:48:00 +02:00
parent 2506df7f23
commit 1c52ccb0a9
3 changed files with 23 additions and 16 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

18
discord_bot.py
View file

@ -1,4 +1,4 @@
import openai from openai import OpenAI
import ffmpeg import ffmpeg
import os import os
import discord import discord
@ -12,9 +12,10 @@ Path("temp/").mkdir(parents=True, exist_ok=True)
config = configparser.ConfigParser() config = configparser.ConfigParser()
config.read('config.ini') config.read('config.ini')
openai.api_key = config['DEFAULT']['OpenAIKey'] # your OpenAI-API Key openai = OpenAI(api_key=config['DEFAULT']['OpenAIKey']) # initiate with your openai key
allowed_senders = config['DEFAULT']['AllowedDiscordSenders'].split(',') # list of allowed Discord accounts allowed_senders = config['DEFAULT']['AllowedDiscordSenders'].split(',') # list of allowed Discord accounts
allowed_channel_names = config['DEFAULT']['AllowedDiscordChannels'].split(',') # allowed channel names allowed_channel_names = config['DEFAULT']['AllowedDiscordChannels'].split(',') # allowed channel names
class TranscribeClient(discord.Client): class TranscribeClient(discord.Client):
async def on_ready(self): async def on_ready(self):
@ -35,13 +36,14 @@ class TranscribeClient(discord.Client):
# convert to mp3 # convert to mp3
input = ffmpeg.input('temp/in') input = ffmpeg.input('temp/in')
out = ffmpeg.output(input,'temp/out.mp3') out = ffmpeg.output(input, 'temp/out.mp3')
ffmpeg.run(out, overwrite_output=True) ffmpeg.run(out, overwrite_output=True)
audio_file = open("temp/out.mp3", "rb") audio_file = open("temp/out.mp3", "rb")
# transcribe using OAI Whisper # transcribe using OAI Whisper
transcript = openai.Audio.transcribe("whisper-1", audio_file) transcript = openai.audio.transcriptions.create(model="whisper-1", file=audio_file)
text = transcript['text']
text = transcript.text
# respond with the transcription # respond with the transcription
await message.channel.send(f'**{attachment.filename}:**') await message.channel.send(f'**{attachment.filename}:**')
@ -54,7 +56,7 @@ async def main():
return return
if __name__=='__main__': if __name__ == '__main__':
token = config['DEFAULT']['DiscordToken'] token = config['DEFAULT']['DiscordToken']
intents = discord.Intents.default() intents = discord.Intents.default()
intents.message_content = True intents.message_content = True

4
requirements.txt Normal file
View file

@ -0,0 +1,4 @@
discord.py==2.3.2
ffmpeg_python==0.2.0
openai==1.27.0
python-telegram-bot==21.1.1

17
telegram_bot.py
View file

@ -13,9 +13,10 @@ Path("temp/").mkdir(parents=True, exist_ok=True)
config = configparser.ConfigParser() config = configparser.ConfigParser()
config.read('config.ini') config.read('config.ini')
openai.api_key = config['DEFAULT']['OpenAIKey'] openai = OpenAI(api_key=config['DEFAULT']['OpenAIKey']) # initiate with your openai key
allowed_senders = config['DEFAULT']['AllowedTelegramSenders'].split(',') allowed_senders = config['DEFAULT']['AllowedTelegramSenders'].split(',')
async def on_msg(update: Update, context: ContextTypes.DEFAULT_TYPE) -> None: async def on_msg(update: Update, context: ContextTypes.DEFAULT_TYPE) -> None:
message = update.message message = update.message
sender_id = message.from_user.id sender_id = message.from_user.id
@ -25,25 +26,25 @@ async def on_msg(update: Update, context: ContextTypes.DEFAULT_TYPE) -> None:
f = await message.effective_attachment.get_file() f = await message.effective_attachment.get_file()
await f.download_to_drive('temp/in') await f.download_to_drive('temp/in')
input = ffmpeg.input('temp/in') input = ffmpeg.input('temp/in')
out = ffmpeg.output(input,'temp/out.mp3') out = ffmpeg.output(input, 'temp/out.mp3')
ffmpeg.run(out, overwrite_output=True) ffmpeg.run(out, overwrite_output=True)
audio_file = open("temp/out.mp3", "rb") audio_file = open("temp/out.mp3", "rb")
transcript = openai.Audio.transcribe("whisper-1", audio_file) transcript = openai.audio.transcriptions.create(model="whisper-1", file=audio_file)
text = transcript['text'] text = transcript.text
for t in textwrap.wrap(text, 4000): for t in textwrap.wrap(text, 4000):
await update.message.reply_text(t) await update.message.reply_text(t)
async def main(): async def main():
return return
if __name__=='__main__': if __name__ == '__main__':
token = config['DEFAULT']['TelegramToken'] token = config['DEFAULT']['TelegramToken']
application = Application.builder().token(token).build() application = Application.builder().token(token).build()
application.add_handler(MessageHandler(filters=None,callback=on_msg)) application.add_handler(MessageHandler(filters=None, callback=on_msg))
application.run_polling() application.run_polling()