transcribe.py

import os
from pytube import YouTube
import openai

openai.api_key = os.environ.get("OPENAI_API_KEY")


class AudioTranscriber:
    def __init__(self, url, audio_path):
        self.url = url
        self.audio_path = audio_path

    def download_audio(self):
        yt = YouTube(self.url)
        video = yt.streams.filter(only_audio=True).first()

        download_path = 'audios/'
        if not os.path.isdir(download_path):
            os.makedirs(download_path)
        out_file = video.download(output_path=download_path)
        base, ext = os.path.splitext(out_file)
        new_file = f"{base}.mp3"
        os.rename(out_file, new_file)
        file_name = os.path.basename(new_file)
        file_path = os.path.join(download_path, file_name)
        return file_path

    def transcribe_audio(self, file):
        transcription = openai.Audio.transcribe("whisper-1", file)
        return transcription;

    def transcribe(self):
        try:
            audio_path = self.download_audio()
            audio_data = open(audio_path, "rb")
            transcription = self.transcribe_audio(audio_data)
            audio_data.close()
            os.remove(audio_path)
            return transcription
        except Exception as e:
            print("Error occurred: ", e)
            return "cannot transcribe"

class Summarizer:
    """
    Summarizer class that uses OpenAI API to generate summaries
    """
    def __init__(self, content):
        self.content = self.replace_non_alphanumeric(content);
        self.max_tokens = 512
        self.max_len = 7500

    def replace_non_alphanumeric(self, string):
        # Use regex to replace all non-alphanumeric characters with an empty space
        return re.sub(r'[^a-zA-Z0-9]', ' ', string)


    def summarize(self):
        # Split text into chunks that are within the token limit
        chunks = self.split_text(self.content);
        summary = ""

        # Generate summary for each chunk of text
        for chunk in chunks:

            prompt = chunk + "\nTl;dr"

            # Generate summary using OpenAI API.
            response = openai.Completion.create(
                engine="text-davinci-003",
                prompt=prompt,
                max_tokens=self.max_tokens,
                n=1,
                stop=None,
                temperature=0.7,
            )

            generated_summary = response.choices[0].text.strip()
            summary += generated_summary

        return summary

    def split_text(self, text):
        chunks = []
        for i in range(0, len(text), self.max_len):
            chunks.append(text[i:i+self.max_len])
        return chunks


if __name__ == "__main__":
    # Test YouTube transcriber
    url = input("Enter YouTube URL: ")
    transcriber = AudioTranscriber(url, "audios/");
    transcribed_text = transcriber.transcribe();
    print("\n\n", transcribed_text);
    # Test summarizer
    summary = Summarizer(transcribed_text).summarize();
    print("\n\n", summary);