youtube_video_transcriber.py

youtube_video_transcriber.py

@@ -0,0 +1,142 @@
+
+import whisper
+from pytube import YouTube
+import gradio as gr
+import os
+import re
+import logging
+from easygoogletranslate import EasyGoogleTranslate
+import nltk
+from nltk.corpus import stopwords
+from nltk.tokenize import word_tokenize, sent_tokenize 
+from typing import Tuple, Optional, List
+ 
+class VideoTranscriber: 
+
+    def __init__(self) -> None:
+
+        """
+        Initializes the VideoTranscriber class by loading the whisper model and downloading NLTK resources.
+        """      
+        self.model = whisper.load_model("base")
+        nltk.download("punkt")
+        nltk.download("stopwords")
+
+    def text_summarizer(text: str) -> str:
+
+        """
+        Summarizes the given text using a simple algorithm based on word frequency.
+
+        Args:
+            text (str): The input text to be summarized.
+
+        Returns:
+            str: The summary of the input text.
+        """      
+        # Tokenize the text into sentences and words
+        sentences = sent_tokenize(text)
+        words = word_tokenize(text)
+
+        # Remove stopwords (common words that don't add much meaning)
+        stop_words = set(stopwords.words("english"))
+        words = [word for word in words if word.lower() not in stop_words]
+
+        # Calculate word frequency
+        word_frequency = {}
+        for word in words:
+            if word not in word_frequency:
+                word_frequency[word] = 1
+            else:
+                word_frequency[word] += 1
+
+        # Calculate sentence scores based on word frequency
+        sentence_scores = {}
+        for sentence in sentences:
+            for word in word_tokenize(sentence.lower()):
+                if word in word_frequency:
+                    if sentence not in sentence_scores:
+                        sentence_scores[sentence] = word_frequency[word]
+                    else:
+                        sentence_scores[sentence] += word_frequency[word]
+
+        # Get the top 'num_sentences' sentences with highest scores
+        summary_sentences = sorted(sentence_scores, key=sentence_scores.get, reverse=True)[:4]
+
+        # Combine the selected sentences to form the summary
+        summary = " ".join(summary_sentences)
+
+        return summary
+
+
+    def get_text(self, url: str, language: str) -> Tuple[str, str]:
+
+        """
+        Transcribes the audio of a YouTube video and translates the result to the specified language.
+
+        Args:
+            url (str): The URL of the YouTube video.
+            language (str): The target language for translation.
+
+        Returns:
+            Tuple[str, str]: A tuple containing the transcribed text and its summary.
+        """      
+        #try:
+        if url != '':
+            output_text_transcribe = ''
+
+        yt = YouTube(url)
+        #video_length = yt.length --- doesn't work anymore - using byte file size of the audio file instead now
+        #if video_length < 5400:
+        video = yt.streams.filter(only_audio=True).first()
+        out_file=video.download(output_path=".")
+
+        file_stats = os.stat(out_file)
+        logging.info(f'Size of audio file in Bytes: {file_stats.st_size}')
+
+        if file_stats.st_size <= 30000000:
+            base, ext = os.path.splitext(out_file)
+            new_file = base+'.mp3'
+            os.rename(out_file, new_file)
+            a = new_file
+            result = self.model.transcribe(a)
+            text =  result['text'].strip()
+            translator = EasyGoogleTranslate(
+                source_language='en',
+                target_language=language,
+                timeout=10
+            )
+            result = translator.translate(text)
+            summary = self.text_summarizer(result)
+            return result, summary
+
+    def gradio_interface(self):
+
+        """
+        Sets up and launches the Gradio interface for YouTube video transcription.
+        """      
+        with gr.Blocks(css="style.css",theme= 'freddyaboulton/test-blue') as demo:
+            with gr.Column(elem_id="col-container"):
+                gr.HTML("""<center><h1 style="color:#fff">YouTube Video Transcriber </h1></center>""")
+                with gr.Row():
+                    with gr.Column(scale=0.4):
+                        language = gr.Dropdown(
+                        ["en","ta","te","hi","ml"], label="Select Language"
+                    )
+                    with gr.Column(scale=0.6):
+                        input_text_url = gr.Textbox(placeholder='Youtube video URL', label='YouTube URL',elem_classes="textbox")
+                with gr.Row():
+                    result_button_transcribe = gr.Button('Transcribe')
+
+                with gr.Row():
+                    output_text_transcribe = gr.Textbox(placeholder='Transcript of the YouTube video.', label='Transcript',lines=10)
+
+                with gr.Row():
+                    output_text_summary = gr.Textbox(placeholder='Summary of the YouTube video transcript.', label='Summary',lines=5)
+
+            result_button_transcribe.click(self.get_text, inputs = [input_text_url,language], outputs = [output_text_transcribe,output_text_summary] )
+
+            demo.launch(share = True)
+
+if __name__ == "__main__":
+    transcriber = VideoTranscriber()
+    transcriber.gradio_interface()