Update app.py

app.py

@@ -37,9 +37,8 @@ class VideoAnalytics:
       """
       # Initialize AzureOpenAI client
       self.client = AzureOpenAI()
-        
-      hf_token =os.getenv("HF_TOKEN")
-      self.mistral_client = InferenceClient("mistralai/Mixtral-8x7B-Instruct-v0.1",token=hf_token)
+      hf_key = os.getenv("HF_TOKEN")
+      self.mistral_client = InferenceClient("mistralai/Mixtral-8x7B-Instruct-v0.1",token=hf_key)
 
       # Initialize transcribed text variable
       self.transcribed_text = ""
@@ -87,40 +86,39 @@ class VideoAnalytics:
             logging.error(f"Error occurred while converting MP3 to WAV: {e}")
             raise e
 
-    def split_audio(self, input_file: str) -> list:
+    def split_audio(self, input_file: str) -> List[AudioSegment]:
         """
-        Split an audio file into segments of a specified length.
+        Split an audio file into segments of fixed length.
 
         Args:
             input_file (str): Path to the input audio file.
 
         Returns:
-            list: List of audio segments.
+            List[AudioSegment]: List containing segments of the input audio.
         """
         try:
-            # Load the audio file
-            audio = AudioSegment.from_file(input_file)
-    
-            # Define segment length in milliseconds (5 minutes = 300,000 milliseconds)
-            segment_length = 60000
-    
-            # Split the audio into segments
-            segments = []
-            for i, start_time in enumerate(range(0, len(audio), segment_length)):
-                # Calculate end time for current segment
-                end_time = start_time + segment_length if start_time + segment_length < len(audio) else len(audio)
-    
-                # Extract segment
-                segment = audio[start_time:end_time]
-    
-                # Append segment to list
-                segments.append(segment)
-    
-            return segments
-        except Exception as e:
-            print(f"An error occurred: {e}")
+          # Load the audio file
+          audio = AudioSegment.from_file(input_file)
+
+          # Define segment length in milliseconds (5 minutes = 300,000 milliseconds)
+          segment_length = 60000
+
+          # Split the audio into segments
+          segments = []
+          for i, start_time in enumerate(range(0, len(audio), segment_length)):
+              # Calculate end time for current segment
+              end_time = start_time + segment_length if start_time + segment_length < len(audio) else len(audio)
+
+              # Extract segment
+              segment = audio[start_time:end_time]
+
+              # Append segment to list
+              segments.append(segment)
+
+          return segments
+        except CouldntDecodeError as e:
+            logging.error(f"Error decoding audio: {e}")
             return []
-            
     # Function to recognize speech in the audio file
     def transcribe_audio(self,path: str,lang: str):
         """Transcribe speech from an audio file."""
@@ -205,6 +203,55 @@ class VideoAnalytics:
           logging.error(f"Error transcribing video: {e}")
           return ""
 
+    def extractive_summary(self,text: str):
+        """
+        Generate an extractive summary of the input text.
+
+        Args:
+            text (str): The input text to be summarized.
+
+        Returns:
+            str: The extractive summary of the input text.
+        """
+        try:
+          article_text =text
+          # Removing Square Brackets and Extra Spaces
+          article_text = re.sub(r'\[[0-9]*\]', ' ', article_text)
+          article_text = re.sub(r'\s+', ' ', article_text)
+          # Removing special characters and digits
+          formatted_article_text = re.sub('[^a-zA-Z]', ' ', article_text )
+          formatted_article_text = re.sub(r'\s+', ' ', formatted_article_text)
+          sentence_list = nltk.sent_tokenize(article_text)
+          stopwords = nltk.corpus.stopwords.words('english')
+
+          word_frequencies = {}
+          for word in nltk.word_tokenize(formatted_article_text):
+              if word not in stopwords:
+                  if word not in word_frequencies.keys():
+                      word_frequencies[word] = 1
+                  else:
+                      word_frequencies[word] += 1
+              maximum_frequncy = max(word_frequencies.values())
+          for word in word_frequencies.keys():
+              word_frequencies[word] = (word_frequencies[word]/maximum_frequncy)
+              sentence_scores = {}
+          for sent in sentence_list:
+              for word in nltk.word_tokenize(sent.lower()):
+                  if word in word_frequencies.keys():
+                      if len(sent.split(' ')) < 30:
+                          if sent not in sentence_scores.keys():
+                              sentence_scores[sent] = word_frequencies[word]
+                          else:
+                              sentence_scores[sent] += word_frequencies[word]
+          import heapq
+          summary_sentences = heapq.nlargest(12, sentence_scores, key=sentence_scores.get)
+
+          summary = ' '.join(summary_sentences)
+          return summary
+        except Exception as e:
+            logging.error(f"Error occurred during summarization: {e}")
+            return ""
+
     def generate_video_summary(self,model) -> str:
         """
         Generate a summary of the transcribe_video.
@@ -216,10 +263,9 @@ class VideoAnalytics:
             # Define a conversation between system and user
             conversation = [
               {"role": "system", "content": "You are a Summarizer"},
-              {"role": "user", "content": f"""summarize the following text delimited by triple backticks.Output must in english.give me a detailed summary.extractive summary working br like extract sentences from given text to return as summary,abstractive summary working be like summary of what about the given text.don't make bullet points write like a passage.
-                        In two format of Outputs given below:
+              {"role": "user", "content": f"""summarize the following text delimited by triple backticks.Output must in english.give me a detailed summary.abstractive summary working be like summary of what about the given text.don't make bullet points write like a passage.
+                        In this format of Outputs given below:
                         Abstractive Summary:
-                        Extractive Summary:
                         ```{self.english_text}```
                 """}
                 ]
@@ -237,12 +283,12 @@ class VideoAnalytics:
           elif model == "Mixtral":
               task = "summary"
               # Generate answer using Mixtral model
-              prompt = f"""<s>[INST] summarize the following text delimited by triple backticks.Output must in english.give me a detailed summary.extractive summary working br like extract sentences from given text to return as summary,abstractive summary working be like summary of what about the given text.don't make bullet points write like a passage.
-                      In two format of Outputs given below:
-                      Abstractive Summary:
-                      Extractive Summary:
+              prompt = f"""<s>[INST]summarize the following text delimited by triple backticks.Output must in english.give me a detailed summary.abstractive summary working be like summary of what about the given text.don't make bullet points write like a passage.
+                        In this format of Outputs given below:
+                        Abstractive Summary:
             ```data:{self.english_text}```[/INST]"""
               result = self.generate(prompt)
+              print("self.english_text",self.english_text)
               return result
 
         except Exception as e:
@@ -546,10 +592,13 @@ class VideoAnalytics:
               input_path = video
             else:
               return "Video Duration Above 10 Minutes,Try Below 10 Minutes Video","","",None,None,None
+          overall_summary = ""
           # Generate summary, important sentences, and topics
           summary = self.generate_video_summary(model)
-          self.write_text_files(summary,"Summary")
-          summary_voice = self.save_audio_with_gtts(summary,"summary.mp3")
+          extractive_summary = self.extractive_summary(self.english_text)
+          overall_summary = summary + "\n\n Extractive Summary: \n\n" + extractive_summary
+          self.write_text_files(overall_summary,"Summary")
+          summary_voice = self.save_audio_with_gtts(overall_summary,"summary.mp3")
           important_sentences = self.extract_video_important_sentence(model)
           self.write_text_files(important_sentences,"Important_Sentence")
           important_sentences_voice = self.save_audio_with_gtts(important_sentences,"important_sentences.mp3")
@@ -558,7 +607,7 @@ class VideoAnalytics:
           topics_voice = self.save_audio_with_gtts(topics,"topics.mp3")
 
           # Return the generated summary, important sentences, and topics
-          return summary,important_sentences,topics,summary_voice,important_sentences_voice,topics_voice
+          return overall_summary,important_sentences,topics,summary_voice,important_sentences_voice,topics_voice
 
         except Exception as e:
             # Log any errors that occur during video analytics
@@ -608,7 +657,7 @@ class VideoAnalytics:
                   result = gr.Textbox(label='Answer',lines=10)
               submit_btn.click(self.main,[video,yt_link,model_selection],[summary,Important_Sentences,Topics,summary_audio,important_sentence_audio,topics_audio])
               question.submit(self.video_qa,[question,model],result)
-        demo.launch(debug = True)
+        demo.launch()
 
 if __name__ == "__main__":
   video_analytics = VideoAnalytics()