update to initial prompting for StyleVehicle

app.py

@@ -80,19 +80,62 @@ if uploaded_file:
 else:
     pdf_text = ""
 
+#"""
+#initial_prompt = f
+#Imagine you are a seasoned researcher specializing in the field of {research_field}. 
+#You are presented with a research paper within your domain. Evaluate its working methodology 
+#and discuss its research impact through concise bullet points. Conclude by summarizing the 
+#research paper and propose three questions for the user based on the paper's context. Finnaly 
+#remeber the research paper context for the next questions.
+
+#Output will be as,
+#Research Paper Title \n
+#Research Summary \n
+#Methodology \n
+#Research Impact \n
+#Suggested Questions
+
+#"""
+
+
 initial_prompt = f"""
 Imagine you are a seasoned researcher specializing in the field of {research_field}. 
-You are presented with a research paper within your domain. Evaluate its working methodology 
-and discuss its research impact through concise bullet points. Conclude by summarizing the 
-research paper and propose three questions for the user based on the paper's context. Finnaly 
-remeber the research paper context for the next questions.
+You are presented with a research paper within your domain. Evaluate its working methodology including model architecture
+which can be Predefined, Custom Predefined, Own Model or other, explain with architecture
+and discuss its research impact through concise bullet points. Research about dataset conditions and domain adaptation technqiues.
+Conclude by summarizing the 
+research paper and propose three questions for the user based on the paper's context. Remember 
+the research paper context for the next questions.
 
 Output will be as,
-Research Paper Title \n
-Research Summary \n
-Methodology \n
-Research Impact \n
-Suggested Questions"""
+Title\n 
+Research Summary\n 
+Methods or Models Used  
+Dataset Used 
+- Name: 
+- Size: 
+  - Number of Images: 
+  - Resolution: 
+  - Instances: 
+Dataset Condition 
+- Day, Night, Mix, Weather: 
+Domain Adaptation Technique:
+- Used (or not) 
+- Method: 
+  - D2N (Day to Night) 
+  - N2D (Night to Day) 
+  - Mix Image Generation 
+Output Image Resolution 
+- Resolution: 
+- Upsampling Technique Used: 
+Experiment Process and Objectives:
+- Steps: 
+- Objectives: 
+Detection/Segmentation Models:
+Performance Improvements:
+- Tradeoff: 
+Suggested Questions:
+"""
 
 if option=='':
     with st.spinner("Processing..."):

video_transcriber.py

@@ -0,0 +1,57 @@
+import moviepy.editor as mp
+import speech_recognition as sr
+import os
+
+def transcribe_audio_chunk(audio_chunk):
+    recognizer = sr.Recognizer()
+    with sr.AudioFile(audio_chunk) as source:
+        audio_data = recognizer.record(source)
+
+    try:
+        text_result = recognizer.recognize_google(audio_data, language='en-US')
+        return text_result
+    except sr.UnknownValueError:
+        print("Speech Recognition could not understand audio")
+        return ""
+    except sr.RequestError as e:
+        print(f"Could not request results from Google Speech Recognition service; {e}")
+        return ""
+
+def transcribe_video(video_path):
+    # Step 1: Extract audio from the video
+    video_clip = mp.VideoFileClip(video_path)
+    audio_clip = video_clip.audio
+    audio_clip.write_audiofile("temp_audio.wav")
+
+    # Step 2: Split the audio into smaller chunks (e.g., 10 seconds each)
+    chunk_duration = 10  # in seconds
+    total_duration = audio_clip.duration
+    chunk_paths = []
+
+    for start_time in range(0, int(total_duration), chunk_duration):
+        end_time = min(start_time + chunk_duration, total_duration)
+        chunk_path = f"temp_audio_chunk_{start_time}_{end_time}.wav"
+        audio_chunk = audio_clip.subclip(start_time, end_time)
+        audio_chunk.write_audiofile(chunk_path)
+        chunk_paths.append(chunk_path)
+
+    # Step 3: Transcribe each audio chunk
+    transcribed_texts = []
+    for chunk_path in chunk_paths:
+        text_result = transcribe_audio_chunk(chunk_path)
+        transcribed_texts.append(text_result)
+
+    # Step 4: Concatenate the transcribed texts
+    final_transcription = " ".join(transcribed_texts)
+    print("Transcription:\n", final_transcription)
+
+    # Clean up temporary files
+    audio_clip.close()
+    video_clip.close()
+    os.remove("temp_audio.wav")
+    for chunk_path in chunk_paths:
+        os.remove(chunk_path)
+
+# Example usage
+video_path = "C:/Users/HP/Downloads/Video/1.mp4"
+transcribe_video(video_path)