import gradio as gr, glob, os, auditok, zipfile, wave, pytube.exceptions, librosa, time from pytube import YouTube from moviepy.editor import VideoFileClip def download_video(url, download_as): try: yt = YouTube(url) except pytube.exceptions.RegexMatchError: raise gr.Error("URL not valid or is empty! Please fix the link or enter one!") video = yt.streams.get_highest_resolution() video.download() video_path = f"{video.default_filename}" video_clip = VideoFileClip(video_path) audio_clip = video_clip.audio if download_as == "wav": audio_clip.write_audiofile("output.wav") elif download_as == "mp3": audio_clip.write_audiofile("output.mp3") audio_clip.close() video_clip.close() for removalmp4 in glob.glob("*.mp4"): os.remove(removalmp4) return "Finished downloading! Please proceed to next tab." def split_audio_from_yt_video(mindur, maxdur, name_for_split_files, show_amount_of_files_and_file_dur): if show_amount_of_files_and_file_dur == True: gr.Warning(f"show_amount_of_files_and_file_dur set to True. This will take longer if your audio file is long.") if not os.path.exists("output.mp3") and not os.path.exists("output.wav"): raise gr.Error("Neither output.mp3 or output.wav exist! Did the video download correctly?") if mindur == maxdur: raise gr.Error(f"Cannot split mindur={mindur} and maxdur={maxdur}, min and max are the same number.") elif mindur > maxdur: raise gr.Error(f"Cannot split mindur={mindur} and maxdur={maxdur}, mindur is higher than maxdur.") elif name_for_split_files == None: raise gr.Error("Split files name cannot be empty!") else: audio_path = "output.wav" if not os.path.exists("output.mp3") else "output.mp3" audio_regions = auditok.split( audio_path, min_dur=mindur, max_dur=maxdur, max_silence=0.3, energy_threshold=45 ) os.remove(audio_path) for i, r in enumerate(audio_regions): filename = r.save(f"{name_for_split_files}-{i+1}.wav") for f in sorted(glob.glob("*.wav")): audio_files = glob.glob("*.wav") zip_file_name = "audio_files.zip" with zipfile.ZipFile(zip_file_name, "w") as zip_file: for audio_file in audio_files: zip_file.write(audio_file, os.path.basename(audio_file)) if show_amount_of_files_and_file_dur == False: for file2 in glob.glob("*.wav"): os.remove(file2) return "Files split successfully!\nCheck below for zipped files.", zip_file_name elif show_amount_of_files_and_file_dur == True: largest_file = ("", 0) total_files = 0 total_length = 0.0 for file_name in glob.glob("*.wav"): file_path = os.path.join(os.getcwd(), file_name) if file_path.lower().endswith(".wav"): try: with wave.open(file_path, 'r') as audio_file: frames = audio_file.getnframes() rate = audio_file.getframerate() duration = frames / float(rate) file_size = os.path.getsize(file_path) if file_size > largest_file[1]: largest_file = (file_name, file_size) total_length += duration total_files += 1 except wave.Error as e: raise gr.Error(f"Error reading file: {e}") length_mins = total_length / 60 for file2 in glob.glob("*.wav"): os.remove(file2) return f"Files split successfully!\nCheck below for zipped files.\n\n{total_files} files created, {length_mins:.2f} minutes total.", zip_file_name def analyze_audio(zip_file_path): with zipfile.ZipFile(zip_file_path, 'r') as zip_ref: zip_ref.extractall('unzipped_files') total_sample_rate = 0 total_files = 0 for file_name in os.listdir('unzipped_files'): if file_name.lower().endswith('.wav'): file_path = os.path.join('unzipped_files', file_name) try: with wave.open(file_path, 'r') as audio_file: sample_rate = audio_file.getframerate() total_sample_rate += sample_rate total_files += 1 except wave.Error as e: print(f"Error reading file: {e}") if total_files > 0: average_sample_rate = total_sample_rate / total_files return f"Average sample rate: {average_sample_rate}" else: return "No average sample rate could be found." def split_wav_or_mp3_file(audiofileuploader, mindur2, maxdur2, name_for_split_files2): if audiofileuploader == None: raise gr.Error("Audio file cannot be empty!") if mindur2 == maxdur2: raise gr.Error(f"Cannot split mindur={mindur} and maxdur={maxdur}, min and max are the same number.") elif mindur2 > maxdur2: raise gr.Error(f"Cannot split mindur={mindur} and maxdur={maxdur}, mindur is higher than maxdur.") elif name_for_split_files2 == None: raise gr.Error("Split files name cannot be empty!") else: audio_path = audiofileuploader audio_regions = auditok.split( audio_path, min_dur=mindur2, max_dur=maxdur2, max_silence=0.3, energy_threshold=45 ) os.remove(audio_path) for i, r in enumerate(audio_regions): filename = r.save(f"{name_for_split_files2}-{i+1}.wav") for f in sorted(glob.glob("*.wav")): audio_files = glob.glob("*.wav") zip_file_name2 = "audio_files.zip" with zipfile.ZipFile(zip_file_name2, "w") as zip_file: for audio_file in audio_files: zip_file.write(audio_file, os.path.basename(audio_file)) for file2 in glob.glob("*.wav"): os.remove(file2) return f"File split successfully!\nCheck below for zipped files.\nAmount created: {len(audio_files)}", zip_file_name2 def get_average_pitch(audio_file): if audio_file == None: raise gr.Error("Audio file cannot be empty!") else: y, sr = librosa.load(audio_file, sr=None) pitches, magnitudes = librosa.piptrack(y=y, sr=sr) mean_pitch = pitches.mean() return f"Average pitch: {mean_pitch:.2f} Hz" def all_in_one_inator(ytvideo, download_yt_video_as, min_duration, max_duration, name_for_outputted_split_files, progress=gr.Progress()): if download_as == "mp3": gr.Warning("MP3 is experimental, especially with this, so caution is advised.") if min_duration == max_duration: raise gr.Error(f"Cannot split mindur={min_duration} and maxdur={max_duration}, min and max are the same number.") elif min_duration > max_duration: raise gr.Error(f"Cannot split mindur={min_duration} and maxdur={max_duration}, mindur is higher than maxdur.") elif name_for_outputted_split_files == None: raise gr.Error("Split files name cannot be empty!") else: try: progress(0, "Downloading video...") yt = YouTube(ytvideo) except pytube.exceptions.RegexMatchError: raise gr.Error("URL not valid or was left empty! Please fix the link or enter one.") video = yt.streams.get_highest_resolution() video.download() video_path = f"{video.default_filename}" video_clip = VideoFileClip(video_path) audio_clip = video_clip.audio if download_yt_video_as == "wav": audio_clip.write_audiofile("output.wav") elif download_yt_video_as == "mp3": audio_clip.write_audiofile("output.mp3") audio_clip.close() video_clip.close() for removemp4 in glob.glob("*.mp4"): os.remove(removemp4) progress(0.5, "Video downloaded! Starting split process...") audio_path = "output.wav" if not os.path.exists("output.mp3") else "output.mp3" audio_regions = auditok.split( audio_path, min_dur=min_duration, max_dur=max_duration, max_silence=0.3, energy_threshold=45 ) os.remove(audio_path) for i, r in enumerate(audio_regions): filename = r.save(f"{name_for_outputted_split_files}-{i+1}.wav") for f in sorted(glob.glob("*.wav")): audio_files = glob.glob("*.wav") zip_file_name = "audio_files.zip" with zipfile.ZipFile(zip_file_name, 'w') as zip_file: for audio_file in audio_files: zip_file.write(audio_file, os.path.basename(audio_file)) for file2 in glob.glob("*.wav"): os.remove(file2) progress(1, "Done! Cleaning up...") time.sleep(2) return "Process done successfully! Check below for zipped files!", zip_file_name def download_video_as_audio_only(yt_video, audio_output_format): try: yt = YouTube(yt_video) except pytube.exceptions.RegexMatchError: raise gr.Error("URL not valid or is empty! Please fix the link or enter one!") video = yt.streams.get_highest_resolution() video.download() video_path = f"{video.default_filename}" video_clip = VideoFileClip(video_path) audio_clip = video_clip.audio if audio_output_format == "wav": audio_clip.write_audiofile("output.wav") audio_clip.close() video_clip.close() for mp4remove in glob.glob("*.mp4"): os.remove(mp4remove) single_zip_name = "only_audio.zip" audio_files = glob.glob("*.wav") if audio_output_format == "wav" else glob.glob("*.mp3") with zipfile.ZipFile(single_zip_name, 'w') as zip_file: for audio_file in audio_files: zip_file.write(audio_file, os.path.basename(audio_file)) for outputwavremoval in glob.glob("*.wav"): if os.path.exists(outputwavremoval): os.remove(outputwavremoval) for outputmp3removal in glob.glob("*.mp3"): if os.path.exists(outputmp3removal): os.remove(outputmp3removal) return "Done! Download the zip file below! This only contains the audio file.", single_zip_name with gr.Blocks(theme='sudeepshouche/minimalist', title="Global Dataset Maker") as app: gr.HTML( "