From 168a66f412d834bd54407e48090d66d10009fe6b Mon Sep 17 00:00:00 2001
From: sindre0830 <sindre0830@gmail.com>
Date: Sun, 6 Mar 2022 17:19:55 +0100
Subject: [PATCH] [#59] Replace Pafy with YT-DLP to fix download speed

---
 NN/API/preprocessing.py | 40 ++--------------------------------------
 NN/API/requirements.txt |  2 +-
 2 files changed, 3 insertions(+), 39 deletions(-)

diff --git a/NN/API/preprocessing.py b/NN/API/preprocessing.py
index 9320041..a576bba 100644
--- a/NN/API/preprocessing.py
+++ b/NN/API/preprocessing.py
@@ -2,9 +2,6 @@
 import dictionary as dict
 # import foreign modules
 import os
-from pathlib import Path
-from pydub import AudioSegment
-import pafy
 
 
 # Downloads an audio file from given URL.
@@ -14,38 +11,5 @@ def downloadAudio(id):
         os.makedirs(dict.NATIVE_DIR)
     # branch if audio file doesn't exist
     if not os.path.isfile(dict.getNativeAudioPath(id)):
-        url = "https://www.youtube.com/watch?v=" + id
-        audiostreams = pafy.new(url).audiostreams
-        # get audio format with best quality
-        best = 0
-        for idx, val in enumerate(audiostreams):
-            temp = int(val.get_filesize())
-            if best == 0 or temp > best:
-                best = idx
-            print(val.bitrate, val.extension, val.get_filesize())
-        tempFilename = id + "." + audiostreams[best].extension
-        # download audio file
-        if os.path.exists(dict.NATIVE_DIR + tempFilename) is False:
-            audiostreams[best].download(filepath=dict.NATIVE_DIR + tempFilename)
-        # convert file to wav format and remove temporary file
-        convertToWav(id, tempFilename)
-        os.remove(dict.NATIVE_DIR + tempFilename)
-
-
-# Attempts to convert a file into wav format.
-def convertToWav(id, file):
-    path = dict.NATIVE_DIR + file
-    newPath = dict.getNativeAudioPath(id)
-    if os.path.exists(newPath) is False and testExt(file):
-        sound = AudioSegment.from_file(path)
-        sound.export(newPath, format="wav")
-    else:
-        pass
-
-
-# Checks if file is a supported audio format.
-def testExt(file):
-    if Path(file).suffix in dict.EXTENSIONS:
-        return True
-    else:
-        return False
+        # download audio file with best quality then convert to wav
+        os.system("yt-dlp -q -f 'ba' -x --audio-format wav https://www.youtube.com/watch?v=" + id + " -o '"+ dict.NATIVE_DIR + "%(id)s.%(ext)s'")
diff --git a/NN/API/requirements.txt b/NN/API/requirements.txt
index cbd2746..7c95ed8 100644
--- a/NN/API/requirements.txt
+++ b/NN/API/requirements.txt
@@ -1,8 +1,8 @@
 flask
 pydub
-pafy
 youtube-dl==2020.12.02
 numpy==1.21
 librosa
 matplotlib
 aubio
+yt-dlp