From 56899816177133cba62ee67a181b7af8e441c638 Mon Sep 17 00:00:00 2001
From: Kwangsoo Yeo <kyeo@picovoice.ai>
Date: Thu, 6 Jul 2023 11:00:13 -0700
Subject: [PATCH] move pytube to ytdlp (#115)

---
 demo/youtube/octotube.py      | 31 +++++++++++++++++--------------
 demo/youtube/requirements.txt |  2 +-
 res/spell-check/dict.txt      |  6 ++++--
 3 files changed, 22 insertions(+), 17 deletions(-)

diff --git a/demo/youtube/octotube.py b/demo/youtube/octotube.py
index 4b9fb15..08e8e70 100644
--- a/demo/youtube/octotube.py
+++ b/demo/youtube/octotube.py
@@ -14,13 +14,14 @@
 import time
 from argparse import ArgumentParser
 from threading import Thread
+from typing import *
 
 import pvoctopus
-from pytube import YouTube
+from yt_dlp import YoutubeDL
 
 
 class ProgressAnimation(Thread):
-    def __init__(self, prefix, step_sec=0.1):
+    def __init__(self, prefix: str, step_sec: float = 0.1):
         self._prefix = prefix
         self._step_sec = step_sec
         self._frames = [
@@ -48,17 +49,19 @@ def stop(self):
         self._done = True
 
 
-def download(url: str, folder: str) -> str:
-    webm_path = os.path.join(folder, '%s.webm' % url.split("watch?v=")[1])
-    if not os.path.exists(webm_path):
-        anime = ProgressAnimation('Downloading %s' % url)
-        anime.start()
-        youtube = YouTube(url)
-        audio_stream = youtube.streams.filter(only_audio=True, audio_codec='opus').order_by('bitrate').last()
-        audio_stream.download(output_path=folder, filename=os.path.basename(webm_path), skip_existing=True)
-        anime.stop()
-
-    return webm_path
+def download(url: str, output_dir: str) -> str:
+    ydl_opts = {
+        'outtmpl': "%(id)s.%(ext)s",
+        'format': 'bestaudio',
+        'paths': {
+            'home': output_dir
+        },
+        'geo_bypass': True
+    }
+    with YoutubeDL(ydl_opts) as ydl:
+        info = ydl.sanitize_info(ydl.extract_info(url, download=False))
+        ydl.download([url])
+        return os.path.join(output_dir, f"{info['id']}.webm")
 
 
 def main():
@@ -70,7 +73,7 @@ def main():
     parser.add_argument('--work-folder', default=os.path.expanduser('~/'))
     args = parser.parse_args()
 
-    webm_path = download(url=args.url, folder=args.work_folder)
+    webm_path = download(url=args.url, output_dir=args.work_folder)
 
     o = pvoctopus.create(access_key=args.access_key)
 
diff --git a/demo/youtube/requirements.txt b/demo/youtube/requirements.txt
index 5cda740..d7b608e 100644
--- a/demo/youtube/requirements.txt
+++ b/demo/youtube/requirements.txt
@@ -1,2 +1,2 @@
 pvoctopus
-pytube
\ No newline at end of file
+yt-dlp
\ No newline at end of file
diff --git a/res/spell-check/dict.txt b/res/spell-check/dict.txt
index c35d8fa..ba31aab 100644
--- a/res/spell-check/dict.txt
+++ b/res/spell-check/dict.txt
@@ -1,4 +1,5 @@
 Autocorrection
+bestaudio
 CMAKE
 Cocoapods
 Compat
@@ -29,15 +30,16 @@ magnifyingglass
 manzana
 octopusdemo
 octotube
+outtmpl
 perroquet
 porco
 porcospino
 pvbase
 pvoctopus
 pvoctopusdemo
-pytube
 recyclerview
 reindex
 wargv
 whitespaces
-xcworkspace
\ No newline at end of file
+xcworkspace
+ytdlp
\ No newline at end of file