From e4085f2518bd96ba71260d48400cc3c19849cbab Mon Sep 17 00:00:00 2001
From: vasiliadi <9432426+vasiliadi@users.noreply.github.com>
Date: Thu, 29 Aug 2024 11:35:34 -0500
Subject: [PATCH] yt-dlp for youtube (#66)

* pytube replaced with yt-dlp

* yt-dlp docs link added

* yt-dlp added
---
 README.md        |  2 +-
 requirements.txt |  2 +-
 streamlit_app.py | 30 +++++++++++++++++++++++-------
 3 files changed, 25 insertions(+), 9 deletions(-)

diff --git a/README.md b/README.md
index cc569b3..d7b3f3a 100644
--- a/README.md
+++ b/README.md
@@ -104,7 +104,7 @@ You need to replace the path to the env_file in `compose.yaml`
 
 |  | Links |
 | ---|--- |
-| Libraries | [streamlit](https://docs.streamlit.io)<br> [replicate](https://replicate.com/docs/get-started/python)<br>[google-generativeai](https://ai.google.dev/gemini-api/docs/get-started/python)<br>[pytube](https://pytube.io/en/latest/) |
+| Libraries | [streamlit](https://docs.streamlit.io)<br> [replicate](https://replicate.com/docs/get-started/python)<br>[google-generativeai](https://ai.google.dev/gemini-api/docs/get-started/python)<br>~~[pytube](https://pytube.io/en/latest/)~~<br>[yt-dlp](https://github.com/yt-dlp/yt-dlp) |
 | Docker | [Docker Best Practices](https://testdriven.io/blog/docker-best-practices/)<br><br>[Docker](https://docs.docker.com/language/python/)<br>[Dockerfile reference](https://docs.docker.com/reference/dockerfile/)<br>[Dockerfile Linter](https://hadolint.github.io/hadolint/)<br><br>[.dockerignore](https://docs.docker.com/build/building/context/#dockerignore-files)<br><br>[Docker Compose](https://docs.docker.com/compose/)<br>[Syntax for environment files in Docker Compose](https://docs.docker.com/compose/environment-variables/env-file/)<br>[Ways to set environment variables with Compose](https://docs.docker.com/compose/environment-variables/set-environment-variables/)<br>[Compose file version 3 reference](https://docs.docker.com/compose/compose-file/compose-file-v3/)|
 | GitHub Actions | [Workflow syntax for GitHub Actions](https://docs.github.com/en/actions/using-workflows/workflow-syntax-for-github-actions)<br>[Publishing images to Docker Hub and GitHub Packages](https://docs.github.com/en/actions/publishing-packages/publishing-docker-images#publishing-images-to-docker-hub-and-github-packages) |
 | Dev Containers | [An open specification for enriching containers with development specific content and settings](https://containers.dev/)<br>[Developing inside a Container](https://code.visualstudio.com/docs/devcontainers/containers) |
diff --git a/requirements.txt b/requirements.txt
index 308476a..bffbe35 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,5 +1,5 @@
 streamlit==1.37.1
 google-generativeai==0.7.2
-pytube==15.0.0
 requests==2.32.3
 replicate==0.32.0
+yt-dlp==2024.8.6
diff --git a/streamlit_app.py b/streamlit_app.py
index 435605e..d95fe45 100644
--- a/streamlit_app.py
+++ b/streamlit_app.py
@@ -6,7 +6,7 @@
 import google.generativeai as genai
 import replicate
 import requests
-from pytube import YouTube
+from yt_dlp import YoutubeDL
 
 # Google Gemini config
 gemini_api_key = os.environ["GEMINI_API_KEY"]
@@ -54,9 +54,18 @@ def download(mode=st.session_state.mode):
             with open(audio_file_name, "wb") as f:
                 f.write(uploaded_file.getbuffer())
         case "YouTube link":
-            YouTube(yt_url).streams.filter(only_audio=True).order_by(
-                "abr"
-            ).asc().first().download(filename=audio_file_name)
+            ydl_opts = {
+                "format": "worstaudio",
+                "outtmpl": "audio",
+                "postprocessors": [
+                    {  # Extract audio using ffmpeg
+                        "key": "FFmpegExtractAudio",
+                        "preferredcodec": "mp3",
+                    }
+                ],
+            }
+            with YoutubeDL(ydl_opts) as ydl:
+                ydl.download(yt_url)
         case "Audio file link":
             downloaded_file = requests.get(audio_link)
             with open(audio_file_name, "wb") as f:
@@ -101,7 +110,11 @@ def summarize(audio_file_name=audio_file_name):
 def transcribe(model_name=st.session_state.model_name):
     match model_name:
         case "whisper-diarization":
-            latest_model_version = replicate_client.models.get("thomasmol/whisper-diarization").versions.list()[0].id
+            latest_model_version = (
+                replicate_client.models.get("thomasmol/whisper-diarization")
+                .versions.list()[0]
+                .id
+            )
             with open(converted_file_name, "rb") as audio:
                 transcription = replicate_client.run(
                     f"thomasmol/whisper-diarization:{latest_model_version}",
@@ -109,7 +122,11 @@ def transcribe(model_name=st.session_state.model_name):
                 )
                 return transcription
         case "incredibly-fast-whisper":
-            latest_model_version = replicate_client.models.get("vaibhavs10/incredibly-fast-whisper").versions.list()[0].id
+            latest_model_version = (
+                replicate_client.models.get("vaibhavs10/incredibly-fast-whisper")
+                .versions.list()[0]
+                .id
+            )
             with open(converted_file_name, "rb") as audio:
                 try:
                     transcription = replicate.run(
@@ -124,7 +141,6 @@ def transcribe(model_name=st.session_state.model_name):
                     st.error("Model error 😫 Try to switch model 👍", icon="🚨")
                     st.stop()
 
-
                 def detected_num_speakers(transcription):
                     speakers = [i["speaker"] for i in transcription[0:-1]]
                     return len(set(speakers))