ytget

Easily get data and download YouTube videos, focused on speed and simplicity.

Also works as a command-line extractor/downloader.

Installation

You can install ytget using pip:

pip install ytget

Some videos/formats will need ffmpeg to download. You can install it from the official website or by running:

sudo apt install ffmpeg

---

Features

• Simple use.
• Quick download and info extraction of youtube videos and playlists.
• Quick youtube search.
• Access to age restricted videos without login.
• Access to your private videos logging into your account.
• Command-line support.

---

Usage

Python

To extract information from a video, create a Video object with the url or query:

from ytget import Video

video = Video("https://www.youtube.com/watch?v=dQw4w9WgXcQ")

# Get info
title = video.title
duration = video.duration
subtitles = video.subtitles
stream_url = video.stream_url
# ...and so on. 
# You can use print(dir(video)) or help(video) to get all available info and parameters.

# Download the video
video.download()

# Change some parameters
video.download(path="downloads/", quality="med", only_audio=True)

You can also search for a query and get the information of all the videos obtained with the Search object:

from ytget import Search, Download
from ytget.utils import formatted_to_seconds

# Get the complete information of all videos
results = Search("never gonna give you up", get_duration_formatted=False).results

# Download all
results.download()

# Or do something with the results before downloading
filtered_results = list(filter(lambda x: x.duration < formatted_to_seconds('3:00'), results))
Download(filtered_results, quality="best", only_video=True, target_fps=30)

# Get simplified information (in case you need speed and don't need to download/get the stream urls)
results = Search("never gonna give you up", get_simple=True).results
for result in results:
    print(result['title'], result['url'])

Get information from a playlist with the Playlist object:

from ytget import Playlist, Fetch, Download

# Get the complete information of all videos
playlist = Playlist("https://www.youtube.com/watch?v=9OFpfTd0EIs&list=PLd9auH4JIHvupoMgW5YfOjqtj6Lih0MKw")

# Download all
playlist.download()

for video in playlist:
    print(video.get('title'), video.get('url'))

# Instead of downloading directly, you can do something with the videos before
videos = list(filter(lambda x: x.title.lower().startswith('a'), playlist.videos))
Download(videos)
        
# If you want to be the most efficient, get only the initial data of each video
videos_info = list(filter(lambda x: x.get('title').lower().startswith('b'), playlist.videos_info))
Download(Fetch(videos_info))

And (as shown in the previous examples) you can also use useful objects like Fetch and Download, and functions like formatted_to_seconds, format_seconds and delete_cache (among others) to make it even easier for you.

You can also use the GenericExtractor to extract info from various sites, though it's recommended to use a dedicated extractor for each website.

Command-line

For more detailed information, use:

ytget --help

Example 1 - Downloading a video and printing its title and url:

ytget https://www.youtube.com/watch?v=dQw4w9WgXcQ --print title url

Example 2 - Searching for a query for videos under 5 minutes long, and without downloading get the data of all the videos and write it to a json file:

ytget "never gonna give you up" --max-duration 5:00 --search --skip-download --print all --write-to-json

Example 3 - Get playlist info (with a maximum of 150 videos) and write to json file their titles, urls and ids:

ytget "https://www.youtube.com/playlist?list=PLd9auH4JIHvupoMgW5YfOjqtj6Lih0MKw" --max-length 150 --print title url video_id --skip-download --write-to-json

---

To Do

• Add playlist support.
• Add channels support.
• Allow some way to download livestreams (fractions).
• Allow search for multiple pages.

Known issues

• Issues related to downloading age restricted videos without logging in.
• When downloading some specific formats the result might be "corrupted". For now this can be fixed by enabling "force_ffmpeg".

PyPi

You can check the PyPi page of the package.

License

This project is licensed under the MIT License.