Allow downloading to buffer

[iuriguilherme]

Checklist

• I'm reporting a feature request
• I've looked through the README
• I've verified that I'm running yt-dlp version 2022.03.08.1. (update instructions)
• I've searched the bugtracker for similar issues including closed ones. DO NOT post duplicates
• I've read the guidelines for opening an issue

Description

Although for the command line interface, almost all the time it is desirable to save the file to a local path. However when using ytdl as a module using the io library to save files to buffers may be desirable.

My use case is a bunch of chatbots which download a video from an URL and then uploads the file back to the user, allowing a interface for people with no access to a CLI or in an environment where it would be difficult to use (such as a locked smartphone).

It would be more efficient if the script can download the file to a io.BufferIO object then upload it without having to make a temporary file in the disk, as it already does with other kind of media (for example with a qr code generator).

[pukkandan]

The filename is what is passed around b/w the various classes/functions instead of a file object, with special handling for - (stdout). That makes this a bit hard to implement

[dirkf]

Indeed.

See also ytdl-org/youtube-dl#30815.

[pukkandan]

Assuming single threaded code, contextlib.redirect_stdout can be used as a workaround

[FirefoxMetzger]

I tried the contextlib.redirect_stdout workaround and it appears to work fine 😄. Here is an example snippet for youtube:

from yt_dlp import YoutubeDL
from contextlib import redirect_stdout
from pathlib import Path

youtube_id = "some-video-id"

ctx = {
    "outtmpl": "-",
    'logtostderr': True
}

buffer = io.BytesIO()
with redirect_stdout(buffer), YoutubeDL(ctx) as foo:
    foo.download([youtube_id])

# write out the buffer for demonstration purposes
Path(f"{youtube_id}.mp4").write_bytes(buffer.getvalue())

Edit: A caveat is that this only seems to work on linux. On Windows the redirect to stdout seems to break and I get a ERROR: unable to open for writing: fileno.

[pukkandan]

Edit: A caveat is that this only seems to work on linux. On Windows the redirect to stdout seems to break and I get a ERROR: unable to open for writing: fileno.

This should now work in windows too

[a-amir01]

@pukkandan, I was trying the above workaround and it works great. But If I want to download the video in a specific format

ctx = {
    "outtmpl": "-",
    'logtostderr': True,
    'format': '600'
}

buffer = io.BytesIO()
with redirect_stdout(buffer), YoutubeDL(ctx) as video:
    video.download([youtube_id])

print(buffer.getbuffer().nbytes / 1000000.0)

By the time it reaches the print statement the buffer is closed.

Traceback (most recent call last):
File "videoConverter.py", line 131, in
print(buffer.getbuffer().nbytes / 1000000.0)
ValueError: I/O operation on closed file.

But this only happens when the field 'format' is present in the ctx object.

Would you happen to know why this is happening and how I can fix it ?

[Grub4K]

You can overwrite the close method on the buffer, so YoutubeDL doesn't close it and you can keep using it as a file, like

old_buffer_close = buffer.close
buffer.close = lambda *_: ...
# YoutubeDL code
old_buffer_close()  # Close when done

[Tom1380]

Any idea on how to do this for subtitles? This is what I got so far:

import io
from contextlib import *
from yt_dlp import *
from pathlib import *

youtube_id = 'https://youtube.com/watch?v=0nwbzoIP14E'

ctx = {
    "outtmpl": "-",
    'logtostderr': True,
    'skip_download': True,
    'writesubtitles': True,
    'sub_langs': 'en.*',
}

buffer = io.BytesIO()
with redirect_stdout(buffer), YoutubeDL(ctx) as foo:
    foo.download([youtube_id])

print(buffer.getvalue())

It's still writing the subtitles to a file and the buffer seems to be empty.

[pukkandan]

You can't #3554