3

I'm intending to use yt-dlp to download a video and then cut the video down afterward using ffmpeg. But to be able to use ffmpeg I am going to have to know the name of the file that yt-dlp produces. I have read through their documentation but I can't seem to find a way of getting the file name back into my program.

LT_Orange
  • 55
  • 2
  • 6
  • the [docs](https://github.com/yt-dlp/yt-dlp#output-template) describe how to pull the filenames or set your own – AudioBaton Oct 21 '22 at 18:56
  • @AudioBaton I've tried to set my own file name but it still still appends .f399 or another number to the end, I can't work out what this number is though? As for how to pull the file name, would you mind sending a small extract so I can try and find the section you are talking about. I've looked through but I seem to be missing it – LT_Orange Oct 21 '22 at 19:10
  • You can get it from an event.. but you have to extract it from a string, I'm using this regex: `/^[\s]*Merging formats into "(.+)"$/` I'm surprised this is not more straightforward. – user1529413 Jul 01 '23 at 13:48

2 Answers2

5

here are the doc examples

the numbers you mentioned (like .f399) I believe are temp only and are eventually removed when the final file is merged.

if you want to get the filename:

import subprocess
someFilename = subprocess.getoutput('yt-dlp --print filename https://www.youtube.com/something')
# then pass someFilename to FFmpeg

to use your own filename:

subprocess.run('yt-dlp -o thisIsMyName https://www.youtube.com/something')
# this will likely download a file named thisIsMyName.webm

but if you are not sure of the file type/extension beforehand and just want to get that:

someFileType = subprocess.getoutput('yt-dlp --print filename -o "%(ext)s" https://www.youtube.com/something')
print(someFileType)

it's not very efficient but to help explain it:

import subprocess

someFileType = subprocess.getoutput('yt-dlp --print filename -o "%(ext)s" https://www.youtube.com/something')
subprocess.run('yt-dlp -o "myFileName.%(ext)s" https://www.youtube.com/something')
subprocess.run(f'ffmpeg -i "myFileName.{someFileType}" outputFile.mp4')
AudioBaton
  • 361
  • 2
  • 6
4

According to the docs, you should be able to pass a progress hook and retrieve the filename there. It should be accessible with the filename attribute, but I have seen instances where it appends the .f399 and/or changes over time. I think it has to do with downloading videos in parts and is actually the temporary filename.

I found that by capturing the whole info_dict dictionary, there is a _filename key that appears to have the final filename.

Here is essentially what I did:

import yt_dlp

final_filename = None

def yt_dlp_monitor(self, d):
    final_filename  = d.get('info_dict').get('_filename')
    # You could also just assign `d` here to access it and see all the data or even `print(d)` as it updates frequently

ydl_opts = {
    "outtmpl": '/whatever/directory/%(uploader)s_%(title)s.%(ext)s',  # this is where you can edit how you'd like the filenames to be formatted
    "progress_hooks": [yt_dlp_monitor]  # here's the function we just defined
    }
with yt_dlp.YoutubeDL(ydl_opts) as ydl:
    info = ydl.extract_info("http://youtu.be/whatever_video)

Now final_filename will just keep updating with the filename every time that hook is called. You could have it only update on d['status'] == 'finished' if you like.

Dale W.
  • 51
  • 5