any way to download selected vids from a substack post?

[abrogard]

There is one particular vid in a substack post that I would like at least a link to but preferably a way to download.

It is the one introduced by simplicius like this:

For the first time, we now have some actual details from a trustworthy source, in this case the German-Russian journalist Alina Lipp. You can listen to her account below and decide for yourself what to believe as far as the details:

But I can't find any way to do anything with it. I tried the firefox addon 'substack video downloader' but it seemed to do nothing at all.

I put the page URL in yt-dlp and it says 'page type newsletter' is not supported.

here is the page: https://simplicius76.substack.com/p/sitrep-5124-the-russian-steamroller?utm_source=pos...m_medium=email


Any help?

[2nHxWW6GkN1l916N3ayz8HQoi]

That site is a complete mess. Can't even distinguish the m3u8s and see which one goes for which video.

import re

import requests


def get_url(source_url, video_index):
    response = requests.get(source_url).content.decode()
    matches = re.findall('data-component-name="VideoEmbedPlayer" id="media-([^"]+)"', response)

    current_index = 0
    if video_index > 0:
        current_index = video_index - 1
        matches = [matches[current_index]]

    video_urls = []
    for media_id in matches:
        current_index += 1
        response = requests.get(
            f'https://simplicius76.substack.com/api/v1/video/upload/{media_id}/src',
            allow_redirects=False
        )

        video_urls.append((current_index, response.headers["Location"]))
    return video_urls


urls = get_url(
    source_url="https://simplicius76.substack.com/p/sitrep-5124-the-russian-steamroller?publication_id=1351274&post_id=144144149&isFreemail=true&r=2ld2be",
    video_index=11
)
for i, u in urls:
    print(f"Video index: {i} // Video url: {u}\n")

Output (shortened the url because it contained a token):

Video index: 11 // Video url: https://stream.mux.com/AyDz00FQ1CWIBpu1yc005Vh12ozEe63Q7s01FOplSjiWRQ/medium.mp4?token=ey...

Your file:
https://www.transfernow.net/dl/20240502rbkStvI5

I tested the python script with other simplicus pages. Seems to be working. Just gotta edit the source page url and the video index in the script. The index starts from 1 and represents what specific video you want from that page. If it's set to 0 or negative, it gets all the videos from that page.

[abrogard]

That's wonderful, really is. Thanks a lot.
It is nice to have that. Now I can spread it around to people/places that wouldn't want to be bothered with Simplicius's posts.
I guess I just run the python script from the command line?
Edit source page URL is fine. But video index? Just manually count them down the page or I could put something in there so's it would just get all of them, the urls for them?

[2nHxWW6GkN1l916N3ayz8HQoi]

I guess I just run the python script from the command line?

You run it like a normal python script. From cmd. Obviously you need to install python before.

python random_script.py

Edit source page URL is fine. But video index? Just manually count them down the page or I could put something in there so's it would just get all of them, the urls for them?

The videos have to be counted from top to bottom if you want something specifically. Your post title contained:

any way to download selected vids from a substack post?

and your first comment:

There is one particular vid in a substack post that I would like

Naturally, I assumed you were only interested in filtering and extracting specific content. Regardless I edited the posted script. All the urls point to mp4, so grab one in browser and right click save as.

[abrogard]

Yep, sure. You did exactly as I was asking for. I'm not complaining. Did I give you that impression? Nope.

Yep I have python. Use it from time to time but I don't 'know' it. I get help from gpt and try to make little utilities and sometimes it works and sometimes it doesn't:

gpt is pretty stupid I find and gets stupider and stupider when fault finding. Which is kinda good now I've learned that. For it makes me step in and start figuring the code which, of course, is what I ought to do.

I'm just old and lazy now. Work averse. I don't want to learn a language just for a little desktop app. Like I download Yahoo transcripts and wanted to make subtitles out of them. Is all. Just for one particular movie. Me and gpt worked for a couple or more days with dozens of attempts and I never finished up with a smooth procedure. Abandoned the whole thing. I vaguely think I found there's places on the web maybe that do it for you.

But that's the sort of thing.

I have military maps covering the whole of ukraine. each map covers a specific long and lat area of course. I wanted a routine that would save me the trouble of googling where a place was and then looking to find which map. ie. a routine where I could input the lat and long for that place and out would come the map number. that was a lengthy try, try, try, try again thing, too.. abandoned in the end....

And back to where we were - sorry about the digression, second glass of wine, late at night, I start to babble, yes I did only want specific content, that one vid. I queried the 'general get them all' thing lest it turned out to be harder to find the index number for the one you want than just counting down from the top. Remembering what you said about it being a total mess. I thought it possible their index numbers were all mixed up and might be easier for me to just get the lot. Not that it would be 'easier' in the end for I'd then have to look through them all and find which one I wanted but I mean it would be then the easiest way to ensure I got the one I wanted.

yep. Hey... thanks for your help.

[2nHxWW6GkN1l916N3ayz8HQoi]

Yep, sure. You did exactly as I was asking for. I'm not complaining. Did I give you that impression? Nope.

I'm not complaining either. No one is complaining. It just seemed funny to me you wanted X and then asked for Y. That's all. No worries.

Remembering what you said about it being a total mess. I thought it possible their index numbers were all mixed up and might be easier for me to just get the lot.

The only thing that is a mess about that site is the m3u8 management mechanism. If you filter in network requests for a m3u8 you have no idea which one is for which video. However, the video index position is set in stone for all pages from top to bottom. That's why I used that to distinguish requested videos since it's a constant (even if it might be tedious to obtain that information, the alternative would have been to input directly the media id, but lots of people don't know how to inspect the html page or what is that even about).

yep. Hey... thanks for your help.

You're welcome.

[abrogard]

Hi again,

I just tried your script for the first time and running python3 got an error

PS C:\WINDOWS\system32> python3 E:\downsubstackvid.py
Traceback (most recent call last):
File "E:\downsubstackvid.py", line 3, in <module>
import requests
ModuleNotFoundError: No module named 'requests'

did i screw something up with the copy paste somehow?


p.p.s.

I can't see how to insert code the way you did else I'd put my current version of your code up.
p.s. looking for the only vid on here this time. https://mcddd.substack.com/p/falling-between-two-stools

I set this:

urls = get_url(
source_url="https://simplicius76.substack.com/p/sitrep-5124-the-russian-steamroller?publication_id=1351274&post_id=1441441 49&isFreemail=true&r=2ld2be",
video_index=0

to 'index=0' because of what you said about the indexes and in order to make it 'general purpose' even if at the cost of downloading too much. In fact I rarely do this and usually there's very few vids.