@quasi meteor
Python help channel opened
#🔒 New to python, wrote a simple file-copying script and it seemed to work fine, until!
winged irisBOT
Python help channel opened
quasi meteor
well darn, that was a long message to have deleted. Any way to recover just the text of the message and I can send the script correctly?
winged irisBOT
mellow pivot
!paste
winged irisBOT
quasi meteor
So, I am trying to copy files from old DVDs from a few decades ago, onto my laptop.
File Explorer was being a pain, and kept hanging randomly when it tried copying certain files.
and even worse, it only copied some of the files, leaving random gaps in my folder!
I didn't want to manually check all the filenames to see what to copy,
so I wrote a quick python script so that I would be able to copy the files faster without messing with File Explorer.
[Note: This script does not validate whether or not the files are equivalent, all it does is check the filenames.]
Something odd is happening. When my script copies .jpg or .mov files, occasionally, it will copy a file that is 2.45 GB!
What a specific number! I cannot open these JPGs, but now, I have dozens of 2.45 GB corrupted JPGs on my computer 😱 whoops
I am guessing somehow, there's just no end of file to the jpgs or something, and it's somehow copying the entire disk in as if it were a jpg?
I am not sure! It's rather funny though 😂
Here is my script: https://paste.pythondiscord.com/BUCQ
(btw, I am new to the writing standards of python so if you have any feedback feel free to leave it too! happy to learn)
DVD Drive: Sandisk Ultra Slim Portable DVD Writer
Destination folder: D:/, m.2 2280 ssd
Is there something I can do to make my code safer and prevent it from attempting to copy files past a certain filesize limit?
Occasionally, the copying process will hang even in my script, and after 10 minutes, it will return
Problem copying IMG_7568.JPG: [Errno 13] Permission denied.
and the file will be copied partially like the attached image.
I would very much love to recover the photos that consistently corrupt like this, since they're baby photos!
Any help appreciated, thank you so much!
sharp folio
well darn, that was a long message to have deleted. Any way to recover just the ...
only "having a moderator copypaste it from the deletion logs"
quasi meteor
Yep, just had a mod help me with that, ty!
sharp folio
So, I am trying to copy files from old DVDs from a few decades ago, onto my lapt...
I feel like this can't possibly be an issue on the python side, but rather with your drive. Not sure it's very fixable, but, hmm
quasi meteor
Also, for more information:
My code assumes that the entire directory is nothing but files, it's not built to handle sub directories or anything like that,
since the disc I am copying from has no subdirectories in it (I might add code for handling that later)
In summary there are two unexpected problems that happen with my code:
-
Occasionally the program will copy files that are multiple gigabytes in size larger than they should be.
The size of every file copied is oddly consistently 2.45 GB. I just checked, and for a previous disk, there was the same problem,
except consistently 1.78 GB. -
Occasionally, the program will hang for 10-30 minutes, then print the error:
Problem copying {filename}. [Errno 13] Permission denied.(see line 102 in script)
I do not understand the "Permission denied" error code exactly, since it seems to copy the file partially all good,
until it hangs. Really odd problems!
sharp folio
(btw, I think you reinvented rsync)
quasi meteor
I feel like this can't possibly be an issue on the python side, but rather with ...
Yeah honestly that might be it, this disk reader was only 24$ so I might have to pitch in for a higher quality one
sharp folio
2.45GB sounds like it might be the size of the DVD disk (or one track of it, something like that)
quasi meteor
oh nice, haha
I might check that out! I figure someone out there has already made something like this before I just wanted to make it myself ;joy
quasi meteor
2.45GB sounds like it might be the size of the DVD disk (or one track of it, som...
yeah exactly, it's exactly the size of all the files on the disc combined
sharp folio
One thing you could try is copying manually - opening source and destination files, reading out, and writing into the other. That's inefficient compared to how copying files works normally, but it seems to me like it's also harder for it to go wrong
quasi meteor
Huh, just realized the files show up that size in the E: drive itself
so yeah this is definitely an issue of the disk reader and not the python script, darn
sharp folio
ah, then my idea won't help
to save the files from this nightmare drive you'll have to know their real sizes, somehow.
which... might actually be possible, since they're images
quasi meteor
I'll keep that idea in mind for the other problem, though, see if it helps with the files that are half chopped off
sharp folio
what format are these images in?
sharp folio
hmm, lemme experiment a little. I suspect if you tell ffmpeg to copy a jpg/mov and the file has garbage data at the end, it'll just ignore it
quasi meteor
||tangent that doesn't matter: one thing I just found out while browsing the docs, apparently .copy() doesn't preserve metadata but .copy2() does,
but somehow, my metadata is being preserved anyway even though I'm using .copy()? That's good since I didn't intend to wipe metadata, but interesting!||
sharp folio
hmm, lemme experiment a little. I suspect if you tell `ffmpeg` to copy a jpg/mov...
it looks like it indeed works this way, though only when not passing -codec copy, which is annoying since it means it'll get reencoded (from jpeg to jpeg)
so you can in theory copy all the image/video files that way, possibly losing some quality from the reencoding
def ffmpeg_transfer(src_path, dest_path):
"The destination path must have an extension (that's how ffmpeg knows what format to output.)"
subprocess.run(
[
"ffmpeg",
"-i",
src_path,
dest_path,
"-y",
]
)
something like that. You'll also need ffmpeg of course (it's a CLI tool; on windows the simplest way to install it would be via Chocolatey, though you can also just download the binaries, unpack them somewhere, and use the absolute path to ffmpeg.exe instead of "ffmpeg" in this snippet).
quasi meteor
ah ye I use ffmpeg occasionally to convert weird files to mp3!
I downloaded chocolatey awhile ago because someone told me to in a tutorial for glsl,
but I don't actually know what it is cause I haven't gotten past the first few minutes lmfao.
I just know it freaked me out when it started running this big intimidating powershell out of nowhere 😂
sharp folio
it's a package manager for windows (one of several existing ones). Lets windows users have the linux experience of installing things by just writing something like choco install ffmpeg.
quasi meteor
ah, ok! so kinda like nuget
good to know
Just realized the sizes of the files are actually mildly different. interesting!
makes me wonder if I downloaded the largest one, there was some potential of like. parsing it somehow, and that magically recovering the entire disc haha.
sharp folio
Just realized the sizes of the files are actually mildly different. interesting!
i suspect each file's size is the size of the entire disk past that file's location, or something
though normally files get fragmented... but maybe on CD drives they don't, since they are typically written in one shot.
sharp folio
makes me wonder if I downloaded the largest one, there was some potential of lik...
i'd go the ffmpeg way, but that's probably possible too, yeah
like, directly copying the disk to a file, and then converting it to iso, or something like that
winged irisBOT
Python help channel closed