what files are modified on upload, and when?
timblaktu
Registered Users Posts: 6 Big grins
Hello. I need to know under what circumstances will Smugmug modify files that are uploaded?
I am working on some improvements to an open-source Smugmug backup application which would rely on comparing the md5 of a local file to the md5 of the corresponding remote file (on SmugMug).
https://github.com/tommyblue/smugmug-backup/issues/21
MD5 comparison is a good general way to compare binary "blobs" and jpg and mp4 are binary blobs. This approach to measuring "equality" would only be valid if Smugmug stores the same file, unaltered. So, I'm simply wanting to know what types of files, and under what circumstances, would Smugmug do this (modify an uploaded file)?
I read here:
https://help.smugmug.com/prepare-videos-for-upload-HycblvkVHG
that:
"Note: We don't keep a copy of the original video you upload; we only store and make high-quality display copies."
Does this mean that ALL video files are ALWAYS modified when they are uploaded? This of course would mean we shouldn't use md5 comparison for videos. Since comparing videos is complex, we may have to settle on rudimentary techniques like comparing filename and size.
Regarding images, I believe that Smugmug does NOT make ANY changes to the uploaded files, EVER, obviously unless the user makes some change to the file using the online editor, like rotating it, etc.. Is this a true statement? If not, what would be the circumstances under which the files would get modified? If SmugMug isn't changing the image data in the file, are they changing the metadata (exit, etc)?
If SmugMug is modifying the files, then we'll have to use an image comparison library, and things get more complicated.
I am working on some improvements to an open-source Smugmug backup application which would rely on comparing the md5 of a local file to the md5 of the corresponding remote file (on SmugMug).
https://github.com/tommyblue/smugmug-backup/issues/21
MD5 comparison is a good general way to compare binary "blobs" and jpg and mp4 are binary blobs. This approach to measuring "equality" would only be valid if Smugmug stores the same file, unaltered. So, I'm simply wanting to know what types of files, and under what circumstances, would Smugmug do this (modify an uploaded file)?
I read here:
https://help.smugmug.com/prepare-videos-for-upload-HycblvkVHG
that:
"Note: We don't keep a copy of the original video you upload; we only store and make high-quality display copies."
Does this mean that ALL video files are ALWAYS modified when they are uploaded? This of course would mean we shouldn't use md5 comparison for videos. Since comparing videos is complex, we may have to settle on rudimentary techniques like comparing filename and size.
Regarding images, I believe that Smugmug does NOT make ANY changes to the uploaded files, EVER, obviously unless the user makes some change to the file using the online editor, like rotating it, etc.. Is this a true statement? If not, what would be the circumstances under which the files would get modified? If SmugMug isn't changing the image data in the file, are they changing the metadata (exit, etc)?
If SmugMug is modifying the files, then we'll have to use an image comparison library, and things get more complicated.
0
Comments
> Thanks for getting in touch. You're correct for both video and photos. Videos are always converted on upload and we do not keep the original that was added to the uploader after the conversion is complete. Photos are not altered during the upload process. We store the original in the same state that it enters the uploader.
> SmugMug account owners do have access to some editing tools in the product and using those tools does indeed change the original file.
> Link
> We're always happy to help. Reach out with any other SmugMug questions that you have.
> We convert the video codec to h264. We do not change the file name, but we do not retain any of the original video metadata.
This implies that the filename would be the only way for me to identify videos in my application, without resorting to complex video analysis.
> Unfortunately storage information is not currently available. The team is aware of this limitation and working toward a resolution, however there is not an eta as of now.
This I was asking bc it's useful to know before running the backup app, so the destination filesystem can be properly sized.