I'm excited to see Uploadr, which shares a name with Aaron Swartz's uploadr.py and does much the same function (syncs a bunch of files from my local computer to Flickr).
It's great that Uploadr does local duplicate detection (do not push the same file twice) and it's great that it automatically creates albums based on the folder name or iPhoto event name that contained the imported pictures. This is good. Super excited that I can pause and resume sync of folders to Flickr; this is way better than the old iPhoto connector.
But there are some problems:
* It looks like Flickr Uploadr, unlike uploadr.py, does not preserve the original filename in a metadata field. It's super hard for me to trivially prove that file DSC_9470.JPG or whatever got uploaded correctly because I can't just search for it.
* I don't love that Flickr Uploadr tries to auto-import stuff from my Desktop; I've got a bunch of screenshots there that were interesting at the time but are not photos I'm trying to preserve.
* Flickr Uploadr does not really do duplicate suppression, although its messaging suggests otherwise. If I run this program from my two laptops which both have some subset of my photos, it will push all the files from my local computer (but the same file no more than once) to the remote location. I have already carefully imported 5 of 11 albums exactly once into Flickr, and Flickr Uploadr is now carefully pushing all the photos from all 11 albums again (but is at least doing it exactly once!).
I have some utility code for de-duplication which hits the Flickr API to (1) download each photo that does not yet have a "checksum" tag, (2) compute its checksum, (3) set that checksum metadata. Then, in a second step, I have some utility code that (1) grabs each photo's checksum tag and (2) deletes all but the oldest version of each "duplicate" photo (same checksum tag).
I was hoping not to have to run this code ever again; I was thinking it was great that Flickr had finally started doing this work in the back end, and I figured that the 2-3 minutes of startup when Flickr Uploadr started doing work was time it was spending downloading checksums to compare against files on my disk. Evidently they have not done this although they have auto-added tags like "food" and "indoor" to my photos…
I'm now looking at files "sorted by upload time, newest" and seeing lots of exact dupes (fairly clear if I then "sort by date taken" and scroll down to the recently uploaded areas) and it looks like Uploadr does not actually do the duplicate suppression I was hoping for.
So Uploadr is currently busy pushing thousands of exact-duplicate copies of photos I already have in Flickr, which differ only in that they lack a checksum tag and all have title "Untitled" vs. the filename-as-title I got with uploadr.py.
I'll have a bunch of junk to clean up, which I am used to doing…
The messaging that Uploadr gives, which gave me a lot of hope, is like
""Of the 10844 photos found, 2643 are new and ready to upload, 4110 are duplicates, 3987 have already been uploaded to Flickr, and 104 are either too small or in the wrong format to upload with this utility.""
Also, it looks like this is currently a very popular tool! I got to see this adorable "fail panda" when I use the Flickr Web UI to click on photos and try to see the photo page: http://i.imgur.com/YomGPfq.jpg
It's great that Uploadr does local duplicate detection (do not push the same file twice) and it's great that it automatically creates albums based on the folder name or iPhoto event name that contained the imported pictures. This is good. Super excited that I can pause and resume sync of folders to Flickr; this is way better than the old iPhoto connector.
But there are some problems:
* It looks like Flickr Uploadr, unlike uploadr.py, does not preserve the original filename in a metadata field. It's super hard for me to trivially prove that file DSC_9470.JPG or whatever got uploaded correctly because I can't just search for it.
* I don't love that Flickr Uploadr tries to auto-import stuff from my Desktop; I've got a bunch of screenshots there that were interesting at the time but are not photos I'm trying to preserve.
* Flickr Uploadr does not really do duplicate suppression, although its messaging suggests otherwise. If I run this program from my two laptops which both have some subset of my photos, it will push all the files from my local computer (but the same file no more than once) to the remote location. I have already carefully imported 5 of 11 albums exactly once into Flickr, and Flickr Uploadr is now carefully pushing all the photos from all 11 albums again (but is at least doing it exactly once!).
I have some utility code for de-duplication which hits the Flickr API to (1) download each photo that does not yet have a "checksum" tag, (2) compute its checksum, (3) set that checksum metadata. Then, in a second step, I have some utility code that (1) grabs each photo's checksum tag and (2) deletes all but the oldest version of each "duplicate" photo (same checksum tag).
I was hoping not to have to run this code ever again; I was thinking it was great that Flickr had finally started doing this work in the back end, and I figured that the 2-3 minutes of startup when Flickr Uploadr started doing work was time it was spending downloading checksums to compare against files on my disk. Evidently they have not done this although they have auto-added tags like "food" and "indoor" to my photos…
I'm now looking at files "sorted by upload time, newest" and seeing lots of exact dupes (fairly clear if I then "sort by date taken" and scroll down to the recently uploaded areas) and it looks like Uploadr does not actually do the duplicate suppression I was hoping for.
So Uploadr is currently busy pushing thousands of exact-duplicate copies of photos I already have in Flickr, which differ only in that they lack a checksum tag and all have title "Untitled" vs. the filename-as-title I got with uploadr.py.
I'll have a bunch of junk to clean up, which I am used to doing…
The messaging that Uploadr gives, which gave me a lot of hope, is like
""Of the 10844 photos found, 2643 are new and ready to upload, 4110 are duplicates, 3987 have already been uploaded to Flickr, and 104 are either too small or in the wrong format to upload with this utility.""
Also, it looks like this is currently a very popular tool! I got to see this adorable "fail panda" when I use the Flickr Web UI to click on photos and try to see the photo page: http://i.imgur.com/YomGPfq.jpg