Tech stuff, API and programming

So I've synced the first 21 files from Adobe Album up to 23, tags and all.

Currently, I remember what photos I've uploaded by putting the string (for example) "album#3" in their caption.

Oh. I'll need an end delimiter too. So album#3 and album#30 won't be returned by the search.

Anyway, each time sync.rb runs, it goes through all the Album files, does a 23 search for that album id in album#id form, and then uploads the photo if it doesn't exist.

I really like this because it keeps all of the sync state on the server. No configuration/history files to worry about on my side (earlier I tried adding a column to the Album Access file but it really screwed things up--I have no idea why).

But now it looks ugly having this album#id in the caption.

So the hack is starting to bother me.

Any ideas on how else I can remember where a photo on 23 came from? Maybe something like a path field, e.g. where it originally was on my harddrive, then I could search on that.

I dunno. I think the answer is that I should just do client-side sync state tracking. But I'd rather not. So I'm throwing this out there as a challenge--good ideas would be appreciated.

Thanks.

Erm...

maybe I'm missing something, but why don't you simply use tags? That's what they are for, after all: a simple way to create metadata. The caption isn't really a useful place for that.

E.g., use tags like "from:flickr", "from:adobe-album", "from:powershot-a70", etc.

Hm...that sounds intriguing. I haven't played much with EXIF, but I like what you're saying.

If I understand it:

- Before uploading each file, make sure its EXIF has a field set like albumId=2
- Search 23 for any photo with albumId=2 in its EXIF
- If it's there, don't upload
- If it's not there, it's new, do an upload.

Just one question: is there an easy way to search the EXIF data? E.g. according to the flickr docs, flickr.photos.search doesn't cover EXIF.

Thanks.

Oh yeah, forgot to show you how to get the photo id. This will do the trick:

http://www.23hq.com/services/rest/?method=test.login&api_key=ignore&username=steffen&password=(password)


/Stc

I think I've got something that's working well.

The EXIF recommendation led me down the right path.

It made me realize each picture has a date taken that is semi-unique down to the second. That is good enough for me.

So, I still have a local cache (backed by ActiveRecord and SQLite this time, so hopefully it will be a bit more robust, but then also a way to re-populate the cache by using photo.search(min_taken_date=d,max_taken_date=d) which should return just the one picture if it's already uploaded.

So, I was able to re-populate the cache with the first 300 images I uploaded and not have to go through the pain of deleting/re-uploading them.

And now the script is starting to upload the rest.

We'll see how it goes.

Thanks for the EXIF tip. Using the date taken is saving me a lot of time.

Very well. The SQLite cache is fast and robust. And re-building it based on max_taken_date/min_taken_date worked quite well too. A bit slow, but not nearly as slow as re-uploading 300 photos.

So now I just need to put it up somewhere. I made some changes to the flickr.rb and sent a diff back to its author. But I haven't done much else--this weekend I'll try to throw it up somewhere on the web.

Granted, its command line, no UI, no installer, requires Ruby, etc., so it may not be popular with non-techies.

Also, thanks a lot for the plus account--so far I've been very pleased with 23. I like the UI better than flickr. I'm not much interested in hanging out in some sort of "cool photo community" that flickr supposedly is. I just want an easy place for friends and relatives to come see photos, and I think 23 fits this quite well.

Finally got around to packaging my Adobe Album -> 23 script up and uploading it.

Binary/source/etc. is at:

http://stephen.exigencecorp.com

Comments/patches/etc. are appreciated.

Thanks,
Stephen