Photos can come from two sources:
- Individual images that are added on a one-off basis. Screenshots, perhaps. I add them with a script by passing the image's filename. The script prompts me for some metadata like title, album/category, event.
- Iphoto. Iphoto has a nice
AlbumData.xmlfile with all the data I need. Title, album, event, rating, x/y location, etcetera. I sync this with a script that investigates the iphoto library and syncs/copies what's needed.
The one big target is a git annex repository. Two things need to end up in there:
- The individual original photo files. They'll have a nice SEO-friendly lowercase name instead of
IMG_1234.JPG-like junk. - A json file (or a couple of them) with metadata on those photo files like title, album and event.
How do I know which photos to sync from iphoto? By having a smart album that tells me which photos are ready. I use smart albums to tell me which photos still need to get a rating, a title, a geotag. So the most safe way is to use the very same mechanism to tell me which photos are ready!
This smart album filter might not be smart enough, so there's room for extra filtering in Python. It'll be extra filtering, though. I won't add photos.
Elements I use from iphoto:
- Events will end up as-is on the website. Those are a handy way to group photos, separate from albums.
- Albums (which I can nest in folders) for grouping. A photo can end up in multiple albums.
I want to be more strict in making photos private. Not all my kids' photos need to be public. And not all my work photos. Some are just for my own enjoyment. Something as simple as an IP-based permission can get me and my family access later on.
How to mark photos private? Well, I'll just add an underscore in front of the title :-)
Videos can come only from one source: they're individual files that are added individually. I make them one at a time. The ones I have now are all over the place. So a script to add them will have to do. It should ask for metadata, perhaps based on the photos' metadata.
The target?
- Individual video files in year-based folders.
- A metadata file (or several) with the title and album/event info.
What if metadata is missing? Perhaps a script that notes it and allows you to add it.
Albums are categories. But I have gone overboard with them on flickr and smugmug. But I think I can simplify it in this new setup.
- I use events, so things like "vacation 2012" are already bundled in an event.
- If I add geotags everywhere, I don't have to split my train photos in "Dutch train photos" and "German train photos. That'll be done automatically once I get around to programming it.
- I can always add tags for some extra grouping later on if I need it.
- I can tag events, too. Perhaps a nice way to group swordfighting stuff?
- Nobody's going to do much browsing through my photos. At least not through multiple levels of folders. The only thing I need to do, really, is to group it for external visitors. "Family", "trains". So that those interested in trains aren't bothered with my cycling photos.
- For special items, I can make proper textual pages in the regular part of my website.
Here's a brainstormed list: ligfiets, familie, (de)construction, trains, model trains, historical trains, work, csr, kerk, school. As an example, "work" contains zestsoftware, python meetings, PhD conferences and so on.
- Add cleanup function that filters out non-existing photos/videos and that also removes them from the albums.
- Add page generation on the server. (Perhaps generate a makefile locally that gets executed on the server... Gna gna gna.)
- Add conversion on the server.
- Add "private? [y/N]" question.
- Add push to s3 on server.