An overview of Soulflyer image management tools
These tools are designed to help speed up photo management and enable a photographer to locate images quickly and easily. The key to this is fast and effective keywording. None of the current photo processing packages seems to do this quite right, although some are better than others. The aim here is to allow commonly used keywords to be applied with only 1 or 2 keystrokes per photo and less commonly used ones with only a few more by making good use of autocompletion and keyboard shortcuts. Finding photos should be similarly easy.
With a little preparation these tools, focused around the hinh-anh photo manager, will do that.
Everyone is different and there is no garuantee that these tools will play nicely with your own workflow. However they are designed to be used as a completely separate step after processing is finished so it should be ok. It makes no difference what tools you use to do the processing, and ratings and keywords will be picked up along with exif and IPTC data if you happen to apply them as part of that processing. As yet there is no provision to write keywords and captions etc back to the master copies but it is on the TODO list.
My workflow goes something like this:
- Photo processing using whatever program is best for the shoot. Usually this is the excellent darktable. Then export to ~/Pictures/fullsize/yyyy/mm/project
- Add watermark info using watermarker
- Export various sizes of pictures, with watermark, using the versions script. This automatically calls save-meta to add the picture details to the database for hinh-anh and other tools.
- Add keywords and refine ratings using hinh-anh
- Use hinh-anh to export a JSON file containing a list of selected pictures for a particular customer. This JSON file is read by an angular script on the soulflyer web site which generates a web page for the customer. The JSON export also triggers the creation of a zip archive of all the selected photos. AngularCustomerPhotos can be found here.
- run
backup-to-soulflyerwhich uses rsync to send the newly exported pics, json files and zip files to the web host machine.
The processed master copies need to be accessible from a single path containing a year/month/project directory structure like this:
If this doesn't match your organisation it should be easy enough to use sym links to create this structure. The year month project thing is a core assumption though and things are pretty much garuanteed to break if the file system doesn't conform to it.
It is possible to work with just this one directory of pictures but it is a little slow when using the keywording tool on a large set of photos. My pictures are all exported in various sizes and these can be used instead when they are available. They too must have the year-month-project structure but it is possible to specify where those trees live. It's also possible to have as many or as few as required. I currently use 4 directories
- thumbs fits in 160px
- small fits in 512px
- medium fits in 1024px
- large fits in 2048px
Normally I just use the medium directory for keywording purposes.
Creates the various sizes of photos for publication and use on the web. It adds the watermark, setting the position and strength according to the values saved in special instructions if they are present. Also triggers save-meta to add the photo details to the database. This is available as part of the publish repository available here. It relies on convert which is part of Imagemagick so you will need to install that and ensure it is on your path.
This is a mac only program that adds the watermark position and strength information to the image details in the database. It's available here. This functionality will eventually be added to hinh-anh so it is available to linux users too. It is possible to set the watermark information in SpecialInstructions by hand, but without the visualisation provided by watermarker this can be a bit fiddly. The format looks like this:
BL14S12X02Y02
This means put the watermark in the bottom left, opacity 14, size 12, x-offset 2, y-offset 2
The watermark is found by the add-watermark script called by versions, and can be set as an option there. TODO:Move add-watermark out of the watermarker repository.
This is used to rate, tag and search photos. Available here.
Superseded by hinh-anh, but still working. Keywords is a seesaw app that can be used to explore keywords and open the associated pics. Available here
This is the heart of image-search and hinh-anh, but provided as a library for use in other situations.
This reads the metadata from an image or a directory of images and stores it in the database. It can be called from the command line or from a script. This is available here.
It is called by the script that creates watermarked versions of the photos, but could also be run as part of the export from whatever processing program is used.
Checks the contents of one of the picture directories against the database and reports on any missing images, possible due to forgetting to export them. Available here
This can be called from the command line to perform simple searches. It can also be run in a clojure repl to do more complex searches. Either way can give a count of results found, open them in your default viewer, or write a list of the results to a file for further processing. This is available here.
Returns a list of all the keywords that are children of a given keyword. Available here
Returns a list of all projects currently in the db. Available here