tskisner / photosort

Convert a directory tree of random images and sort by date and/or exif information.

photosort

Convert a directory tree of random images and sort by date and/or exif information.

Motivation

The purpose of this tool is to take directories of badly organized photos (think iPhoto masters directory) and sort all photos based on the actual creation datestamp in the EXIF metadata. Duplicates (based on the file's checksum) are ignored. The resulting output is a directory tree of photos sorted by year / month / day.

Details

When parsing the creation date, EXIF information is checked. If the image is a RAW format, dcraw is used to extract the embedded jpeg and extract the EXIF tags from that. The image file name is kept, and the image is copied to a new directory tree based on the creation date.

Metadata

During execution, a small sqlite DB is created to store file checksums, etc.
This speeds up re-scanning of directories when looking for new files to import.
NO OTHER METADATA is kept. For example, iPhoto event names, albums, original directory names, etc are not copied to the output directory. If you want to tag these new pristine images, I recommend using the tagging feature of Darktable.

Dependencies and Installation

You should have a fairly recent version of python3 installed. Additionally, you will need some external command-line tools:

1. exiftool (REQUIRED)
2. dcraw (if using RAW images)
3. ffmpeg (if converting old videos)
4. ImageMagick (if exporting images to reduced resolutions)

Usage

There are several small commandline scripts provided.

Sorting

Decide where you would like to output the sorted files, then run (for example):

$> phts_sync --indir /path/to/crazy/photos --outdir ~/happy/sorted/photos

And then wait for a while. If the photosync.db file in the output directory ever gets corrupted, just use the --reindex option to phts_sync to regenerate it. You can run the command above whenever there are new photos added to the input directory, or run it multiple times on different input directories and the same output directory. Only non-duplicates will be copied to the output directory.

Verify

When indexing, files are checksummed and a short version of that hash is put into the filename and the database. This makes it possible to verify that the contents of the image files have not changed. This should detect some kinds of disk corruption where the file size and access times are correct, but some bytes of the file have been changed:

$> phts_verify --outdir ~/happy/sorted/photos

Albums

Albums are handy. I typically use symbolic links for these, and there is a useful script included here for this purpose (phts_album). Usually I specify the album directory options and then for the file names I select the files I want using a file browser and drag them to my terminal window. For many Linux desktop environments this copies the paths to the files into the terminal, and then simply press enter.

FAQ

• But I want to save my hard work invested in tagging my photos using a proprietary software package. Can't you do something?

Sadly, no. This is the eternal false promise of fancy closed-source applications. They make some operations easier, but you lose control over your own data.

• Can't you add feature XXXX?

Why don't you implement it and send me a pull request?