A script to create albums from a Google Photos Takeout in PhotoPrism
PhotoPrism does not yet support transferring albums from Google Photos. Once a library has been fully transferred, this script will scrape the necessary data from a Google Takeout of an album and then, using the PhotoPrism API, create a new PhotoPrism album and populate it with the matching photos.
Note: You must import or index the photos from your Google Photos Takeout before running this script! This script does not upload any photos for you.
- Download the desired albums, or your whole collection, via Google Takeout.
- Add your Takeout photos or the original photos to PhotoPrism's import or originals directory.
- Import or index your files.
- Run the script in the Takeout directory alongside the albums and respond to any prompts.
Note: If photos were uploaded from an original source in step 2 (not Google Takeout),
use --match name
as described below. PhotoPrism's sidecar yml files will be required.
Note: This script relies on metadata.json files created by Google Takeout. Depending on your language settings from Google Photos those json files might be named differently (e.g. "Metadaten.json"). In this case just rename the files to "metadata.json" before running the script.
Pipe the output to tee to watch it run and save the output to a file:
./transfer-album.sh [options] | tee output.log
Then, use grep to list the photos that couldn't be matched:
grep WARN output.log > failures.log
Usage: transfer-album.sh <options>
-a, --import-all Import all photo albums (default)
-n, --album-name [name] Specify a single album name to import
-d, --takeout-dir [dir] Specify an alternate Google Takeout directory
Defaults to the current working directory
-s, --sidecar-dir [dir] Specify the sidecar directory (name matching only)
-m, --match [option] Set the method used to match/identify photos
Valid options: hash/name - Default matching: hash
-b, --batching [option] Set to true/false to enable/disable batch submitting
to the API. When false, photos are submitted
the the API one at a time. (default: true)
-c, --config [file] Specify an optional configuration file
-h, --help Display this help
If --album-name
is not specified, all albums will be imported.
- Google Photos automatically generates some albums (ie, hangouts albums, albums by year, or albums by date).
These will be skipped. Force import of one of these albums by specifying it with
--album-name
.
If --takeout-dir
is not specified, it will use the current working directory.
- In this case, the script can be run either in a single album directory, or in the base Google Photos
directory. If it finds a
metadata.json
in the given directory it will import it as a single album, otherwise it will try to import each subdirectory as an album.
If --match
is not specified, hash mode will be used.
- Use hash matching if you've uploaded photos from your Google Takeout and the files in PhotoPrism and Google Photos are identical. This method is significantly faster.
- Use name matching if you've uploaded original photos from another source and you just want to re-create your google Photos albums in PhotoPrism. PhotoPrism's sidecar directory, or a copy containing the yml files, must be available to the script.
If --batching
is not specified, it defaults to true.
Disabling batching will submit photos to the API one at a time as matches are found.
The script will prompt interactively for all the information it needs.
An optional config file can be specified with --config
to define
the site URL, username, and password.
An example config.conf
:
API_USERNAME=admin
API_PASSWORD=your really good password
SITE_URL=https://photos.example.com
- For each album, it creates a new PhotoPrism album with the title and description from
the album's
metadata.json
.
Then, in hash matching mode
- It scans all files in the album's directory, hashing any non-JSON files.
- It looks up each file in the database by its hash using PhotoPrism's files API.
- If it finds a match, it adds the photo's UID to the album's batching list.
- When all files are processed or every time it has gathered 999 files, an API request is sent to the server to add the gathered photos to the album.
Or, in name matching mode
- It scans the json files in the Google Takeout directory, pulling out the title field.
- It scans the yml files in the PhotoPrism sidecar directory, attempting to find a matching filename.
- If it finds a match, it pulls the photo's UID from the yml file and adds it to the album's batching list.
- When all files are processed or every time it has gathered 999 files, an API request is sent to the server to add the gathered photos to the album.