Anchorage consists of a Python library and CLI to save your bookmark collection in bulk, forever: online in the Internet Archive or locally, using ArchiveBox.
Anchorage will automatically retrieve your bookmark collection from your browser of choice, filter out duplicates, local files as well as entries matching filters of your own making, and archive the chosen ones.
Read on to get started. The full Python API documentation is available here.
As the internet ages link rot takes over larger and larger swathes of it, from the tiny to the mighty, from the trivial to the best pieces you ever found: all lost forever. Anchorage is an attempt to make it as easy as possible for you to save the little corner of it you're most fond of, for your own peace of mind and the enjoyment of us all :)
A working Docker install is the only requirement, beyond Python and Anchorage's dependencies. Without Docker: Docker is used to run ArchiveBox, via a provided docker-compose file. Without Docker Anchorage will not be able to archive your collection locally, but it will still be able to save it online in the Internet Archive.
Anchorage can be installed using pip as any Python package. Its dependencies will be downloaded automatically.
pip install anchorage
To access a browser's bookmarks file, Anchorage stores its location in its configuration file:
~/.anchorage/config.toml
There's an example config.toml
in this repo for reference.
To add a new browser simply add a new top-level key, followed by its bookmark file paths. Anchorage only needs the path in your operating system to work.
[<browser name>]
linux = <path>
macos = <path>
windows = <path>
Importantly:
- Linux and MacOS paths are stored in full.
- Windows paths are stored from the
AppDatadirectory.
The default config.toml contains the bookmark file paths for Google Chrome, Mozilla Firefox and Microsoft Edge
and Edge Beta for Windows only. To use Anchorage in Linux or MacOS add the bookmark file path of your
browser of choice to your config.toml.
The config file can be edited just as any other. New browsers will automatically be listed in the CLI.
Importantly:
- Set unknown bookmark file paths to "?". That way the CLI will recognize those as unknown and behave appropriately.
The CLI will guide you through retrieving your bookmarks from your browser of choice, applying filters to you bookmark collection and archiving your bookmarks in the Internet Archive or locally, using ArchiveBox.
To start the CLI open your shell and type
anchorage
You will be asked whether you're ready to proceed. On the ok it will ensure all dependencies are present.
If a config file is found, you will be prompted to choose whether to keep the current config or overwrite it with the default one.
You will be prompted to choose which browser to retrieve your bookmark collection from. The browser
choices are sourced from config.toml. Refer to section 3 for
editing it to add a missing browser or enter the path to the bookmarks file of your browser, if it's missing
(equal to "?").
Filters can be applied to your bookmark collection before archiving. Any or all of four filters can be chosen, one specific for URLs:
Local files: remove local URLs (say, PDFs stored in your computer) from the collection.
and three general:
Match string: remove bookmark URLs, names or bookmark directories matching a provided string or any string in a string list.Match substring: remove bookmark URLs, names or bookmark directories containing a provided string or any string in a string list.Regex: remove bookmark URLs, names or bookmark directories matching a provided regex formula.
For each you will be prompted to choose to apply it to any or all of the previous.
You will be then asked to choose whether to archive your collection online or locally.
By default websites will not be archived if a previous image exists in The Internet Archive. This is to save time: we rest easy as a those sites are saved already at some point. In case you want to save a current snapshot of the colection, you will be prompted whether to override this and archive all sites in the collection regardless. This may take significantly longer. Based on your choice, you will be given an estimate of the archive time.
To archive your collection locally you will be prompted for an archive directory.
After a last confirmation the process will begin. A progress bar will inform you of how far the process is from finishing, how many bookmarks have been saved and provide a dynamic estimate of the time remaining before the process is finished.
The full documentation of the Anchorage API is available in the docs site.
Generate the Anchorage config file with the init command.
from anchorage import init
init()
Three methods are relevant:
path(<browser>): obtain the path to your chosen browser's bookmarks file (in your OS) fromconfig.toml.load(<path>): read your chosen browser's JSON or JSONLZ4 bookmarks file and return a Python dictionary.bookmarks(<dict>): create an instance of thebookmarksclass.
The bookmarks class creates a second bookmarks dictionary more suitable for our intent, and contains methods
to filter and loop through the collection. Filters can be applied as seen below.
from anchorage import path, load, bookmarks
collection = bookmarks(load(path(<browser name>)),
drop_local_files= <boolean>,
drop_dirs= <string or list of strings>,
drop_names= <string or list of strings>,
drop_urls= <string or list of strings>,
drop_dirs_subs= <string or list of strings>,
drop_names_subs= <string or list of strings>,
drop_urls_subs= <string or list of strings>,
drop_dirs_regex= <string>,
drop_names_regex= <string>,
drop_urls_regex= <string>
)
Input: bookmarks instance or bookmark dictionary returned by load.
from anchorage import anchor_online
anchor_online(bookmarks, overwrite=<bool>)
The overwrite parameter determines whether to save snapshots of sites already present in the
Internet Archive or not.
from anchorage import anchor_locally
anchor_locally(bookmarks, archive=<dir>)
The archive parameter specifies the directory in which to create the local archive.
Running the ArchiveBox default NGINX server can be done with the following command.
from anchorage import server
server()