Repo for all AS Python Related work
Clone or download
pobocks Release mostly covering some weirdness with ASpace return types
Contains fixes/workarounds for:

  - /users route not having permissions
  - search routes which DON'T return raw Solr still return weird
  wrapped objects
  - A number of places that don't use wrap_json_object discovered
  while fixing previous issues
Latest commit 37f35e6 Sep 16, 2018


A client library for working with the ArchivesSpace API


As institutions have adopted ArchivesSpace, a variety of practitioners and institutions have written scripts making use of the backend API to accomplish various bulk tasks not supported (yet) by the interface. ArchivesSnake is intended to be a comprehensive client library, to reduce duplication of effort and simplify scripting ArchivesSpace.

Detailed API Doc

When you've read through this, please check out the detailed API Docs. The most important classes to understand are:

  • asnake.aspace.ASpace
  • asnake.client.ASnakeClient
  • asnake.jsonmodel.JSONModelObject (and its subclasses ComponentObject and TreeNode)
  • asnake.jsonmodel.JSONModelRelation

Scripts and Projects using ASnake

Here are some examples scripts and projects that make use of ASnake

For more examples on working with ASnake, please check the Wiki page.

Other API Scripts

Here are listed several examples of non-ASnake scripts that operate on ArchivesSpace - please feel free to submit your own via pull request!


ArchivesSnake has the following requirements.

  • Python 3.4 or higher
  • ability to install packages via pip (Pipenv is recommended for development)


ArchivesSnake is available from pypi:

pip3 install ArchivesSnake

If you want to install from git:

git clone
cd ArchivesSnake
python3 sdist
pip3 install dist/ArchivesSnake-0.2.0.tar.gz

This is assuming a standard Python 3 install, which provides pip3 and python3 commands. If your environment doesn't let you successfully run either command, please consult the documentation for your version of Python and/or your operating system.

You'll need an internet connection to fetch ASnake's dependencies.


Low level API

The low level API allows full access to the ArchivesSpace API; it's essentially "what if requests knew enough about an ASpace instance to manage authorization, turn uris into full URLs, and handle paged resources.

To start, here's a simple, fairly complete example - fetching the JSON representation of all the repositories from an ArchivesSpace instance and saving it to a variable.

from asnake.client import ASnakeClient

client = ASnakeClient(baseurl="",
repos = client.get("repositories").json()

# do what thou wilt with some repos

The return values from these methods are raw requests.models.Response objects - you have to call .json() on them to get the actual JSON object.

There's also a get_paged method, which handles index methods (repositories, repositories/:id/resources, etc) and returns JSON for each object in the collection.

for repo in client.get_paged('repositories'):

The ASnakeClient class is a convenience wrapper over the requests module. The additional functionality it provides is:

  • handles configuration,
  • handles and persists authorization across multiple requests
  • prepends a baseurl to API paths.

The latter means that this:


is equivalent to:


In addition to saving typing, the result of this is that the url fragments used as identifiers in ArchivesSpace ref objects can often (always?) be passed directly to these methods, e.g.:

uri = client.get('repositories/2').json()['agent_representation']['ref']
client.get(uri) # gets the agent!

Abstraction Layer

The other way to use ASnake right now is a higher level, more convenient abstraction over the whole API. It lets you ignore some of the low level details of the API, though you still need to know its structure.

There are three base classes involved: an ASpace class that represents the instance of ArchivesSpace you're connecting to, a JSONModelObject class that represents individual objects, and a JSONModelRelation class that represents routes that return groups of objects. Both JSONModel classes have subtypes for representing various exceptional cases in the API.

To use it, import the asnake.aspace.ASpace class.


JSONModelObjects wrap a single ASpace JSONModel object. Method calls on JSONModelObjects will return either the value stored in the object's JSON representation, or will try to make a call to the API to fetch a subsidiary route.

So, for a JSONModelObject named obj wrapping this JSON:

    "jsonmodel_type": "repository",
    "uri": "/repositories/2",
    "name": "International Repository of Pancakes",
} would return "International Repository of Pancakes", and obj.resources would return a JSONModelRelation of the route /repositories/2/resources


JSONModelObjects representing resource or classification trees or nodes within those trees have specialized representation; specifically, they support two specialized properties:

a_tree.record # this returns the JSONModel object pointed to by that tree or node

a_tree.walk # this returns a generator that returns the record, followed by all records in the tree below it

# Usage example: printing a resource and all its subsidiary objects
for record in a_tree.walk:


JSONModelRelation objects "wrap" an API route representing either a collection of objects or an intermediate route (a route such as /agents that has child routes but no direct results. A JSONModelRelation can be iterated over like a list, like so:

for repo in aspace.repositories:
    # do stuff with repo which is a JSONModelObject

You can get the wrapped JSON by doing:


If you know the id of a particular thing in the collection, you can also treat JSONModelRelation objects as functions and pass the ids to get that specific thing, like so.

aspace.repositories(101) # repository with id 101

If you need to pass parameters to a route, you can add them using the with_params method; here's an example using the /repositories/:repo_id/search route to find published resources within a repository:

repo = aspace.repositories(101)
for resource in"primary_type:resource", fq="publish:true"):
    # do things with published resources from repo 101

A short full example using ASnake to print the title for all finding aids in ArchivesSpace, for example:

from asnake.aspace import ASpace

aspace = ASpace()

for repo in aspace.repositories:
    for resource in repo.resources:

Currently, the ASpace interface is effectively read-only; if you need to create or update records (or just do something we haven't implemented yet), you'll have to drop down to the low-level interface - for convenience, the ASnakeClient used by an ASpace object is accessible like so:


For example, if you were really excited about archival data, and wanted to add an interrobang (‽) to the end of every resource's title, you'd do:

for repo in aspace.repositories:
    for resource in repo.resources:
        res_json = resource.json()
        res_json['title'] = res_json['title'] + '', json=res_json)


As per the example above, you can configure the client object by passing it arguments during creation.

Allowed configuration values are:

Setting Description Default Value
baseurl The location (including port if not on port 80) of your archivesspace backend http://localhost:4567
username Username for authorization admin
password Password for authorization admin
retry_with_auth Whether to respond to 403 errors by trying to authorize and retrying True
logging_config Hash with various config values for the logging subsystem see below

The logging config allows the following settings, none of which are present by default:

Setting Description Notes
default_config A default configuration to start from, see Logging for more info
stream stream to be printed to (e.g. sys.stdin, sys.stdout, an open file) cannot be combined with stream
filename name of file to be printed to cannot be combined with filename
filemode mode to apply to file, as per open ('w' for write, 'a' for append, etc) only useful combined with filename
level level to log at (e.g. 'INFO', 'DEBUG', 'WARNING')

You can also define a configuration file, formatted in the YAML markup language. By default, ASnake looks for a file called .archivessnake.yml in the home directory of the user running it. If an environment variable ASNAKE_CONFIG_FILE is set, ASnake will treat it as a filename and search there.

An example configuration file:

baseurl: http://localhost:4567
username: admin
password: admin
retry_with_auth: false
    default_config: INFO_TO_STDERR

Default values corresponding to the admin account of an unaltered local development instance of ASpace are included as fallback values.


ArchivesSnake uses structlog combined with the stdlib logging module to provide configurable logging with reasonable defaults. By default, log level is INFO, logging's default formatting is suppressed, and the log entries are formatted as line-oriented JSON and sent to standard error. As logging in ArchivesSnake is by default universally below INFO level, in general the log will be silent unless you change configuration. All of this can be configured; note that configuration must happen prior to creating an asnake.client.ASnakeClient or asnake.aspace.ASpace object

from asnake.client import ASnakeClient
import asnake.logging as logging

logging.setup_logging(level='DEBUG') # logging takes several arguments, provides defaults, etc

# NOW it is safe to initialize any ASnake stuff
client = ASnakeClient()

There are a number of provided configurations, available in dict asnake.logging.configurations and exposed as toplevel constants in the module (e.g. asnake.logging.DEBUG_TO_STDERR, asnake.logging.DEFAULT_CONFIG). Log level and the stream/filename to be printed to can be overriden by passing level and either the stream or filename arguments to setup_logging. Mode of a file can be controlled by passing filemode.

For example:

logging.setup_logging(filename="my_precious.log", filemode="a") # write to my_precious.log, appending if it already exists
logging.setup_logging(stream=sys.stdout, level="DEBUG") # log to stdout, showing all log entry levels

The provided configurations are:

Configuration Names Level Output To Notes
INFO_TO_FILE INFO ~/archivessnake.log
DEBUG_TO_FILE DEBUG ~/archivessnake.log

By setting the ASNAKE_LOG_CONFIG environment variable to one of these names, you can set that config as the default.

To directly get ahold of a logger for use in your own application, you can call asnake.logging.get_logger. An example of using this to print your own logs to a file:

import asnake.logging as logging
logger = logging.get_logger("my_script_log")

# do stuff"my_event_name", property1="a property", anything={"json": "serializable"})
# do more stuff

This will leave the following in my_cool_logfile.log (pretty-printed below, but all on one line in practice):

{ "property1": "a property",
  "anything": {"json": "serializable"},
  "event": "my_event_name",
  "logger": "my_script_log",
  "level": "info",
  "timestamp": "2018-07-18T00:06:49.636482Z"


Documentation is generated using Sphinx with the Read the Docs Theme, and is available here