DataTrails Python Client
The standard DataTrails Python Client.
Please note that the canonical API for DataTrails is always the REST API documented at https://docs.datatrails.ai
Support
This package currently is tested against Python versions 3.8,3.9,3.10,3.11 and 3.12.
The current default version is 3.8 - this means that this package will not use any features specific to versions 3.9 and later.
After End of Life of a particular Python version, support is offered on a best effort basis. We may ask you to update your Python version to help solve the problem, if it cannot be reasonably resolved in your current version.
Installation
Use standard python pip utility:
python3 -m pip install datatrails-archivist
If your version of python3 is too old an error of this type or similar will be emitted:
Note
ERROR: Could not find a version that satisfies the requirement datatrails-archivist (from versions: none) ERROR: No matching distribution found for datatrails-archivist
Example
You can then use the examples code to create assets (see examples directory):
"""Create an asset in DataTrails with User Token.
The module contains two functions: main and create_asset. Main function parses in
a url to the Archivist and credentials, which is a user authorization.
The main function would initialize an archivist connection using the url and
the credentials, called "arch", then call arch.assets.create() and the asset will be created.
"""
from os import getenv
from archivist.archivist import Archivist
def create_asset(arch):
"""Create an asset using Archivist Connection.
Args:
arch: archivist connection.
Returns:
newasset: a new asset created.
"""
attrs = {
"arc_display_name": "display_name", # Asset's display name in the user interface
"arc_description": "display_description", # Asset's description in the user interface
"arc_display_type": "display_type", # Arc_display_type is a free text field
# allowing the creator of
# an asset to specify the asset
# type or class. Be careful when setting this:
# assets are grouped by type and
# sharing policies can be
# configured to share assets based on
# their arc_display_type.
# So a mistake here can result in asset data being
# under- or over-shared.
"some_custom_attribute": "value" # You can add any custom value as long as
# it does not start with arc_
}
#
# There are 3 alternatives
#
# 1. Create the asset:
# The first argument is the attributes of the asset
return arch.assets.create(attrs=attrs)
#
# 2. alternatively one can wait for the asset to be confirmed in the
# immutable store.
# The second argument is wait for confirmation:
# If @confirm@ is True then this function will not
# return until the asset is confirmed.
#
# Confirmation guarantees that 3rd parties can retrieve and cryptographically
# verify your Assets, which can take a few seconds to propagate. It is typically
# not necessary to wait unless your workflow involves near-real-time
# communication with 3rd parties and the 3rd party needs immediate cryptographic
# verification of your new Asset.
return arch.assets.create(attrs=attrs, confirm=True)
#
# 3. lastly if some work can be done whilst the asset is confirmed then this call
# can be replaced by a two-step alternative:
asset = arch.assets.create(props=props, attrs=attrs)
# ... do something else here
# and then wait for confirmation
return arch.assets.wait_for_confirmation(asset['identity']))
def main():
"""Main function of create asset.
Parse in user input of url and client id/secrets and use them to
create an example archivist connection and create an asset.
"""
# client id and client secret is obtained from the appidp endpoint - see the
# application registrations example code in examples/applications_registration.py
#
# client id is an environment variable. client_secret is stored in a file in a
# directory that has 0700 permissions. The location of this file is set in
# the client_secret_filename environment variable.
client_id = getenv("DATATRAILS_APPREG_CLIENT")
client_secret_file = getenv("DATATRAILS_APPREG_SECRET_FILENAME")
with open(client_secret_file, mode="r", encoding="utf-8") as tokenfile:
client_secret = tokenfile.read().strip()
# Initialize connection to Archivist. max_time is the time to wait for confirmation
# of an asset or event creation - the default is 300 seconds but one can optionally
# specify a different value.
with arch = Archivist(
"https://app.datatrails.ai",
(client_id, client_secret),
max_time=300,
) as arch:
# Create a new asset
asset = create_asset(arch)
print("Asset", asset)
if __name__ == "__main__":
main()
Notebooks
Some jupyter notebooks are available to exercise the examples code. These examples can be downloaded from python.datatrails.ai and run in a local install of jupyter notebook such as jupyterLabDesktop.
Please consult https://python.datatrails.ai/notebooks.html for details.
File Story Runner
You can run scenarios - a sequence of steps - from a python dictionary or from a yaml or json file.
Python
from logging import getLogger
from pyaml_env import parse_config
from sys import exit as sys_exit
from sys import stdout as sys_stdout
from archivist import about
from archivist.archivist import Archivist
from archivist.parser import common_parser, endpoint
LOGGER = getLogger(__name__)
def run(arch: Archivist, args):
LOGGER.info("Using version %s of datatrails-archivist", about.__version__)
LOGGER.info("Namespace %s", args.namespace)
with open(args.yamlfile, "r", encoding="utf-8") as y:
arch.runner(parse_config(data=y)
sys_exit(0)
def main():
parser = common_parser("Executes the archivist runner from a yaml file")
parser.add_argument(
"yamlfile", help="the yaml file describing the steps to conduct"
)
args = parser.parse_args()
arch = endpoint(args)
run(arch, args)
parser.print_help(sys_stdout)
sys_exit(1)
Command Line
This functionality is also available with the CLI tool archivist_runner, which is bundled with version v0.10 onwards of the datatrails-archivist.
You can verify the installation by running the following:
archivist_runner -h
Which will show you the available options when using archivist_runner.
To use the archivist_runner command you will need the following:
A Client ID and Client Secret by creating an App Registration
The YAML file with the operations you wish to run
The URL of your DataTrails instance, this is typically https://app.datatrails.ai
Example usage:
archivist_runner \
-u https://app.datatrails.ai \
--client-id <your-client-id> \
--client-secret <your-client-secret> \
functests/test_resources/richness_story.yaml
Example Yaml Snippet
This is an example of creating an asset and creating an event for that asset. The yaml file consists of a list of steps.
Each step consists of control parameters (specified in the 'step' dictionary) and the yaml representation of the request body for an asset or event.
The confirm: field is a control variable for the PythonSDK that ensures that the asset or event is confirmed before returning. This is optional and is only required 3rd parties need to immediately retrieve and cryptographically verify your Assets, which can take a few seconds to propagate. It is typically not necessary to wait unless your workflow involves near-real-time communication with 3rd parties and the 3rd party needs instant cryptographic verification of your new Asset.
Note
The name of the asset is important. The value of the name is carried forward for every operation - in this case the name of the asset is 'radiation bag 1'.
Arguments to the archivist are usually strings - in this example radioactive is 'true' which archivist will treat as a boolean.
---
# Demonstration of applying a Richness compliance policy to an asset that undergoes
# events that may or may not make the asset compliant or non-compliant.
#
# The operation field is a string that represents the method bound to an endpoint and
# the args and kwargs correspond to the arguments to such a method.
#
# NB the assets and events endpoints require all values to be strings. Other values may
# be of the correct type such as confirm which is a boolean.
#
steps:
# note the values to the assets.create method are string representations of boolean
# and numbers
- step:
action: ASSETS_CREATE
description: Create an empty radiation bag with id 1.
asset_label: radiation bag 1
behaviours:
- RecordEvidence
attributes:
arc_display_name: radiation bag 1
radioactive: "true"
radiation_level: "0"
weight: "0"
# setup the radiation bags to have a varing amount of radiactive waste
# note the values to the events.create method are string representations of boolean
# and numbers
- step:
action: EVENTS_CREATE
description: Create Event adding 3 rads of radiation to bag 1, increasing its weight by 1kg.
asset_label: radiation bag 1
operation: Record
behaviour: RecordEvidence
event_attributes:
arc_description: add waste to bag
arc_evidence: see attached conformance report
conformance_report: blobs/e2a1d16c-03cd-45a1-8cd0-690831df1273
asset_attributes:
radiation_level: "3"
weight: "1"
Logging
Follows the Django model as described here: https://docs.djangoproject.com/en/3.2/topics/logging/
The base logger for this package is rooted at "archivist" with subloggers for each endpoint:
Note
- archivist.archivist
sublogger for archivist submodule
- archivist.assets
sublogger for assets submodule
and for other endpoints.
Logging is configured by either defining a root logger with suitable handlers, formatters etc. or by using dictionary configuration as described here: https://docs.python.org/3/library/logging.config.html#logging-config-dictschema
A recommended minimum configuration would be:
import logging
logging.dictConfig({
"version": 1,
"disable_existing_loggers": False,
"handlers": {
"console": {
"class": "logging.StreamHandler",
},
},
"root": {
"handlers": ["console"],
"level": "INFO",
},
})
For convenience this has been encapsulated in a convenience function set_logger
which should be called before anything else:
from archivist.logger import set_logger
from archivist.archivist import Archivist
set_logger("DEBUG")
client_id = getenv("DATATRAILS_APPREG_CLIENT")
client_secret_file = getenv("DATATRAILS_APPREG_SECRET_FILENAME")
with open(client_secret_file, mode="r", encoding="utf-8") as tokenfile:
client_secret = tokenfile.read().strip()
arch = Archivist(
"https://app.datatrails.ai",
(client_id, client_secret),
max_time=300,
)
Development
For instructions on contributing to the DataTrails SDK see DEVELOPMENT.md.