Open-source Python library that allows developers to leverage the nth.community service platform and API to implement secure, privacy-preserving data collaborations within their web services and applications.
This library is a client-side component and Python API for the nth.community secure, privacy-preserving data collaboration service platform. Together, this library and the nth.community service platform make it possible to define and execute data workflows (called collaborations) that operate on encrypted input data without decrypting it.
This open-source library supports a very limited set of input data types (non-negative 32-bit integers and single-column tables of strings) and operations (intersection of tables, row count of an intersection of tables, and summation of integers). The secure data collaboration workflows enabled by nth.community protect contributor inputs by relying on secure multi-party computation protocols, including private set intersection (via the oblivious library) and additive secret sharing (via the additive library).
Designed to be integrated easily into full-stack web applications, this library organizes secure multi-party computation workflows into a familiar structure that resembles public-key cryptographic systems. In order to improve portability of data and to allow programmers to leverage native Python features, data structures in the library are derived from built-in Python types and can be converted in a straightforward way to and from ubiquitous formats such as JSON.
The package is available on PyPI:
python -m pip install nthcommunity
The library can be imported in the usual ways:
import nthcommunity from nthcommunity import *
A secure, privacy-preserving data collaboration begins when a recipient party is created. This party is responsible for defining the collaboration data workflow, distributing contribution keys to contributors, and performing the computation on the encrypted data contributions in concert with the nth.community service platform:
>>> from nthcommunity import * >>> r = recipient()
Individual contributor objects encapsulate individual contributors. Two contributors are defined below:
>>> c_a = contributor() >>> c_b = contributor()
Each contributor is assigned a unique identifier when the contributor object is created:
>>> c_a.identifier() '57728767-55bf-4833-ab36-d3733a0e8448'
A collaboration is a tree-like data structure that defines how the overall result of a collaborative data workflow is computed. Every leaf node in the data structure must indicate which contributor must supply the encrypted data corresponding to that node. The internal, non-leaf nodes represent data operations such as count, intersection, and summation. The collaboration workflow defined below computes the size of the intersection between two tables (one from each contributor):
>>> w = count(intersection(table(contributor=c_a), table(contributor=c_b)))
The recipient can (by leveraging the nth.community service platform API) generate a dictionary that maps each of the contributor identifiers to their respective contribution key:
>>> id_to_key = r.generate(w)
Each of these individual keys id_to_key[c_a.identifier()] and id_to_key[c_b.identifier()] can be delivered to their corresponding contributor. Note that the collaboration class is derived from dict, making conversion to JSON straightforward (using either the built-in json library or the wrapper method used below):
>>> print(id_to_key[c_a.identifier()].to_json(indent=2))
{
"type": "operation",
"value": "count",
"arguments": [
{
"type": "operation",
"value": "intersection",
"arguments": [
{
"type": "table",
"contributor": {
"type": "contributor",
"identifier": "6a8dd844-6003-4475-aea7-b4182400bb90"
},
"public": "tdeJzrPIEpsFcykqnlN4M/hiP8gnNAJiSBRysLIutwo="
}
]
}
],
"material": {
"public": "F/RIe6jQ38Uk1CUdW7JKwd7Q9b+J+HlnOARSM70zERg=",
"scalar": "A4f8MRxKsxZnKiScaJPw/O6uPzqBfREeNaPAaOdlnAU="
},
"certificate": "U4HFZcQ1hCplmec50gEDYnMxHn/ILclPmR6KC04uz1i5sJBwpEq0HB+tMlRPzWZu6dpPNuAjOzlIOshSgw7EXA=="
}
Each contributor can then encrypt their data contribution using their key:
>>> table_a = [['a'], ['b'], ['c'], ['d']] >>> enc_a = c_a.encrypt(id_to_key[c_a.identifier()], table_a) >>> table_b = [['b'], ['c'], ['d'], ['e']] >>> enc_b = c_b.encrypt(id_to_key[c_b.identifier()], table_b)
As with contributor keys, encrypted contributions can be converted to and from JSON in a straightforward way:
>>> payload = enc_a.to_json(indent=2)
>>> print(payload)
{
"type": "operation",
"value": "count",
"arguments": [
{
"type": "operation",
"value": "intersection",
"arguments": [
{
"type": "table",
"contributor": {
"type": "contributor",
"identifier": "6a8dd844-6003-4475-aea7-b4182400bb90"
},
"public": "tdeJzrPIEpsFcykqnlN4M/hiP8gnNAJiSBRysLIutwo=",
"value": [
[
"0Ch4F9+RVGu+e0dmK+ydFvOZSLld/xMk4xcXX8U2TGY5/CUiwCtPb0bM6nwJrbZPVBIAWa5tgVCGR2Tu8LdfalxDwpGqyNuWPcUtQP8T6P8=",
"F1CZ5kRyYpEoZhlwTLu1gShLsH1HDZxQnQnypgDncyG+HGCrHn+UtKb9B4bvDaxbb8Nl92TV+U5ouE8MePbnFkslfwZyD6g5hWh4SOv1UhWl4uOuW4fSX8rnLsVg5VKc2Q=="
],
[
"4noLjc9aSpY1owBlSHv9IYZr6bvxkQUhhKOhV9Ty2kZTJP7UbXgSyxjsqEr1qbEWXQ1vP1uWS+JMN+rWqtpblcitqcg/epuPYWvJi6vr9vE=",
"JC1L0t4JeVDr/+40c/JAsnpTqQUZLELZ5hJukDlX1Uxxf5MqZXglNc1aR9p+A5fvIZPj1P3YHOMS1Cbjpy1R3iZNUMsTOkn7pjhkZdkGE+XzHVRKz9TQdmkQZ2hk+13INQ=="
],
[
"TmQcw0iFgnnUofuJj4hhof6oqNp1iQ94E+jqNCdQT3i8kcL4D7I33lJdLyiUwAQHm1BE6SrU+ZS+jOrfle5mfY91PuuxFsB6UBlcc8MP+Do=",
"puSkEfdRdNiIO668GG5jwV/zp84OhFN9rs5UjOoYNEUHif1BzrNDMq8huDJxFW8k78GbjMPZ/D4Yn5S3bjg/CwmGFPb99o4p7OwBYP1pGuds8smzUaVqcPt5FCCJxZWGYQ=="
],
[
"PcBEJUe6IjiJ5EKUBTfeebzlLo2s291sAEdi9UnSQ3O1yg+ttOVxQQNN2QVvc4V8Swj6J+FBbC+S3+av1yV/57yW88J8YJYBTLlK0awWYqw=",
"ubwDp7Up8bRvznbOSmKlTDJ4MlC1JIz8zFaOHQIdfiCK1LPCKeyqIOaX/BPRbmNUkGQA9raLdsgUvWH9TQzGd3c7nvmbjltL9e4WWofM72+fBdj+/Dgg16+YAdEY6fXoog=="
]
]
}
]
}
]
}
>>> collaboration.from_json(payload) is not None
True
Once all the encrypted contributions are received, the recipient can (by again leveraging the nth.community service platform API) execute the workflow on the encrypted contributions to obtain a result:
>>> result = r.evaluate({c_a.identifier(): enc_a, c_b.identifier(): enc_b})
>>> result["value"]
3
The documentation can be generated automatically from the source files using Sphinx:
cd docs python -m pip install -r requirements.txt sphinx-apidoc -f -E --templatedir=_templates -o _source .. ../setup.py && make html
All unit tests are executed and their coverage is measured when using pytest (see setup.cfg for configuration details):
python -m pip install pytest pytest-cov python -m pytest
Style conventions are enforced using Pylint:
python -m pip install pylint python -m pylint nthcommunity test/test_nthcommunity.py
In order to contribute to the source code, open an issue or submit a pull request on the GitHub page for this library.
Beginning with version 0.1.0, the version number format for this library and the changes to the library associated with version number increments conform with Semantic Versioning 2.0.0.
This library can be published as a package on PyPI by a package maintainer. Install the wheel package, remove any old build/distribution files, and package the source into a distribution archive:
python -m pip install wheel rm -rf dist *.egg-info python setup.py sdist bdist_wheel
Next, install the twine package and upload the package distribution archive to PyPI:
python -m pip install twine python -m twine upload dist/*