Skip to main content

Python library that serves as an API for common cryptographic primitives used to implement OPRF, OT, and PSI protocols.

Project description

Python library that serves as an API for common cryptographic primitives used to implement OPRF, OT, and PSI protocols.

PyPI version and link. Read the Docs documentation status. GitHub Actions status. Coveralls test coverage summary.


This library provides pure-Python implementations, Python wrappers for libsodium and mcl, and additional utility methods for cryptographic primitives that are often used to implement oblivious pseudorandom function (OPRF), oblivious transfer (OT), and private set intersection (PSI) protocols.

Installation and Usage

This library is available as a package on PyPI:

python -m pip install oblivious

It is possible to install the library together with packages that bundle dynamic/shared libraries, such as rbcl and/or mclbn256 (note that in some environments, the brackets and/or commas may need to be escaped):

python -m pip install oblivious[rbcl]
python -m pip install oblivious[mclbn256]
python -m pip install oblivious[rbcl,mclbn256]

The library can be imported in the usual ways:

import oblivious
from oblivious import ristretto
from oblivious import bn254


This library supports concise construction of elliptic curve points and scalars. The examples below use the ristretto module that provides data structures for working with the Ristretto group:

>>> from oblivious.ristretto import point, scalar
>>> p = point.hash('abc'.encode()) # Point derived from a hash of a string.
>>> s = scalar() # Random scalar.
>>> t = scalar.from_int(0) # Scalar corresponding to the zero residue.

Built-in Python operators are overloaded to support point operations (such as addition, subtraction, negation, and equality) and scalar operations (such as multiplication by a scalar and inversion of scalars):

>>> q = s * p
>>> p == (~s) * q
>>> p == ((~s) * s) * p
>>> p + q == q + p
>>> t * p == p - p

The point and scalar classes have common conversion methods that correspond to those supported by bytes objects (and in some cases, these classes are themselves derived from bytes):

>>> hex = '35c141f1c2c43543de9d188805a210abca3cd39a1e986304991ceded42b11709'
>>> s = scalar.from_hex(hex)
>>> s.hex()

In addition, Base64 conversion methods are included to support concise encoding and decoding of point and scalar objects:

>>> s.to_base64()
>>> s == scalar.from_base64('NcFB8cLENUPenRiIBaIQq8o805oemGMEmRzt7UKxFwk=')

For more information and background about the underlying mathematical structures and primitives found in the ristretto module, consult materials about Curve25519, the Ristretto group, and the related Ed25519 system.

Using Pure Python or a Shared/Dynamic Library

Each module within this library can export two variants of its primitives and data structures: one corresponding to pure-Python implementations and another corresponding to shared/dynamic library wrappers.

For example, the ristretto module exports two container classes/namespaces: python and sodium. These encapsulate pure-Python implementations and shared/dynamic library (i.e., libsodium) wrappers, respectively, of all operations and classes available in the ristretto module. This makes it possible to explicitly choose whether an operation requires only Python or also requires the presence of a compiled copy of libsodium on the host system.

The example below uses pure-Python implementations of the scalar multiplication operation (relying on the ge25519 library):

>>> from oblivious.ristretto import python
>>> p = python.point.hash('abc'.encode())
>>> s = python.scalar.hash('123'.encode())
>>> (s * p).to_base64()

To check whether an instance of the libsodium shared/dynamic library has been loaded successfully, the check below can be performed:

>>> from oblivious.ristretto import sodium
>>> sodium is not None # Was the dynamic/shared library loaded?

In the example below, the scalar multiplication operation invokes a binding for the crypto_scalarmult_ristretto255 function exported by libsodium:

>>> p = sodium.point.hash('abc'.encode())
>>> s = sodium.scalar.hash('123'.encode())
>>> (s * p).to_base64()

The class methods exported by the ristretto module directly (e.g., the method __add__ within the class point that is imported via the statement from oblivious.ristretto import point) correspond either (A) to libsodium wrappers if an instance of libsodium is found and loaded or (B) to pure-Python implementations if all attempts to load a working instances of libsodium fail. The ordered list below summarizes what definitions are exported under various conditions and the ordered sequence of attempts to locate and load an instance of libsodium.

  1. Under all conditions, the wrapper class python is defined and encapsulates a pure-Python variant of every low-level operation and class available in the ristretto module. As a starting default, all classes exported directly by the ristretto module correspond to the pure-Python implementations.

  2. If a shared/dynamic library instance of libsodium is found on the system and successfully loaded during one of the attempts below, then the wrapper class sodium is defined:

    1. the built-in ctypes.util.find_library function is able to locate 'sodium' or 'libsodium' and it is loaded successfully;

    2. a file or libsodium.dll under the paths specified by the PATH and LD_LIBRARY_PATH environment variables is found and loaded successfully; or

    3. the optional rbcl package is installed and the compiled subset of libsodium included in that package is loaded successfully.

  3. If sodium is not None, then the sodium class encapsulates libsodium wrappers for low-level operations and for every class exported by the ristretto module. Furthermore, those classes exported directly by the library are redefined to use the bindings available in the loaded instance of libsodium. The python class is still exported, as well, and all operations and class methods encapsulated within python remain as-is (i.e., pure-Python implementations).

The classes within the bn254 module (both those that are pure-Python implementations and those that are wrappers for functions in the mcl library) are organized in a similar manner. More information is available in the documentation for the bn254 module.


All installation and development dependencies are fully specified in pyproject.toml. The project.optional-dependencies object is used to specify optional requirements for various development tasks. This makes it possible to specify additional options (such as docs, lint, and so on) when performing installation using pip:

python -m pip install .[docs,lint]


The documentation can be generated automatically from the source files using Sphinx:

python -m pip install .[docs]
cd docs
sphinx-apidoc -f -e -E --templatedir=_templates -o _source .. && make html

Testing and Conventions

All unit tests are executed and their coverage is measured when using pytest (see the pyproject.toml file for configuration details, and note that unit tests that require rbcl and/or mclbn256 are skipped if the corresponding optional package is not installed):

python -m pip install .[test]
python -m pytest

Concise unit tests are implemented with the help of fountains; new reference specifications for the tests in a given testing module can be generated by running that testing module directly:

python test/
python test/

Style conventions are enforced using Pylint:

python -m pip install .[lint]
python -m pylint src/oblivious test/ test/


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. First, install the dependencies required for packaging and publishing:

python -m pip install .[publish]

Ensure that the correct version number appears in pyproject.toml, and that any links in this README document to the Read the Docs documentation of this package (or its dependencies) have appropriate version numbers. Also ensure that the Read the Docs project for this library has an automation rule that activates and sets as the default all tagged versions. Create and push a tag for this version (replacing ?.?.? with the version number):

git tag ?.?.?
git push origin ?.?.?

Remove any old build/distribution files. Then, package the source into a distribution archive:

rm -rf build dist src/*.egg-info
python -m build --sdist --wheel .

Finally, upload the package distribution archive to PyPI:

python -m twine upload dist/*

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

oblivious-7.0.0.tar.gz (51.9 kB view hashes)

Uploaded Source

Built Distribution

oblivious-7.0.0-py3-none-any.whl (37.6 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page