Lightweight Partially Homomorphic Encryption Library for Python

Common errors

• TypeError: unsupported operand type(s) for *: 'Ciphertext' and 'Ciphertext'

  cause You are attempting to multiply two `Ciphertext` objects using an algorithm (e.g., Paillier) that is not multiplicatively homomorphic for ciphertexts.
  fix Ensure the chosen algorithm supports the desired homomorphic operation. For multiplicative operations on ciphertexts, consider algorithms like RSA. For scalar multiplication (ciphertext by a known plain constant), many additively homomorphic schemes like Paillier support it. Refer to the 'Summary of Homomorphic Features' in the LightPHE documentation.

• Message does not get decrypted in Python with LightPHE

  cause This generic error often occurs if the cryptosystem was built with one algorithm but operations or decryption are attempted in a way incompatible with that algorithm's properties, or if the keys are mismatched/corrupted. Another common cause is attempting an operation that, while syntactically correct, breaks the homomorphic properties for the specific scheme.
  fix Double-check that the `algorithm_name` used during `LightPHE` initialization is appropriate for your desired operations. Ensure that the `encrypt` and `decrypt` calls use the same `LightPHE` instance (or correctly serialized/deserialized keys). Verify that intermediate homomorphic operations respect the limitations of the chosen scheme (e.g., not performing multiplication if only addition is supported).

Warnings

• gotcha Attempting homomorphic operations not supported by the chosen algorithm will raise an exception. For example, Paillier is additively homomorphic but not multiplicatively homomorphic for two ciphertexts.
  Fix: Consult the library's documentation (e.g., GitHub README table) to verify which homomorphic operations (addition, multiplication, scalar multiplication, XOR) are supported by each specific algorithm (RSA, Paillier, ElGamal, etc.) before performing operations.
• gotcha Decryption for some algorithms, like Exponential ElGamal or Elliptic Curve ElGamal, involves solving the discrete logarithm problem. While the library handles this, very large plaintexts might impact practical decryption times or even theoretical feasibility if not carefully considered.
  Fix: Be mindful of the plaintext value range and the computational implications for algorithms relying on discrete logarithm for decryption, especially in performance-sensitive applications.

Install

• pip install lightphe Install latest version

Imports

• LightPHE
  wrong:

  from lightphe.schemes import Paillier

  correct:

  from lightphe import LightPHE

  The `LightPHE` class acts as a factory to build cryptosystems for various algorithms; direct import of internal scheme modules is not the intended public API.

Quickstart

This quickstart demonstrates how to initialize a partially homomorphic cryptosystem (Paillier), encrypt two plaintexts, perform a homomorphic addition on their ciphertexts, and decrypt the result to confirm it matches the sum of the original plaintexts.

from lightphe import LightPHE

# Build an additively homomorphic cryptosystem (e.g., Paillier)
# Other options: 'RSA', 'ElGamal', 'Exponential-ElGamal', 'Damgard-Jurik', etc.
cs = LightPHE(algorithm_name = "Paillier")

# Define plaintexts
m1 = 17
m2 = 21

# Encrypt plaintexts (private key is not required for encryption)
c1 = cs.encrypt(m1)
c2 = cs.encrypt(m2)

# Perform homomorphic addition on ciphertexts
c3 = c1 + c2

# Decrypt the result to get the sum of plaintexts
d3 = cs.decrypt(c3)

print(f"Plaintext 1: {m1}")
print(f"Plaintext 2: {m2}")
print(f"Ciphertext 1: {c1}")
print(f"Ciphertext 2: {c2}")
print(f"Homomorphic Sum (c1 + c2): {c3}")
print(f"Decrypted Sum: {d3}")

assert d3 == m1 + m2, "Decrypted sum does not match plaintext sum"
print("Assertion successful: Decrypted sum matches expected value.")