TransCrypto

Basic cryptography primitives implementation, a companion to "Criptografia, Métodos e Algoritmos".

Started in July/2025, by Daniel Balparda. Since version 1.0.2 it is PyPI package:

TransCrypto

License

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License here.

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.

Design assumptions / Disclaimers

The library is built to have reference, reliable, simple implementations of math and crypto primitives (e.g. RawEncrypt()/RawSign() and friends plus all the low-level primality and modular arithmetic). The issue is not that the library is unsafe, it is that the library is full of places that allow you to shoot yourself in the foot if you don't know what you are doing.
The library also has advanced top-level methods that are cryptographically safe and might be used in real-world scenarios (e.g. Encrypt()/Sign() and friends).
All library methods' int are tailored to be efficient with arbitrarily large integers.
Everything should work, as the library is extensively tested, but not necessarily the most efficient or safe for real-world cryptographic use. For real-world crypto you might consider other optimized/safe libraries that were built to be resistant to malicious attacks.
All operations in this library may be vulnerable to timing attacks. This may be a problem to your use-case or not depending on the situation.

All that being said, extreme care was taken that this is a good library with a solid safe implementation. Have fun!

CLI Apps

SafeTrans/safetrans: Safe cryptographic operations;
TransCrypto/transcrypto: Does all the operations but allows you to shoot yourself in the foot;
Profiler/profiler: Measure transcrypto performance.

Programming API

Install

To use in your project just do:

pip3 install transcrypto

and then from transcrypto import rsa (or other parts of the library) for using it.

Known dependencies:

Base Library

Humanized Sizes (IEC binary)

from transcrypto import utils

utils.HumanizedBytes(512)                 # '512 B'
utils.HumanizedBytes(2048)                # '2.00 KiB'
utils.HumanizedBytes(5 * 1024**3)         # '5.00 GiB'

Converts raw byte counts to binary-prefixed strings (B, KiB, MiB, GiB, TiB, PiB, EiB). Values under 1024 bytes are returned as integers with B; larger values use two decimals.

standard: 1 KiB = 1024 B, 1 MiB = 1024 KiB, …
errors: negative inputs raise InputError

Humanized Decimal Quantities (SI)

# Base (unitless)
utils.HumanizedDecimal(950)               # '950'
utils.HumanizedDecimal(1500)              # '1.50 k'

# With a unit (trimmed and attached)
utils.HumanizedDecimal(1500, ' Hz ')      # '1.50 kHz'
utils.HumanizedDecimal(0.123456, 'V')     # '0.1235 V'

# Large magnitudes
utils.HumanizedDecimal(3_200_000)         # '3.20 M'
utils.HumanizedDecimal(7.2e12, 'B/s')     # '7.20 TB/s'

Scales by powers of 1000 using SI prefixes (k, M, G, T, P, E). For values <1000, integers are shown as-is; small floats show four decimals. For scaled values, two decimals are used and the unit (if provided) is attached without a space (e.g., kHz).

unit handling: unit is stripped; <1000 values include a space before the unit ('950 Hz')
errors: negative or non-finite inputs raise InputError

Humanized Durations

utils.HumanizedSeconds(0)                 # '0.00 s'
utils.HumanizedSeconds(0.000004)          # '4.000 µs'
utils.HumanizedSeconds(0.25)              # '250.000 ms'
utils.HumanizedSeconds(42)                # '42.00 s'
utils.HumanizedSeconds(3661)              # '1.02 h'
utils.HumanizedSeconds(172800)            # '2.00 d'

Chooses an appropriate time unit based on magnitude and formats with fixed precision:

< 1 ms: microseconds with three decimals (µs)
< 1 s: milliseconds with three decimals (ms)
< 60 s: seconds with two decimals (s)
< 60 min: minutes with two decimals (min)
< 24 h: hours with two decimals (h)
≥ 24 h: days with two decimals (d)
special case: 0 → '0.00 s'
errors: negative or non-finite inputs raise InputError

Execution Timing

A flexible timing utility that works as a context manager, decorator, or manual timer object.

from transcrypto import base
import time

Context manager

with base.Timer('Block timing'):
    time.sleep(1.2)
# → logs: "Block timing: 1.20 s" (default via logging.info)

Starts timing on entry, stops on exit, and reports elapsed time automatically.

Decorator

@base.Timer('Function timing')
def slow_function():
    time.sleep(0.8)

slow_function()
# → logs: "Function timing: 0.80 s"

Wraps a function so that each call is automatically timed.

Manual use

tm = base.Timer('Inline timing', emit_print=True)
tm.Start()
time.sleep(0.1)
tm.Stop()   # prints: "Inline timing: 0.10 s"

Manual control over Start() and Stop() for precise measurement of custom intervals.

Key points

Label: required, shown in output; empty labels raise InputError
Output:
- emit_log=True → logging.info() (default)
- emit_print=True → direct print()
- Both can be enabled
Format: elapsed time is shown using HumanizedSeconds()
Safety:
- Cannot start an already started timer
- Cannot stop an unstarted or already stopped timer (raises Error)

Serialization Pipeline

These helpers turn arbitrary Python objects into compressed and/or encrypted binary blobs, and back again — with detailed timing and size logging.

from transcrypto import base

Serialize

data = {'x': 42, 'y': 'hello'}

# Basic serialization
blob = base.Serialize(data)

# With compression and encryption
blob = base.Serialize(
    data,
    compress=9,               # compression level (-22..22, default=3)
    key=my_symmetric_key      # must implement SymmetricCrypto
)

# Save directly to file
base.Serialize(data, file_path='/tmp/data.blob')

Serialization path:

obj → pickle → (compress) → (encrypt) → (save)

At each stage:

Data size is measured using HumanizedBytes
Duration is timed with Timer
Results are logged once at the end

Compression levels:

compress uses zstandard; see table below for speed/ratio trade-offs:

Level	Speed	Compression ratio	Typical use case
-5 to -1	Fastest	Poor (better than no compression)	Real-time or very latency-sensitive
0…3	Very fast	Good ratio	Default CLI choice, safe baseline
4…6	Moderate	Better ratio	Good compromise for general persistence
7…10	Slower	Marginally better ratio	Only if storage space is precious
11…15	Much slower	Slight gains	Large archives, not for runtime use
16…22	Very slow	Tiny gains	Archival-only, multi-GB datasets

Errors: invalid compression level is clamped to range; other input errors raise InputError.

DeSerialize

# From in-memory blob
obj = base.DeSerialize(data=blob)

# From file
obj = base.DeSerialize(file_path='/tmp/data.blob')

# With decryption
obj = base.DeSerialize(data=blob, key=my_symmetric_key)

Deserialization path:

data/file → (decrypt) → (decompress if Zstd) → unpickle

Compression is auto-detected via Zstandard magic numbers.
All steps are timed/logged like in Serialize.

Constraints & errors:

Exactly one of data or file_path must be provided.
file_path must exist; data must be at least 4 bytes.
Wrong key or corrupted data can raise CryptoError.

Cryptographically Secure Randomness

These helpers live in base and wrap Python’s secrets with additional checks and guarantees for crypto use-cases.

from transcrypto import base

Fixed-size random integers

# Generate a 256-bit integer (first bit always set)
r = base.RandBits(256)
assert r.bit_length() == 256

Produces a crypto-secure random integer with exactly n_bits bits (≥ 8). The most significant bit is guaranteed to be 1, so entropy is ~n_bits−1 — negligible for large crypto sizes.

errors: n_bits < 8 → InputError

Uniform random integers in a range

# Uniform between [10, 20] inclusive
n = base.RandInt(10, 20)
assert 10 <= n <= 20

Returns a crypto-secure integer uniformly distributed over the closed interval [min_int, max_int].

constraints: min_int ≥ 0 and < max_int
errors: invalid bounds → InputError

In-place secure shuffle

deck = list(range(10))
base.RandShuffle(deck)
print(deck)   # securely shuffled order

Performs an in-place Fisher–Yates shuffle using secrets.randbelow. Suitable for sensitive data ordering.

constraints: sequence length ≥ 2
errors: shorter sequences → InputError

Random byte strings

# 32 random bytes
b = base.RandBytes(32)
assert len(b) == 32

Generates n_bytes of high-quality crypto-secure random data.

constraints: n_bytes ≥ 1
errors: smaller values → InputError

Computing the Greatest Common Divisor

>>> from transcrypto import base
>>> base.GCD(462, 1071)
21
>>> base.GCD(0, 17)
17

The function is O(log(min(a, b))) and handles arbitrarily large integers. To find Bézout coefficients (x, y) such that ax + by = gcd(a, b) do:

>>> base.ExtendedGCD(462, 1071)
(21, -2, 1)
>>> 462 * -2 + 1071 * 1
21

Use-cases:

modular inverses: inv = x % m when gcd(a, m) == 1
solving linear Diophantine equations
RSA / ECC key generation internals

Fast Modular Arithmetic

from transcrypto import modmath

m = 2**256 - 189    # a large prime modulus

# Inverse ──────────────────────────────
x = 123456789
x_inv = modmath.ModInv(x, m)
assert (x * x_inv) % m == 1

# Division (x / y) mod m ──────────────
y = 987654321
z = modmath.ModDiv(x, y, m)      # solves z·y ≡ x (mod m)
assert (z * y) % m == x % m

# Exponentiation ──────────────────────
exp = modmath.ModExp(3, 10**20, m)   # ≈ log₂(y) time, handles huge exponents

Chinese Remainder Theorem (CRT) – Pair

from transcrypto import modmath

# Solve:
#   x ≡ 2 (mod 3)
#   x ≡ 3 (mod 5)
x = modmath.CRTPair(2, 3, 3, 5)
print(x)             # 8
assert x % 3 == 2
assert x % 5 == 3

Solves a system of two simultaneous congruences with pairwise co-prime moduli, returning the least non-negative solution x such that:

x ≡ a1 (mod m1)
x ≡ a2 (mod m2)
0 ≤ x < m1 * m2

Requirements:
- m1 ≥ 2, m2 ≥ 2, m1 != m2
- gcd(m1, m2) == 1 (co-prime)
Errors:
- invalid modulus values → InputError
- non co-prime moduli → ModularDivideError

This function is a 2-modulus variant; for multiple moduli, apply it iteratively or use a general CRT solver.

Modular Polynomials & Lagrange Interpolation

# f(t) = 7t³ − 3t² + 2t + 5  (coefficients constant-term first)
coefficients = [5, 2, -3, 7]
print(modmath.ModPolynomial(11, coefficients, 97))   # → 19

# Given three points build the degree-≤2 polynomial and evaluate it.
pts = {2: 4, 5: 3, 7: 1}
print(modmath.ModLagrangeInterpolate(9, pts, 11))   # → 2

Primality testing & Prime generators, Mersenne primes

modmath.IsPrime(2**127 - 1)              # True  (Mersenne prime)
modmath.IsPrime(3825123056546413051)     # False (strong pseudo-prime)

# Direct Miller–Rabin with custom witnesses
modmath.MillerRabinIsPrime(961748941, witnesses={2,7,61})

# Infinite iterator of primes ≥ 10⁶
for p in modmath.PrimeGenerator(1_000_000):
  print(p)
  if p > 1_000_100:
    break

# Secure random 384-bit prime (for RSA/ECC experiments)
p384 = modmath.NBitRandomPrimes(384).pop()

for k, m_p, perfect in modmath.MersennePrimesGenerator(0):
  print(f'p = {k:>8}  M = {m_p}  perfect = {perfect}')
  if k > 10000:          # stop after a few
    break

Cryptographic Hashing

Simple, fixed-output-size wrappers over Python’s hashlib for common digest operations, plus file hashing.

from transcrypto import base

SHA-256 hashing

h = base.Hash256(b'hello world')
assert len(h) == 32                       # bytes
print(h.hex())                            # 64 hex chars

Computes the SHA-256 digest of a byte string, returning exactly 32 bytes (256 bits). Suitable for fingerprints, commitments, or internal crypto primitives.

SHA-512 hashing

h = base.Hash512(b'hello world')
assert len(h) == 64                       # bytes
print(h.hex())                            # 128 hex chars

Computes the SHA-512 digest of a byte string, returning exactly 64 bytes (512 bits). Higher collision resistance and larger output space than SHA-256.

File hashing

# Default SHA-256
fh = base.FileHash('/path/to/file')
print(fh.hex())

# SHA-512
fh2 = base.FileHash('/path/to/file', digest='sha512')

Hashes a file from disk in streaming mode. By default uses SHA-256; digest='sha512' switches to SHA-512.

constraints:
- digest must be 'sha256' or 'sha512'
- full_path must exist
errors: invalid digest or missing file → InputError

Symmetric Encryption Interface

SymmetricCrypto is an abstract base class that defines the byte-in / byte-out contract for symmetric ciphers.

Metadata handling — if the algorithm uses a nonce or tag, the implementation must handle it internally (e.g., append it to ciphertext).
AEAD modes — if supported, associated_data must be authenticated; otherwise, a non-None value should raise InputError.

class MyAES(base.SymmetricCrypto):
    def Encrypt(self, plaintext: bytes, *, associated_data=None) -> bytes:
        ...
    def Decrypt(self, ciphertext: bytes, *, associated_data=None) -> bytes:
        ...

Crypto Objects General Properties (`CryptoKey`)

Cryptographic objects all derive from the CryptoKey class and will all have some important characteristics:

Will be safe to log and print, i.e., implement safe __str__() and __repr__() methods (in actuality repr will be exactly the same as str). The __str__() should always fully print the public parts of the object and obfuscate the private ones. This obfuscation allows for some debugging, if needed, but if the secrets are "too short" then it can be defeated by brute force. For usual crypto defaults the obfuscation is fine. The obfuscation is the fist 4 bytes of the SHA-512 for the value followed by an ellipsis (e.g. c9626f16…).
It will have a _DebugDump() method that does print secrets and can be used for debugging only.
Can be easily serialized to bytes by the blob property and to base-64 encoded str by the encoded property.
Can be serialized encrypted to bytes by the Blob(key=[SymmetricCrypto]) method and to encrypted base-64 encoded str by the Encoded(key=[SymmetricCrypto]) method.
Can be instantiated back as an object from str or bytes using the Load(data, key=[SymmetricCrypto] | None) method. The Load() will decide how to build the object and will work universally with all the serialization options discussed above.

Example:

from transcrypto import base, rsa, aes

priv = rsa.RSAPrivateKey.New(512)  # small key, but good for this example
print(str(priv))                   # safe, no secrets
# ▶ RSAPrivateKey(RSAPublicKey(public_modulus=pQaoxy-QeXSds1k9WsGjJw==, encrypt_exp=AQAB), modulus_p=f18141aa…, modulus_q=67494eb9…, decrypt_exp=c96db24a…)

print(priv._DebugDump())  # UNSAFE: prints secrets
# ▶ RSAPrivateKey(public_modulus=219357196311600536151291741191131996967, encrypt_exp=65537, modulus_p=13221374197986739361, modulus_q=16591104148992527047, decrypt_exp=37805202135275158391322585315542443073, remainder_p=9522084656682089473, remainder_q=8975656462800098363, q_inverse_p=11965562396596149292)

print(priv.blob)
# ▶ b"(\xb5/\xfd \x98\xc1\x04\x00\x80\x04\x95\x8d\x00\x00\x00\x00\x00\x00\x00\x8c\x0ftranscrypto.rsa\x94\x8c\rRSAPrivateKey\x94\x93\x94)\x81\x94]\x94(\x8a\x11'\xa3\xc1Z=Y\xb3\x9dty\x90/\xc7\xa8\x06\xa5\x00J\x01\x00\x01\x00\x8a\t\xa1\xc4\x83\x81\xc8\xc1{\xb7\x00\x8a\t\xc7\x8a5\xf0Qq?\xe6\x00\x8a\x10A$&\x82!\x1cy\x89r\xef\xeb\xa7_\x04q\x1c\x8a\t\x01\xbc\xbb\x8a\x8b=%\x84\x00\x8a\x08;\x94#s\xff\xef\x8f|\x8a\t,\x9c\xe2z\x9a7\x0e\xa6\x00eb."

print(priv.encoded)
# ▶ KLUv_WBwAIELAIAElWUBAAAAAAAAjA90cmFuc2NyeXB0by5yc2GUjA1SU0FQcml2YXRlS2V5lJOUKYGUXZQoikHf1EvsmZedAZve7TrLmobLAwuRIr_77TLG6G_0fsLGThERVJu075be8PLjUQYnLXcacZFQ5Fb1Iy1WtiE985euAEoBAAEAiiFR9ngiXMzkf41o5CRBY3h0D4DJVisDDhLmAWsiaHggzQCKIS_cmQ6MKXCtROtC7c_Mrsi9A-9NM8DksaHaRwvy6uTZAIpB4TVbsLxc41TEc19wIzpxbi9y5dW5gdfTkRQSSiz0ijmb8Xk3pyBfKAv8JbHp8Yv48gNZUfX67qq0J7yhJqeUoACKIbFb2kTNRzSqm3JRtjc2BPS-FnLFdadlFcV4-6IW7eqLAIogFZfzDN39gZLR9uTz4KHSTaqxWrJgP8-YYssjss6FlFKKIIItgCDv7ompNpY8gBs5bibN8XTsr-JOYSntDVT5Fe5vZWIu

key = aes.AESKey(key256=b'x' * 32)
print(key)
# ▶ AESKey(key256=86a86df7…)

encrypted = priv.Blob(key=key)
print(priv == rsa.RSAPrivateKey.Load(encrypted, key=key))
# ▶ True

AES-256 Symmetric Encryption

Implements AES-256 in GCM mode for authenticated encryption and decryption, plus an ECB mode helper for fixed-size block encoding. Also includes a high-iteration PBKDF2-based key derivation from static passwords.

Key creation

from transcrypto import aes

# From raw bytes (must be exactly 32 bytes)
key = aes.AESKey(key256=b'\x00' * 32)

# From a static password (slow, high-iteration PBKDF2-SHA256)
key = aes.AESKey.FromStaticPassword('correct horse battery staple')
print(key.encoded)  # URL-safe Base64

Length: key256 must be exactly 32 bytes
FromStaticPassword():
- Uses PBKDF2-HMAC-SHA256 with fixed salt and ~2 million iterations
- Designed for interactive password entry, not for password databases

AES-256 + GCM (default)

data = b'secret message'
aad  = b'metadata'

# Encrypt (returns IV + ciphertext + tag)
ct = key.Encrypt(data, associated_data=aad)

# Decrypt
pt = key.Decrypt(ct, associated_data=aad)
assert pt == data

Security:
- Random 128-bit IV (iv) per encryption
- Authenticated tag (128-bit) ensures integrity
- Optional associated_data is authenticated but not encrypted
Errors:
- Tag mismatch or wrong key → CryptoError

AES-256 + ECB (unsafe, fixed block only)

# ECB mode is for 16-byte block encoding ONLY
ecb = key.ECBEncoder()

block = b'16-byte string!!'
ct_block = ecb.Encrypt(block)
pt_block = ecb.Decrypt(ct_block)
assert pt_block == block

# Hex helpers
hex_ct = ecb.EncryptHex('00112233445566778899aabbccddeeff')

ECB mode:
- 16-byte plaintext ↔ 16-byte ciphertext
- No padding, no IV, no integrity — do not use for general encryption
- associated_data not supported

Key points:

GCM mode is secure for general use; ECB mode is for special low-level operations
Static password derivation is intentionally slow to resist brute force
All sizes and parameters are validated with InputError on misuse

RSA (Rivest-Shamir-Adleman) Public Cryptography

https://en.wikipedia.org/wiki/RSA_cryptosystem

This implementation is raw RSA, no OAEP or PSS! It works on the actual integers. For real uses you should look for higher-level implementations.

By default and deliberate choice the encryption exponent will be either 7 or 65537, depending on the size of phi=(p-1)*(q-1). If phi allows it the larger one will be chosen to avoid Coppersmith attacks.

from transcrypto import rsa

# Generate a key pair
priv = rsa.RSAPrivateKey.New(2048)        # 2048-bit modulus
pub  = rsa.RSAPublicKey.Copy(priv)        # public half
print(priv.public_modulus.bit_length())   # 2048

# Safe Encrypt & decrypt
msg = b'xyz'
cipher = pub.Encrypt(msg, associated_data=b'aad')
plain  = priv.Decrypt(cipher, associated_data=b'aad')
assert plain == msg

# Safe Sign & verify
signature = priv.Sign(msg)  # can also have associated_data, optionally
assert pub.Verify(msg, signature)

# Raw Encrypt & decrypt
msg = 123456789  # (Zero is forbidden by design; smallest valid message is 1.)
cipher = pub.RawEncrypt(msg)
plain  = priv.RawDecrypt(cipher)
assert plain == msg

# Raw Sign & verify
signature = priv.RawSign(msg)
assert pub.RawVerify(msg, signature)

# Blind signatures (obfuscation pair) - only works on raw RSA
pair = rsa.RSAObfuscationPair.New(pub)

blind_msg = pair.ObfuscateMessage(msg)            # what you send to signer
blind_sig = priv.RawSign(blind_msg)               # signer’s output

sig = pair.RevealOriginalSignature(msg, blind_sig)
assert pub.RawVerify(msg, sig)

El-Gamal Public-Key Cryptography

https://en.wikipedia.org/wiki/ElGamal_encryption

This is raw El-Gamal over a prime field — no padding, no hashing — and is not DSA. For real-world deployments, use a high-level library with authenticated encryption and proper encoding.

from transcrypto import elgamal

# Shared parameters (prime modulus, group base) for a group
shared = elgamal.ElGamalSharedPublicKey.New(256)
print(shared.prime_modulus)
print(shared.group_base)

# Public key from private
priv = elgamal.ElGamalPrivateKey.New(shared)
pub  = elgamal.ElGamalPublicKey.Copy(priv)

# Safe Encrypt & decrypt
msg = b'xyz'
cipher = pub.Encrypt(msg, associated_data=b'aad')
plain  = priv.Decrypt(cipher, associated_data=b'aad')
assert plain == msg

# Safe Sign & verify
signature = priv.Sign(msg)  # can also have associated_data, optionally
assert pub.Verify(msg, signature)

# Raw Encryption
msg = 42
cipher = pub.RawEncrypt(msg)
plain = priv.RawDecrypt(cipher)
assert plain == msg

# Raw Signature verify
sig = priv.RawSign(msg)
assert pub.RawVerify(msg, sig)

Key points:

Security parameters:
- Recommended prime_modulus bit length ≥ 2048 for real security
- Random values from base.RandBits
Ephemeral keys:
- Fresh per encryption/signature
- Must satisfy gcd(k, p-1) == 1
Errors:
- Bad ranges → InputError
- Invalid math relationships → CryptoError
Group sharing:
- Multiple parties can share (p, g) but have different (individual_base, decrypt_exp)

DSA (Digital Signature Algorithm)

https://en.wikipedia.org/wiki/Digital_Signature_Algorithm

This is raw DSA over a prime field — no hashing or padding. You sign/verify integers modulo q (prime_seed). For real use, hash the message first (e.g., SHA-256) and then map to an integer < q.

from transcrypto import dsa

# Shared parameters (p, q, g)
shared = dsa.DSASharedPublicKey.New(p_bits=1024, q_bits=160)
print(shared.prime_modulus)  # p
print(shared.prime_seed)     # q  (q | p-1)
print(shared.group_base)     # g

# Individual key pair
priv = dsa.DSAPrivateKey.New(shared)
pub  = dsa.DSAPublicKey.Copy(priv)

# Safe Sign & verify
msg = b'xyz'
signature = priv.Sign(msg)  # can also have associated_data, optionally
assert pub.Verify(msg, signature)

# Raw Sign & verify (message must be 1 ≤ m < q)
msg = 123456789 % shared.prime_seed
sig = priv.RawSign(msg)
assert pub.RawVerify(msg, sig)

ranges:
- 1 ≤ message < q
- signatures: (s1, s2) with 2 ≤ s1, s2 < q
errors:
- invalid ranges → InputError
- inconsistent parameters → CryptoError

Security notes

Choose large parameters (e.g., p ≥ 2048 bits, q ≥ 224 bits) for non-toy settings.
In practice, compute m = int.from_bytes(Hash(message), 'big') % q before calling Sign(m).

Advanced: custom primes generator

# Generate primes (p, q) with q | (p-1); also returns m = (p-1)//q
p, q, m = dsa.NBitRandomDSAPrimes(p_bits=1024, q_bits=160)
assert (p - 1) % q == 0

Used internally by DSASharedPublicKey.New(). Search breadth and retry caps are bounded; repeated failures raise CryptoError.

Public Bidding

This is a way of bidding on some commitment (the secret) that can be cryptographically proved later to not have been changed. To do that the secret is combined with 2 nonces (random values, n1 & n2) and a hash of it is taken (H=SHA-512(n1||n2||secret)). The hash H and one nonce n1 are public and divulged. The other nonce n2 and the secret are kept private and will be used to show secret was not changed since the beginning of the process. The nonces guarantee the secret cannot be brute-forced or changed after-the-fact. The whole process is as strong as SHA-512 collisions.

from transcrypto import base

# Generate the private and public bids
bid_priv = base.PrivateBid512.New(secret)    # this one you keep private
bid_pub = base.PublicBid512.Copy(bid_priv)   # this one you publish

# Checking that a bid is genuine requires the public bid and knowing the nonce and the secret:
print(bid_pub.VerifyBid(private_key, secret_bid))  # these come from a divulged private bid
# of course, you want to also make sure the provided private data matches your version of it, e.g.:
bid_pub_expected = base.PublicBid512.Copy(bid_priv)
print(bid_pub == bid_pub_expected)

SSS (Shamir Shared Secret)

https://en.wikipedia.org/wiki/Shamir's_secret_sharing

This is the information-theoretic SSS but with no authentication or binding between share and secret. Malicious share injection is possible! Add MAC or digital signature in hostile settings. Use at least 128-bit modulus for non-toy deployments.

from transcrypto import sss

# Generate parameters: at least 3 of 5 shares needed,
# coefficients & modulus are 128-bit primes
priv = sss.ShamirSharedSecretPrivate.New(minimum_shares=3, bit_length=128)
pub  = sss.ShamirSharedSecretPublic.Copy(priv)   # what you publish

print(f'threshold        : {pub.minimum}')
print(f'prime mod        : {pub.modulus}')
print(f'poly coefficients: {priv.polynomial}')         # keep these private!

# Safe Issuing shares

secret = b'xyz'
# Generate 5 shares, each has a copy of the encrypted secret
five_shares = priv.MakeDataShares(secret, 5)
for sh in five_shares:
  print(sh)

# Raw Issuing shares

secret = 0xC0FFEE
# Generate an unlimited stream; here we take 5
five_shares = list(priv.RawShares(secret, max_shares=5))
for sh in five_shares:
  print(f'share {sh.share_key} → {sh.share_value}')

A single share object looks like sss.ShamirSharePrivate(minimum=3, modulus=..., share_key=42, share_value=123456789).

# Safe Re-constructing the secret
secret = b'xyz'
five_shares = priv.MakeDataShares(secret, 5)
subset = five_shares[:3]                   # any 3 distinct shares
recovered = subset[0].RecoverData(subset)  # each share has the encrypted data, so you ask it to join with the others
assert recovered == secret

# Raw Re-constructing the secret
secret = 0xC0FFEE
five_shares = list(priv.RawShares(secret, max_shares=5))
subset = five_shares[:3]          # any 3 distinct shares
recovered = pub.RawRecoverSecret(subset)
assert recovered == secret

If you supply fewer than minimum shares you get a CryptoError, unless you explicitly override:

try:
  pub.RawRecoverSecret(five_shares[:2])        # raises
except Exception as e:
  print(e)                                  # "unrecoverable secret …"

# Force the interpolation even with 2 points (gives a wrong secret, of course)
print(pub.RawRecoverSecret(five_shares[:2], force_recover=True))

# Checking that a share is genuine

share = five_shares[0]
ok = priv.RawVerifyShare(secret, share)       # ▶ True
tampered = sss.ShamirSharePrivate(
    minimum=share.minimum,
    modulus=share.modulus,
    share_key=share.share_key,
    share_value=(share.share_value + 1) % share.modulus)
print(priv.RawVerifyShare(secret, tampered))  # ▶ False

Appendix: Development Instructions

Setup

If you want to develop for this project, first install python 3.13 and Poetry, but to get the versions you will need, we suggest you do it like this (Linux):

sudo apt-get update
sudo apt-get upgrade
sudo apt-get install git python3 python3-pip pipx python3-dev python3-venv build-essential software-properties-common

sudo add-apt-repository ppa:deadsnakes/ppa  # install arbitrary python version
sudo apt-get update
sudo apt-get install python3.13

sudo apt-get remove python3-poetry
python3.13 -m pipx ensurepath
# re-open terminal
pipx install poetry
poetry --version  # should be >=2.1

poetry config virtualenvs.in-project true  # creates .venv inside project directory
poetry config pypi-token.pypi <TOKEN>      # add your personal PyPI project token, if any

or this (Mac):

brew update
brew upgrade
brew cleanup -s

brew install git python@3.13  # install arbitrary python version

brew uninstall poetry
python3.13 -m pip install --user pipx
python3.13 -m pipx ensurepath
# re-open terminal
pipx install poetry
poetry --version  # should be >=2.1

poetry config virtualenvs.in-project true  # creates .venv inside project directory
poetry config pypi-token.pypi <TOKEN>      # add your personal PyPI project token, if any

Now install the project:

git clone https://github.com/balparda/transcrypto.git transcrypto
cd transcrypto

poetry env use python3.13  # creates the venv
poetry install --sync      # HONOR the project's poetry.lock file, uninstalls stray packages
poetry env info            # no-op: just to check

poetry run pytest -vvv
# or any command as:
poetry run <any-command>

To activate like a regular environment do:

poetry env activate
# will print activation command which you next execute, or you can do:
source .venv/bin/activate                        # if .venv is local to the project
source "$(poetry env info --path)/bin/activate"  # for other paths

pytest  # or other commands

deactivate

Updating Dependencies

To update poetry.lock file to more current versions do poetry update, it will ignore the current lock, update, and rewrite the poetry.lock file.

To add a new dependency you should do:

poetry add "pkg>=1.2.3"  # regenerates lock, updates env (adds dep to prod code)
poetry add -G dev "pkg>=1.2.3"  # adds dep to dev code ("group" dev)
# also remember: "pkg@^1.2.3" = latest 1.* ; "pkg@~1.2.3" = latest 1.2.* ; "pkg@1.2.3" exact

If you manually added a dependency to pyproject.toml you should very carefully recreate the environment and files:

rm -rf .venv .poetry poetry.lock
poetry env use python3.13
poetry install

Remember to check your diffs before submitting (especially poetry.lock) to avoid surprises!

When dependencies change, always regenerate requirements.txt by running:

poetry export --format requirements.txt --without-hashes --output requirements.txt

Creating a New Version

# bump the version!
poetry version minor  # updates 1.6 to 1.7, for example
# or:
poetry version patch  # updates 1.6 to 1.6.1
# or:
poetry version <version-number>
# (also updates `pyproject.toml` and `poetry.lock`)

# publish to GIT, including a TAG
git commit -a -m "release version 1.0.2"
git tag 1.0.2
git push
git push --tags

# prepare package for PyPI
poetry build
poetry publish

If you changed the CLI interface at all, in any tool, run:

./tools/generate_docs.sh

You can find the 10 top slowest tests by running:

poetry run pytest -vvv -q --durations=30

poetry run pytest -vvv -q --durations=30 -m "not slow"      # find slow > 0.1s
poetry run pytest -vvv -q --durations=30 -m "not veryslow"  # find veryslow > 1s

poetry run pytest -vvv -q --durations=30 -m slow      # check
poetry run pytest -vvv -q --durations=30 -m veryslow  # check

You can search for flaky tests by running all tests 100 times, or more:

poetry run pytest --flake-finder --flake-runs=100
poetry run pytest --flake-finder --flake-runs=500 -m "not veryslow"
poetry run pytest --flake-finder --flake-runs=10000 -m "not slow"

You can instrument your code to find bottlenecks:

$ source .venv/bin/activate
$ which transcrypto
/path/to/.venv/bin/transcrypto  # place this in the command below:
$ pyinstrument -r html -o dsa_shared.html -- /path/to/.venv/bin/transcrypto -p rsa-key rsa new
$ deactivate

Hint: 85%+ is inside MillerRabinIsPrime()/gmpy2.powmod()...

Name		Name	Last commit message	Last commit date
Latest commit History 85 Commits
.vscode		.vscode
src/transcrypto		src/transcrypto
tests		tests
tools		tools
.gitignore		.gitignore
LICENSE		LICENSE
README.md		README.md
poetry.lock		poetry.lock
profiler.md		profiler.md
pyproject.toml		pyproject.toml
requirements.txt		requirements.txt
safetrans.md		safetrans.md
transcrypto.md		transcrypto.md

License

balparda/transcrypto

Folders and files

Latest commit

History

Repository files navigation

TransCrypto

License

Design assumptions / Disclaimers

CLI Apps

Programming API

Install

Base Library

Humanized Sizes (IEC binary)

Humanized Decimal Quantities (SI)

Humanized Durations

Execution Timing

Context manager

Decorator

Manual use

Key points

Serialization Pipeline

Serialize

DeSerialize

Cryptographically Secure Randomness

Fixed-size random integers

Uniform random integers in a range

In-place secure shuffle

Random byte strings

Computing the Greatest Common Divisor

Fast Modular Arithmetic

Chinese Remainder Theorem (CRT) – Pair

Modular Polynomials & Lagrange Interpolation

Primality testing & Prime generators, Mersenne primes

Cryptographic Hashing

SHA-256 hashing

SHA-512 hashing

File hashing

Symmetric Encryption Interface

Crypto Objects General Properties (CryptoKey)

AES-256 Symmetric Encryption

Key creation

AES-256 + GCM (default)

AES-256 + ECB (unsafe, fixed block only)

RSA (Rivest-Shamir-Adleman) Public Cryptography

El-Gamal Public-Key Cryptography

DSA (Digital Signature Algorithm)

Security notes

Advanced: custom primes generator

Public Bidding

SSS (Shamir Shared Secret)

Appendix: Development Instructions

Setup

Updating Dependencies

Creating a New Version

About

Topics

Resources

License

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Languages

Crypto Objects General Properties (`CryptoKey`)

Packages