Module conditioning 

Centralized entropy conditioning module.

ALL post-processing of raw entropy lives here — no conditioning code should exist in individual source implementations. Sources produce raw bytes; this module is the single, auditable gateway for any transformation.

Architecture

Source → Raw Bytes → Conditioning Layer (this module) → Output

Conditioning Modes

• Raw: No processing. XOR-combined bytes pass through unchanged. Preserves the actual hardware noise signal for research.
• VonNeumann: Debias only. Removes first-order bias without destroying the noise structure. Output is shorter than input (~25% yield).
• Sha256: Full SHA-256 conditioning with counter and timestamp mixing. Produces cryptographically strong output but destroys the raw signal.

Most QRNG APIs (ANU, Outshift/Cisco) apply DRBG post-processing that makes output indistinguishable from PRNG. The Raw mode here is what makes openentropy useful for researchers studying actual hardware noise.

Structs

MinEntropyReport

Min-entropy analysis report with individual estimator results.

QualityReport

Enums

ConditioningMode

Conditioning mode for entropy output.

Functions

collision_estimate

Collision estimator — NIST SP 800-90B Section 6.3.2. Measures average distance between repeated values. Returns estimated min-entropy bits per sample.

compression_estimate

Compression estimator — NIST SP 800-90B Section 6.3.4. Uses Maurer’s universal statistic to estimate entropy via compression. Returns estimated min-entropy bits per sample.

condition

Apply the specified conditioning mode to raw entropy bytes.

grade_min_entropy

Grade a source based on its min-entropy (H∞) value.

markov_estimate

Markov estimator — NIST SP 800-90B Section 6.3.3. Models first-order dependencies between consecutive samples. Returns estimated min-entropy bits per sample.

mcv_estimate

Most Common Value (MCV) estimator — NIST SP 800-90B Section 6.3.1. Estimates min-entropy with upper bound on p_max using confidence interval. Returns (min_entropy_bits_per_sample, p_max_upper_bound).

min_entropy

Min-entropy estimate: H∞ = -log2(max probability). More conservative than Shannon — reflects worst-case guessing probability. Returns bits per sample (0.0 to 8.0 for byte-valued data).

min_entropy_estimate

Combined min-entropy estimate using multiple estimators. Takes the minimum (most conservative) across all methods. Returns a MinEntropyReport with individual and combined estimates.

quick_min_entropy

Quick min-entropy (just the combined estimate, no full report).

quick_quality

Quick quality assessment.

quick_shannon

Quick Shannon entropy in bits/byte for a byte slice.

sha256_condition

SHA-256 condition with explicit state, sample, counter, and extra data. Returns (new_state, 32-byte digest).

sha256_condition_bytes

SHA-256 chained conditioning: stretches or compresses raw bytes to exactly n_output bytes using counter-mode hashing.

t_tuple_estimate

t-Tuple estimator — NIST SP 800-90B Section 6.3.5. Estimates entropy from most frequent t-length tuple. Returns estimated min-entropy bits per sample.

von_neumann_debias

Von Neumann debiasing: extract unbiased bits from a biased stream.

xor_fold

XOR-fold: reduce data by XORing the first half with the second half.