Skip to content

LZGraphs

API Reference

MuteJester/LZGraphs

API Reference¶

Complete reference documentation for all LZGraphs classes and functions.

Core¶

LZGraph — Main graph class (all variants)
SimulationResult — Output of simulate()
PgenDistribution — Analytical PGEN distribution

Analysis & Utilities¶

Module Functions — jensen_shannon_divergence, k_diversity, etc.
CLI Tool — lzg command reference
Exceptions — Error handling

Import Patterns¶

Classes¶

from LZGraphs import LZGraph, SimulationResult, PgenDistribution

Functions¶

from LZGraphs import (
    jensen_shannon_divergence,
    k_diversity,
    saturation_curve,
    lz76_decompose,
    set_log_level,
)

Exceptions¶

from LZGraphs import LZGraphError, NoGeneDataError, ConvergenceError, CorruptFileError

The `LZGraph` Class¶

There is a single graph class with a variant parameter:

from LZGraphs import LZGraph

# Amino acid positional (most common)
graph = LZGraph(sequences, variant='aap')

# Nucleotide double positional
graph = LZGraph(sequences, variant='ndp')

# Naive (no positional encoding)
graph = LZGraph(sequences, variant='naive')

All variants share the same methods. See the LZGraph reference for the full API.

Quick Method Reference¶

Category	Method	Description
Scoring	`lzpgen(seq)`	Log-probability of sequence(s)
Simulation	`simulate(n)`	Generate n sequences
Diversity	`effective_diversity()`	exp(Shannon entropy) = D(1)
	`hill_number(alpha)`	Hill diversity number D(α)
	`hill_numbers(orders)`	Multiple Hill numbers
	`hill_curve()`	Full diversity curve
	`diversity_profile()`	Entropy, diversity, uniformity
Occupancy	`predicted_richness(depth)`	Expected unique seqs at depth
	`predicted_overlap(d_i, d_j)`	Expected shared sequences
	`richness_curve(depths)`	Richness at multiple depths
	`predict_sharing(draws)`	Sharing spectrum across donors
Distribution	`pgen_moments()`	Mean/var of log-PGEN
	`pgen_distribution()`	Analytical Gaussian mixture
	`pgen_diagnostics()`	Check proper distribution
	`pgen_dynamic_range()`	Dynamic range in orders of mag
Perplexity	`sequence_perplexity(seq)`	Single sequence perplexity
	`repertoire_perplexity(seqs)`	Average repertoire perplexity
	`path_entropy_rate(seqs)`	Entropy rate (bits/token)
Graph Ops	`union(other)` / `a \\| b`	Sum edge counts
	`intersection(other)` / `a & b`	Shared edges, min counts
	`difference(other)` / `a - b`	Subtract edge counts
	`weighted_merge(other, α, β)`	Linear combination
	`posterior(seqs)`	Bayesian posterior update
Features	`feature_aligned(query)`	Aligned feature vector
	`feature_mass_profile()`	Position-based mass profile
	`feature_stats()`	15-element stats vector
IO	`save(path)`	Save to `.lzg` binary format
	`LZGraph.load(path)`	Load from `.lzg` file
Info	`summary()`	Structural summary dict
	`n_nodes`, `n_edges`	Graph size
	`variant`, `is_dag`	Graph properties
	`has_gene_data`	Whether gene data is available

Version Information¶

import LZGraphs
print(LZGraphs.__version__)  # 3.0.2