Python API

TL;DR

Use the unified public entry points first: fit_model, sample_model, and filter_model. They route by ensemble, family, and constraint.

Main imports

from menobis.models import Constraint, Ensemble, ModelFamily, fit_model, sample_model
from menobis.filtering import filter_model
from menobis.data.io import read_edges, write_edges

Model selectors

Selector	Values
`ModelFamily`	`ME`, `B`, `W`
`Ensemble`	`GRAND_CANONICAL`, `CANONICAL`, `MICROCANONICAL`
`Constraint`	`STRENGTH`, `STRENGTH_COST`, `STRENGTH_EDGES`, `STRENGTH_DEGREE`, `DEGREE_EVENTS`

Core functions

Function	Purpose
`fit_model(...)`	solve model parameters from constraints
`sample_model(...)`	sample a sparse `EdgeTable` from a fit or ME stubs
`filter_model(edges, ...)`	classify edges against an independent null

Minimal fit/sample/filter

fit = fit_model(
    family=ModelFamily.ME,
    constraint=Constraint.STRENGTH,
    strength_out=strength_out,
    strength_in=strength_in,
    self_loops=False,
)

sample = sample_model(
    family=ModelFamily.ME,
    constraint=Constraint.STRENGTH,
    fit=fit,
    seed=42,
)

result = filter_model(
    edges,
    family=ModelFamily.ME,
    constraint=Constraint.STRENGTH,
    fit=fit,
)

Common data/result types

Type	Module	Meaning
`EdgeTable`	`menobis.data.frames`	sparse `source`, `target`, `weight` arrays
`ProbabilityTable`	`menobis.data.frames`	sparse custom probabilities/rates
`FitResult`	`menobis.models`	base fit protocol with diagnostics
`StrengthFit`	`menobis.models`	strength multipliers `x`, `y`
`StrengthCostFit`	`menobis.models`	`x`, `y`, and `gamma`
`StrengthEdgesFit`	`menobis.models`	`x`, `y`, and global edge multiplier
`StrengthDegreeFit`	`menobis.models`	strength and degree multipliers
`DegreeEventsFit`	`menobis.models`	degree occupation plus positive-weight parameter
`FilterResult`	`menobis.filtering`	upper/lower/compatible/absent classifications

Analysis helpers

Function	Purpose
`directed_strengths(edges)`	out/in strengths
`directed_degrees(edges)`	out/in binary degrees
`compute_all_stats(edges)`	strengths, degrees, Y2, nearest-neighbour stats
`weight_distribution(edges)`	occupation-count histogram
`clustering_coefficient(edges)`	binary clustering
`weighted_clustering_coefficient(edges)`	occupation-weighted clustering helper

Synthetic fixtures

Use these in examples and tests to avoid infeasible arbitrary constraints:

from menobis.utilities.synthetic import (
    derive_synthetic_constraints,
    generate_pa_geographic_network,
)