personalitygen

personalitygen generates simulated character personalities for games, storytelling, simulations, and tests. It supports conventional Big Five (OCEAN) profiles and Adaptive Bifurcated Big Five (ABBF) signed-vector profiles.

Intent and scope

• Generate full Big Five profiles with sub-trait components and aggregate scores.
• Generate ABBF profiles as signed 5D vectors with dominant poles.
• Bias outputs by life stage using tuned Gaussian distributions (child, young adult, adult).
• Derive a conflict-resolution style from trait weights, plus mapped concern-for-self/others.
• Project Big Five profiles into ABBF vectors for systems that want both model shapes.
• Support deterministic generation by accepting a seeded random source.
• Stay lightweight and dependency-free (pure Python).

Model overview

• Big Five traits: openness, conscientiousness, extraversion, agreeableness, neuroticism.
• Each trait is composed of three sub-traits and an aggregate score.
• Life stage changes sampling means. Standard deviation is fixed per trait.
• Conflict-resolution style is selected from avoiding, obliging, integrating, dominating, or compromising based on trait scores.
• ABBF profiles use five signed axes in chart order: order, chaos, cooperation, conflict, and competition.
• Positive ABBF values select the chart's left pole; negative values select the chart's right pole.

Usage

from personalitygen import BigFivePersonality, LifeStage

personality = BigFivePersonality.random(LifeStage.ADULT)
print(personality.traits)
print(personality.conflict_resolution)

If you want deterministic output, pass a seeded random number generator:

import random
from personalitygen import BigFiveTraits, LifeStage

rng = random.Random(42)
traits = BigFiveTraits.random(LifeStage.YOUNG_ADULT, rng=rng)
print(traits)

Custom random sources need only provide uniform(minimum, maximum). They may return either inclusive endpoint and must return a finite value inside the requested range. A fixed draw sequence is reproducible within one package version; the same integer seed is not guaranteed to produce identical profiles after a later model or algorithm change.

ABBF profiles can be generated directly or projected from Big Five traits:

from personalitygen import AdaptiveBifurcatedProfile

profile = AdaptiveBifurcatedProfile.random()
print(profile.vector)
print(profile.dominant_poles(threshold=0.2))

projected = AdaptiveBifurcatedProfile.from_big_five(traits)
print(projected.cosine_similarity(profile))

Development

This package targets Python 3.11+.

pdm install --group dev
pdm run check

Deeper architecture, quality, and maintenance guidance lives in the repository documentation. Simulation recipes live in the usage guide, and runnable examples live in examples/.