NietzscheDB: Multi-Manifold Architecture

NietzscheDB is the first database to implement Geometric Perspectivism. It doesn't treat geometry as a fixed property of the data, but as a dynamic lens applied at query time.

The Unified Poincaré Layer

All embeddings in NietzscheDB are stored in the Poincaré Ball model ($|x| < 1$). This is the "anatomical" storage layer because it is the most efficient at representing hierarchical structures.

However, the engine provides 4 specialized projections:

1. Poincaré Ball (Hierarchy)

• Use case: General similarity and taxonomic depth.
• Metric: Geodesic distance $d(u,v) = \text{arcosh}(1 + \frac{2|u-v|^2}{(1-|u|^2)(1-|v|^2)})$.
• Logic: Objects near the boundary are more specific; objects near the center are more abstract.

2. Klein Model (Rational Reasoning)

• Use case: Collinearity checks and logical chain verification.
• Logic: In Klein space, hyperbolic geodesics are straight lines. This transforms complex manifold navigation into O(1) linear checks.
• Operation: PROJECT_TO_KLEIN(u) = \frac{u}{1+\|u\|^2}.

3. Minkowski Spacetime (Causality)

• Use case: Causal auditing and temporal scrubbing.
• Metric: Lorentzian interval $ds^2 = -c^2\Delta t^2 + |\Delta x|^2$.
• Operation: The database treats the created_at timestamp as the 4th dimension ($t$) and the Poincaré embedding as the spatial vector ($x$).
• Agential Role: Ensure that a refined concept (effect) always falls within the "future light cone" of its foundational concept (cause).

4. Riemann Sphere (Synthesis)

• Use case: Combining conflicting ideas (Hegelian Dialectic).
• Logic: Two points on opposite sides of the Poincaré ball represent total antithesis. Projecting them onto the Riemann Sphere allows for a spherical midpoint that represents a "shallower" synthesis point.

Geometric Transition Table

Implementation

The nietzsche-hyp-ops crate handles these transitions using high-precision f64 arithmetic, ensuring that cascaded roundtrip errors remain below 1e-4.