steeldb/text/mod.rs
1//! Native text projection primitives (multilingual, CPU-only), ported from the TS ingest pipeline.
2//! • `model2vec` — the EN/JA/KO static span embedder (reuses the shipped `potion.f32` artifact).
3//! • `ot` — Sinkhorn optimal transport + k-means codebook for auto ontology sensing.
4//! The SPLADE / SPO-tagger ONNX stages plug in next via the `ort` crate (ONNX Runtime), running the
5//! same exported `.onnx` graphs unchanged.
6
7pub mod gazetteer;
8#[cfg(feature = "embed")]
9pub mod model2vec;
10pub mod ot;
11
12pub use gazetteer::{GazHit, Gazetteer};
13#[cfg(feature = "embed")]
14pub use model2vec::{cosine, Model2Vec};
15
16#[cfg(feature = "onnx")]
17pub mod splade;
18#[cfg(feature = "onnx")]
19pub mod tagger;
20
21#[cfg(feature = "onnx")]
22pub use splade::{FacetTerm, SpladeProjector};
23#[cfg(feature = "onnx")]
24pub use tagger::{Span, SpoTagger};