zer-schema 1.0.0

Schema inference and Fellegi-Sunter model registry for the zer entity-resolution library
Documentation

zer-schema

Schema inference and Fellegi-Sunter model registry for the zer entity-resolution library.

This crate handles two concerns: automatically detecting the schema of an incoming dataset, and persisting trained Fellegi-Sunter parameters so they can be reused across pipeline runs without re-running EM.

What it provides

Item Description
SchemaInferrer Detects FieldKind from column names and value patterns; no manual schema definition needed
SchemaFingerprint Compact SHA-256 identity for a schema plus its data distribution
SchemaRegistry sled-backed persistent store for trained ModelArtifacts
ModelArtifact Serialisable container for m/u parameters, thresholds, and metadata
StartupMode Decision enum: load exact match, warm-start EM, or run full EM from priors

Usage

[dependencies]
zer = { version = "1.0" }  # includes zer-schema
# or directly:
zer-schema = "0.1"

use zer_schema::{SchemaInferrer, SchemaRegistry};

let schema   = SchemaInferrer::default().infer_from_csv_headers(&headers)?;
let registry = SchemaRegistry::open("./model-store")?;
let mode     = registry.lookup_startup_mode(&schema, &fingerprint)?;

Part of the zer ecosystem

zer · GitHub