zer-schema 1.0.2

Schema inference and Fellegi-Sunter model registry for the zer entity-resolution library
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33

# 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.

- **Documentation**: [docs.zal-analytics.ch](https://docs.zal-analytics.ch)
- **Website**: [www.zal-analytics.ch](https://www.zal-analytics.ch)
- **Support & feedback**: [info@zal-analytics.ch](mailto:info@zal-analytics.ch)

## 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 `ModelArtifact`s |
| `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

```rust
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)?;
```

## License

Apache-2.0 ยท [GitHub](https://github.com/ZAL-Analytics/zer)