Guides

Practical guides for building and tuning identity resolution with Kanoniv.

Spec Authoring

• Writing Your First Spec: Build a complete identity spec from scratch, step by step
• Matching Strategy: Choose the right rules, algorithms, thresholds, and blocking for your use case
• Tuning Match Quality: Diagnose false positives and false negatives, iterate on precision and recall

Migration

• Migrating to Kanoniv: Side-by-side migration from Dedupe, Splink, or hand-rolled scripts

Versioning

• Versioning Overview: Understand the four versioning layers — specs, rules, entities, and temporal
• Spec Versioning: Track changes with plan_hash and diff()
• Rule Versioning: Auto-incrementing rule versions, testing, and rollback
• Entity History & Audit: Audit trail, locking, reverting, temporal matching
• ML Training Data: Pin training datasets to spec versions for reproducible ML

Reference

These pages complement the guides with detailed specifications:

• Spec Reference: Complete YAML schema for every spec key
• Rules: All rule types, algorithms, weights, and blocking strategies
• Decision: Scoring methods, thresholds, and conflict strategies
• Survivorship: Field-level strategies for golden record construction
• Python SDK: Core SDK — validate, plan, diff, and reconcile locally
• Kanoniv Cloud: Managed identity graph with real-time resolution API