GoldenMatch

Suite orchestration for Identity Graph

GoldenPipe now orchestrates the v1.15 Identity Graph. One CLI / Python / Airflow path runs Check -> Flow -> Match -> Identity end-to-end and maintains a durable identity store with stable entity_ids across runs.

Quickstart

goldenpipe run customers.csv \
  --identity-path .goldenmatch/identity.db \
  --identity-source-pk customer_id \
  --identity-dataset customers

import goldenpipe as gp

result = gp.run(
    "customers.csv",
    identity_opts={
        "path": ".goldenmatch/identity.db",
        "source_pk_column": "customer_id",
        "dataset": "customers",
    },
)
print(result.artifacts["identity_summary"])

What's new

• New stage goldenmatch.identity_resolve registered at the goldenpipe.stages entry-point. Wraps goldenmatch.identity.resolve_clusters.
• CLI flags on goldenpipe run: --identity-path, --identity-dataset, --identity-source-pk, --identity-weak-threshold.
• gp.run(source, identity_opts={...}) Python entry-point.
• New Airflow DAG examples/airflow/golden_suite_identity_graph.py — daily Check→Flow→Match→Identity with S3-synced identity store.
• Public docs at packages/python/goldenpipe/docs/identity-graph.md.
• DedupeStage surfaces two new artifacts: scored_pairs and matchkey_used. Backwards-compatible.

Requires

goldenmatch >= 1.15.0. No other-package version bumps.

Tests

12 new (8 stage + 4 CLI). Full suite: 136 passing, 0 regressions.

🤖 Generated with Claude Code