Database Relationship Schematic
This schematic shows the main relational path for the Earthbond build: tenant/auth scope, raw source registration, job execution, canonical well-centric truth, semantic projection, and audit traceability.
How To Read
core: tenancy, users, and access gates.raw: source bundle and object registry before canonicalization.ops: operational truth, canonical facts, and decision-support outputs.semantic: ontology-style entity projection from canonical truth.audit: immutable event and evidence trail.
Key Foreign-Key Paths
| Path | Why It Matters |
|---|---|
raw.source_bundles.bundle_id -> raw.source_objects.bundle_id -> raw.extracted_fields.source_object_id | Proves how one staged bundle becomes object-level registration and then field-level extraction. |
ops.upload_sessions.upload_id -> ops.well_logs.source_upload_id -> ops.well_interpretations.source_well_log_id -> ops.well_pay_events.interpretation_id | Explains how a raw upload can become interpreted log intervals and payload candidates. |
ops.canonical_wells.canonical_well_id -> canonical_well_aliases / canonical_well_locations / canonical_well_source_links | Shows the well identity hub and why the build is well-centric after normalization. |
ops.canonical_wells.canonical_well_id -> formation_tops / completion_intervals / production_records / technical_reports / witsml rows | Shows how static and dynamic evidence attach to one canonical well identity. |
ops.canonical_wells -> semantic.entities -> semantic.entity_links | Shows that ontology projection is downstream from canonical truth, not a replacement for it. |
audit.event_ledger and audit.evidence_packs | Preserve the replay and defense trail across the workflow. |
Boundary
This schematic is intentionally selective. It is designed for decision review, not exhaustive DDL browsing. The full table inventory, IDs, row counts, and column-level provenance remain in the storage/DB audit outputs and the data-platform storage manual.