FEAT: unified data store layout, acquisition gateway, and StoreClient API

[RBhupi]

PR tags (required)

Select all that apply.

• code-change
• scientific/algorithmic
• design/architecture

What changed

2-4 bullets in plain language.

• Implemented the core data store architecture (adapt init layout, run registry, acquisition gateway, immutable object store, and StoreClient query surface).
• Added canonicalization of source variables at ingest via reader.field_map and established deterministic schemas with SchemaLedger and single-home DDL ratchets.
• Refactored runtime components, dashboard, and renderers to consume data via the StoreClient and resolve tracking fields through run provenance.
• Updated user documentation, design guides, and layer-stack dependencies to enforce store-only consumer access.

Why

1-2 lines on the user/problem impact.

Establishes a unified, immutable data store and query layer that eliminates fragmented pre-store persistence, guaranteeing schema consistency and repeatable tracking provenance across runs.

Target version (required)

Aiming for: v0.2

Validation

Mark each item as done or N/A. Add one short evidence line for each item.

• I ran ruff
• I ran import-linter
• I ran tests

AI usage & manual verification

• AI used: yes
• Model/tooling: Claude Fabel 5
• Tasks delegated to AI: coding
• Manual verification performed: Manual verification is done using GUI.

Reviewer guide

Where should review start, and what are the key files/areas?

• Entry point: adapt/store/ layout definitions and StoreClient interface implementation.
• Key files/areas (1-3):
  1. StoreClient query surface and run provenance resolution.
  2. Acquisition gateway and ingest canonicalization (reader.field_map).
  3. SchemaLedger DDL single-home ratchet and frozen product tables.
• Expected review time: 4 days

Impact note (optional): Breaking change — removes pre-store persistence layer; all downstream consumers must read exclusively through StoreClient.