v0.5.0
v0.5.0
This release turns the project into a Dockerized Obsidian warehouse simulation that can be run from a fresh clone.
Highlights
- Added lifecycle-aware synthetic vault data with realistic source and observed timestamps.
- Added deterministic synthetic vault generation and virtual-day vault simulation.
- Added Docker Compose simulation services for seed generation, live vault browsing, Airflow, live web UI, and live MCP.
- Added an Airflow DAG that advances virtual time and runs ingest, dbt run, and dbt test once per minute.
- Added DuckDB landing tables and dbt medallion models over parsed Obsidian content.
- Added incremental context marts for people, companies, projects, mentions, decisions, risks, open loops, person context, and project context.
- Improved local Docker startup reliability for fresh named volumes.
- Fixed entity matching so similarly named projects such as
Project Atlas 1andProject Atlas 16do not leak into each other's query results.
Verification
- Fresh clone Docker simulation smoke test passed.
- Airflow DAG completed successfully.
- dbt run passed with 25 models.
- dbt test passed with 90 tests.
- Python test suite passed with 34 tests.