Vlad — Soviet Statistical Corpus: Robert-DB + corpus-inventory bundle
=====================================================================

Every file here is REAL metadata from the Vlad (USSR) project. NOTHING is
fabricated, and NO harmonized economic time series are included -- because none
are built yet. Displayed source-document titles are translated into English with
the Russian original preserved.

Part 1 — Robert-database ingestion (the table-level catalog)
------------------------------------------------------------
The validated knowledge base was ingested by the Robert Database Framework v1.0
into a canonical per-project SQLite database. These files are the honest catalog
of WHAT WAS EXTRACTED (per-table metadata -- NOT the cell/observation values):

  vlad_tables_catalog.csv          one row per agent-enriched (Tier-A/B) table:
                                     title (EN + RU), units, geography, period,
                                     start_year, topic + geo facet (taxonomy),
                                     two-axis quality (obs_status x V/H/L/R/X),
                                     confidence, dimensions, source document.
  vlad_tables_catalog.parquet      same, as Apache Parquet.
  vlad_concordances_catalog.csv    175 curated year-series families (linked
                                     tables across years -> would-be panels).
  vlad_agg_tables_by_topic.csv     enriched-table counts by taxonomy topic.
  vlad_agg_tables_by_decade.csv    enriched-table counts by start decade.
  vlad_agg_tables_by_geography.csv enriched-table counts by geographic level.
  vlad_agg_quality_two_axis.csv    the Robert two-axis quality cross-tab.

Part 2 — Corpus inventory (per-cluster measurements)
----------------------------------------------------
  anu_series_inventory.csv      full 266-cluster inventory (one row per cluster)
  vlad_panels.csv             panel-capable clusters (flat CSV)
  vlad_richest.csv            richest-by-tables clusters (flat CSV)
  vlad_project_totals.csv     every headline count + an explicit scope tag
  vlad_anu_clusters.csv       the 5 flagship Anu targets (A-E) -- SCOPE/ROADMAP

What is NOT here (because it is not built)
------------------------------------------
  No harmonized economic SERIES. No observation/cell VALUES. The Anu
  data-construction pipeline that would turn the panel-capable families into
  year x category x value panels is SCOPED, not constructed. The Cluster A fiscal
  pilot (1856-1990) is first in line. See ussr.heterodata.org/methodology.

Also not shipped: the 4.4 GB HDARP knowledge base and ~33 GB of outputs. Those
live in the project repository, not this compact showcase.

Sources: the canonical USSR Robert database (Robert Database Framework v1.0,
robertdb.sqlite), the reconciled campaign-status record (2026-06-07), the Anu
pipeline scoping document, and ANU_SERIES_INVENTORY.csv.
