Add legacy database structural reference for non-Windows machines

docs/LEGACY_DATABASES.md documents all three source Access databases
(every table, column, type, and known data-quality quirk) generated
from a live read of the real files, so no Windows/Access driver is
needed to understand their structure going forward.

New migration/ tooling: catalog_schema.py connects to the real files
and walks every table (including excluded scratch tables);
render_catalog_md.py renders that into the doc's appendix. Raw output
checked in at migration/catalog.json so the doc can be regenerated
without touching Access again.

migration/render_catalog_md.py

@@ -0,0 +1,98 @@
+"""
+Renders catalog.json (from catalog_schema.py) into a Markdown appendix:
+one section per source file, one subsection per table, with a column
+table. Excluded (scratch/template/report) tables are marked and collapsed
+to a one-liner rather than a full column dump, since they carry no data
+worth documenting column-by-column.
+
+Usage:
+    python render_catalog_md.py catalog.json --output appendix.md
+"""
+
+from __future__ import annotations
+
+import argparse
+import json
+from pathlib import Path
+
+SOURCE_TITLES = {
+    "utilities": "UTILITIES.accdb",
+    "seguros": "SEGUROS 16_be.mdb",
+    "scothia": "SCOTHIA.mdb",
+}
+
+
+def render_table(name: str, entry: dict) -> str:
+    lines = [f"#### `{name}`"]
+
+    if entry.get("excluded"):
+        lines.append("")
+        lines.append(f"**Excluded from migration** — {entry.get('row_count', '?')} rows. Not detailed here; see the exclusion rationale table above.")
+        lines.append("")
+        return "\n".join(lines)
+
+    if "error" in entry:
+        lines.append("")
+        lines.append(f"**Could not read this table:** `{entry['error']}`")
+        lines.append("")
+        return "\n".join(lines)
+
+    lines.append("")
+    lines.append(f"Rows: {entry['row_count']} | Columns: {entry['column_count']}")
+    lines.append("")
+    lines.append("| Column | Type | Nullable |")
+    lines.append("|---|---|---|")
+    for col in entry["columns"]:
+        type_label = col["access_type"] or f"_(inferred: {col['pandas_dtype']})_"
+        nullable = "yes" if col["nullable"] else ("no" if col["nullable"] is False else "—")
+        col_name = col["name"].replace("|", "\\|")
+        lines.append(f"| `{col_name}` | {type_label} | {nullable} |")
+    lines.append("")
+    return "\n".join(lines)
+
+
+def render_source(source_name: str, source_data: dict) -> str:
+    title = SOURCE_TITLES.get(source_name, source_name)
+    lines = [f"### {title}", ""]
+
+    tables = source_data["tables"]
+    included = sorted(t for t, e in tables.items() if not e.get("excluded"))
+    excluded = sorted(t for t, e in tables.items() if e.get("excluded"))
+
+    lines.append(f"{len(tables)} tables total — {len(included)} included in migration, {len(excluded)} excluded (scratch/template/report tables).")
+    lines.append("")
+
+    for table_name in included:
+        lines.append(render_table(table_name, tables[table_name]))
+
+    if excluded:
+        lines.append("#### Excluded tables")
+        lines.append("")
+        lines.append("| Table | Rows |")
+        lines.append("|---|---|")
+        for table_name in excluded:
+            entry = tables[table_name]
+            lines.append(f"| `{table_name}` | {entry.get('row_count', '?')} |")
+        lines.append("")
+
+    return "\n".join(lines)
+
+
+def main() -> None:
+    parser = argparse.ArgumentParser(description=__doc__)
+    parser.add_argument("catalog", type=Path)
+    parser.add_argument("--output", type=Path, default=Path("appendix.md"))
+    args = parser.parse_args()
+
+    catalog = json.loads(args.catalog.read_text(encoding="utf-8"))
+
+    sections = ["## Appendix: full table catalog", "", "Generated by `migration/catalog_schema.py` + `migration/render_catalog_md.py`. Regenerate after any change to the source files.", ""]
+    for source_name in ["utilities", "seguros", "scothia"]:
+        sections.append(render_source(source_name, catalog[source_name]))
+
+    args.output.write_text("\n".join(sections), encoding="utf-8")
+    print(f"Wrote {args.output}")
+
+
+if __name__ == "__main__":
+    main()