Phase 1 project prototype

src/admin_analytics/cli.py

@@ -107,10 +107,83 @@ def irs990(
 
 
 @ingest_app.command()
-def cpi() -> None:
-    """Ingest BLS CPI-U data. (Not yet implemented.)"""
-    typer.echo("CPI-U ingestion is not yet implemented (Sprint 3).")
-    raise typer.Exit(1)
+def cpi(
+    force: Annotated[
+        bool, typer.Option("--force", help="Re-download even if file exists")
+    ] = False,
+) -> None:
+    """Ingest BLS CPI-U data."""
+    conn = get_connection()
+    ensure_schema(conn)
+
+    from admin_analytics.bls.download import download_cpi_file
+    from admin_analytics.bls.loader import load_cpi
+
+    typer.echo("Downloading BLS CPI-U data...")
+    file_path = download_cpi_file(force=force)
+
+    typer.echo("Loading CPI-U data into database...")
+    count = load_cpi(conn, file_path)
+    typer.echo(f"CPI-U ingestion complete: {count} monthly observations loaded.")
+    conn.close()
+
+
+@ingest_app.command()
+def scrape() -> None:
+    """Scrape UD staff directory pages for admin headcounts."""
+    conn = get_connection()
+    ensure_schema(conn)
+
+    from admin_analytics.scraper.directory import scrape_all
+    from admin_analytics.scraper.loader import load_scrape
+    from admin_analytics.scraper.classify import OVERHEAD_CATEGORIES, NON_OVERHEAD_CATEGORIES
+
+    typer.echo("Scraping UD staff directory pages...")
+    entries = scrape_all()
+
+    typer.echo("Loading scraped data into database...")
+    count = load_scrape(conn, entries)
+
+    # Summary by unit and category
+    typer.echo(f"\nLoaded {count} staff entries.\n")
+    unit_counts: dict[str, dict[str, int]] = {}
+    for e in entries:
+        unit_counts.setdefault(e.unit, {})
+        unit_counts[e.unit][e.category] = unit_counts[e.unit].get(e.category, 0) + 1
+
+    for unit, cats in sorted(unit_counts.items()):
+        total = sum(cats.values())
+        overhead = sum(v for k, v in cats.items() if k in OVERHEAD_CATEGORIES)
+        typer.echo(f"  {unit}: {total} staff ({overhead} overhead)")
+        for cat, n in sorted(cats.items(), key=lambda x: -x[1]):
+            typer.echo(f"    {cat}: {n}")
+
+    conn.close()
+
+
+@app.command()
+def dashboard(
+    port: Annotated[int, typer.Option(help="Port to serve on")] = 8050,
+    debug: Annotated[bool, typer.Option(help="Enable Dash debug mode")] = True,
+) -> None:
+    """Launch the analytics dashboard."""
+    from admin_analytics.dashboard.app import create_app
+
+    dash_app = create_app()
+    typer.echo(f"Starting dashboard at http://localhost:{port}/")
+    dash_app.run(debug=debug, port=port)
+
+
+@app.command()
+def validate() -> None:
+    """Run data validation checks and print a report."""
+    conn = get_connection()
+    ensure_schema(conn)
+
+    from admin_analytics.validation import format_report
+
+    typer.echo(format_report(conn))
+    conn.close()
 
 
 @ingest_app.command(name="all")
@@ -124,3 +197,6 @@ def ingest_all(
 ) -> None:
     """Ingest all data sources."""
     ipeds(year_range=year_range, component="all", force=force)
+    irs990(year_range=year_range, force=force)
+    cpi(force=force)
+    scrape()