feat: AR-House initial commit

2026-07-03 12:24:58 -04:00
commit 047c05287a
216 changed files with 127552 additions and 0 deletions
@@ -0,0 +1,82 @@
+"""Full HUD pipeline: scrape FL → persist → classify."""
+from __future__ import annotations
+import io, sys, time
+from pathlib import Path
+
+sys.stdout = io.TextIOWrapper(sys.stdout.buffer, encoding="utf-8", errors="replace")
+ROOT = Path(__file__).resolve().parent.parent
+sys.path.insert(0, str(ROOT))
+
+
+def main() -> int:
+    print("=" * 70)
+    print("HUD Homestore FULL PIPELINE (FL only)")
+    print("=" * 70)
+
+    from scrapers.hud_homestore import run_scraper_to_db
+    from deals_db import init_db, list_deals
+    init_db()
+
+    def log(m): print(f"  {m}")
+
+    t0 = time.perf_counter()
+    summary = run_scraper_to_db(states=["FL"], auto_classify=True, status_cb=log)
+    elapsed = time.perf_counter() - t0
+
+    print()
+    print("=" * 70)
+    print(f"PIPELINE SUMMARY (elapsed {elapsed:.0f}s = {elapsed/60:.1f} min)")
+    print("=" * 70)
+    for k, v in summary.items():
+        if k == "errors":
+            print(f"  {k}: ({len(v)} items)")
+            for e in v[:3]:
+                print(f"    - {e}")
+        else:
+            print(f"  {k}: {v}")
+
+    # Show classification breakdown for HUD source
+    print()
+    print("--- HUD deals by classification ---")
+    hud = list_deals(source="hud_homestore", limit=200)
+    by_class = {}
+    for d in hud:
+        cs = d.get("classification_status") or "(unclassified)"
+        by_class[cs] = by_class.get(cs, 0) + 1
+    for cs in sorted(by_class.keys()):
+        print(f"  {cs}: {by_class[cs]}")
+
+    print()
+    print("--- TOP 10 HUD deals by classification_score ---")
+    top = sorted(hud, key=lambda d: (d.get("classification_score") or 0), reverse=True)[:10]
+    print(f"{'#':<3} {'Score':<6} {'Status':<20} {'Strategy':<14} {'Price':<10} {'Beds':<5} Address (county)")
+    print("-" * 130)
+    import json as _json
+    for i, d in enumerate(top, 1):
+        score = d.get("classification_score") or 0
+        cls = d.get("classification_status") or "?"
+        strat = d.get("classification_strategy") or "?"
+        price = d.get("listing_price")
+        price_str = f"${price:,.0f}" if price else "N/A"
+        beds = d.get("beds")
+        addr = (d.get("address") or "?")[:70]
+        county = d.get("county") or ""
+        print(f"{i:<3} {score:<6} {cls:<20} {strat:<14} {price_str:<10} {beds!s:<5} {addr} ({county})")
+
+    # Print 3 sample reasons
+    print()
+    print("--- Sample reasons (top 3) ---")
+    for i, d in enumerate(top[:3], 1):
+        print(f"\n  [{i}] {d.get('case_number')} — {d.get('classification_status')} score {d.get('classification_score')}")
+        try:
+            reasons = _json.loads(d.get("classification_reasons") or "[]")
+            for r in reasons:
+                print(f"     - {r}")
+        except Exception:
+            pass
+
+    return 0
+
+
+if __name__ == "__main__":
+    sys.exit(main())