fine

finder/storage.py

@@ -10,8 +10,16 @@ from transform import map_property_type, normalize_postcode
 log = logging.getLogger("rightmove")
 
 
-def write_parquet(properties: list[dict], path: Path) -> None:
-    """Write sale properties list to parquet with server-ready column names."""
+def write_parquet(
+    properties: list[dict], path: Path, price_history: dict | None = None
+) -> None:
+    """Write sale properties list to parquet with server-ready column names.
+
+    ``price_history`` is the persistent listing-id -> [{date, price, reason}]
+    store (see finder/price_history.py); each listing's accrued asking-price
+    series is embedded as a ``price_history`` list column. Absent id -> empty
+    list, so pre-instrumentation runs and tests simply write empty series.
+    """
     if not properties:
         log.warning("No properties to write to %s", path)
         return
@@ -95,6 +103,14 @@ def write_parquet(properties: list[dict], path: Path) -> None:
     asking_prices = [p["price"] if p["price"] > 0 else None for p in properties]
     listing_statuses = ["For sale"] * len(properties)
 
+    # Accrued asking-price history per listing (oldest -> newest). Look up with
+    # the SAME normalisation the store keys on (str(id).strip(), matching
+    # price_history.update_history and scraper dedup); an unseen id -> empty.
+    history_lookup = price_history or {}
+    price_history_col = [
+        history_lookup.get(str(p.get("id")).strip(), []) for p in properties
+    ]
+
     df = pl.DataFrame(
         {
             "Bedrooms": [p["Bedrooms"] for p in properties],
@@ -144,6 +160,7 @@ def write_parquet(properties: list[dict], path: Path) -> None:
             "Listing date": listing_dates,
             "Listing status": listing_statuses,
             "Asking price": asking_prices,
+            "price_history": price_history_col,
         },
         schema={
             "Bedrooms": pl.Int32,
@@ -169,6 +186,11 @@ def write_parquet(properties: list[dict], path: Path) -> None:
             "Listing date": pl.Datetime("us"),
             "Listing status": pl.Utf8,
             "Asking price": pl.Int64,
+            "price_history": pl.List(
+                pl.Struct(
+                    {"date": pl.Utf8, "price": pl.Int64, "reason": pl.Utf8}
+                )
+            ),
         },
     )