reuse query params when exporting to immoweb and allow filtering from available date

crawler/1_dump_listings.py

@@ -1,29 +1,11 @@
 import asyncio
 from dataclasses import dataclass
 import pathlib
-from rec.query import ListingType, listing_query, FurnishType
+from rec.query import listing_query, QueryParameters
 from rec.districts import get_districts
 from data_access import Listing
 
 
-@dataclass(frozen=True)
-class QueryParameters:
-    listing_type: ListingType
-    min_bedrooms: int
-    max_bedrooms: int
-    min_price: int
-    max_price: int
-    district_names: set[str]
-    radius: float = 0
-    page_size: int = 500  # items per page
-    max_days_since_added: int = 30
-    furnish_types: list[FurnishType] | None = None
-
-    # The values below are not supported by rightmove
-    # hence we apply them after fetching
-    # available from; council tax
-
-
 async def dump_listings(
     parameters: QueryParameters,
     data_dir: pathlib.Path = pathlib.Path("data/rs/"),

crawler/data_access.py

@@ -272,10 +272,7 @@ class Listing:
         return (now - ds).days
 
     @property
-    def last_seen(self) -> int | None:
-        if not self.path_last_seen_listing().exists():
-            return None
-
+    def last_seen(self) -> int:
         with open(self.path_last_seen_listing(), "r") as f:
             datetime_str = json.load(f)
             dt = datetime.datetime.fromisoformat(datetime_str)

crawler/main.py

@@ -1,4 +1,5 @@
 import asyncio
+from datetime import datetime
 import json
 import os
 import pathlib
@@ -8,7 +9,7 @@ import importlib
 from rec.districts import get_districts
 from data_access import Listing
 import csv_exporter
-from rec.query import ListingType, FurnishType
+from rec.query import ListingType, FurnishType, QueryParameters
 from rec.routing import API_KEY_ENVIRONMENT_VARIABLE, TravelMode
 from ui_exporter import export_immoweb as export_immoweb_ui
 from functools import wraps
@@ -75,6 +76,12 @@ def listing_filter_options(func):
         ),
         multiple=True,
     )
+    @click.option(
+        "--available-from",
+        help="Let date available from",
+        default=None,
+        type=click.DateTime(),
+    )
     @wraps(func)
     def wrapper(*args, **kwargs):
         return func(*args, **kwargs)
@@ -113,9 +120,10 @@ def dump_listings(
     max_price: int,
     type: str,
     furnish_types: list[str],
+    available_from: datetime | None,
 ):
     data_dir: str = ctx.obj["data_dir"]
-    query_parameters = dump_listings_module.QueryParameters(
+    query_parameters = QueryParameters(
         listing_type=ListingType[type],
         district_names=set(district),
         min_bedrooms=min_bedrooms,
@@ -123,6 +131,7 @@ def dump_listings(
         min_price=min_price,
         max_price=max_price,
         furnish_types=[FurnishType[furnish_type] for furnish_type in furnish_types],
+        let_date_available_from=available_from,
     )
     click.echo(
         f"Running dump_listings for districts {district}, data dir {data_dir} and parameters: "
@@ -257,10 +266,34 @@ def export_csv(ctx: click.core.Context, output_file: str, columns: tuple[str]):
         resolve_path=True,
     ),
 )
+@listing_filter_options
 @click.pass_context
-def export_immoweb(ctx, output_file: str):
-    click.echo(f"Exporting data to {output_file}")
-    asyncio.run(export_immoweb_ui(ctx, output_file))
+def export_immoweb(
+    ctx,
+    output_file: str,
+    district: list[str],
+    min_bedrooms: int,
+    max_bedrooms: int,
+    min_price: int,
+    max_price: int,
+    type: str,
+    furnish_types: list[str],
+    available_from: datetime | None,
+):
+    query_parameters = QueryParameters(
+        listing_type=ListingType[type],
+        district_names=set(district),
+        min_bedrooms=min_bedrooms,
+        max_bedrooms=max_bedrooms,
+        min_price=min_price,
+        max_price=max_price,
+        furnish_types=[FurnishType[furnish_type] for furnish_type in furnish_types],
+        let_date_available_from=available_from,
+    )
+    click.echo(
+        f"Exporting data to {output_file} that matches the query parameters: {query_parameters}"
+    )
+    asyncio.run(export_immoweb_ui(ctx, output_file, query_parameters))
 
 
 if __name__ == "__main__":

crawler/rec/query.py

@@ -1,8 +1,11 @@
 # from diskcache import Cache
 import asyncio
+from dataclasses import dataclass
+from datetime import datetime
 import enum
 from typing import Any
 import aiohttp
+from data_access import Listing
 
 
 class ListingType(enum.StrEnum):
@@ -16,6 +19,57 @@ class FurnishType(enum.StrEnum):
     PART_FURNISHED = "partFurnished"
 
 
+@dataclass(frozen=True)
+class QueryParameters:
+    listing_type: ListingType
+    min_bedrooms: int
+    max_bedrooms: int
+    min_price: int
+    max_price: int
+    district_names: set[str]
+    radius: float = 0
+    page_size: int = 500  # items per page
+    max_days_since_added: int = 30
+    furnish_types: list[FurnishType] | None = None
+    # The values below are not supported by rightmove
+    # hence we apply them after fetching
+    # available from; council tax
+    let_date_available_from: datetime | None = None
+    last_seen_days: int = 14
+
+
+async def filter_listings(
+    listings: list[Listing],
+    query_parameters: QueryParameters,
+) -> list[Listing]:
+    """
+    Filter listings based on the provided query parameters.
+    """
+    filtered_listings = []
+    for listing in listings:
+        if (
+            listing.bedrooms > query_parameters.max_bedrooms
+            or listing.bedrooms < query_parameters.min_bedrooms
+        ):
+            continue
+        if (
+            listing.price < query_parameters.min_price
+            or listing.price > query_parameters.max_price
+        ):
+            continue
+        if listing.last_seen > query_parameters.last_seen_days:
+            continue
+        if (
+            listing.letDateAvailable is not None
+            and query_parameters.let_date_available_from is not None
+            and listing.letDateAvailable < query_parameters.let_date_available_from
+        ):
+            continue
+        filtered_listings.append(listing)
+
+    return filtered_listings
+
+
 headers = {
     "Host": "api.rightmove.co.uk",
     # 'Accept-Encoding': 'gzip, deflate, br',

crawler/ui_exporter.py

@@ -1,47 +1,53 @@
-
 import json
 import pathlib
 
 from data_access import Listing
+from rec.query import QueryParameters, filter_listings
 
 
-async def export_immoweb(ctx, output_file: str):
-    data_dir = ctx.obj['data_dir']
-    output_file_path = pathlib.Path(output_file) 
+async def export_immoweb(
+    ctx,
+    output_file: str,
+    query_parameters: QueryParameters | None = None,
+):
+    data_dir = ctx.obj["data_dir"]
+    output_file_path = pathlib.Path(output_file)
     listing_paths = sorted(list(pathlib.Path(data_dir).glob("*/listing.json")))
     # listing_paths = listing_paths[:10]
     listings = Listing.get_all_listings([str(path) for path in listing_paths])
-    
+    if query_parameters is not None:
+        listings = await filter_listings(listings, query_parameters)
+
     # Convert listings to immoweb format
     immoweb_listings = []
     for listing in listings:
         immoweb_listing = {
-            'type': 'Feature',
-            'properties': {
-                'city': 'London', # change me
-                'country': 'United Kingdom',
-                'qm': await listing.sqm_ocr(),
-                'qmprice': round(await listing.price_per_sqm(), 2),
-                'rooms': listing.bedrooms,
-                'total_price': listing.price,
-                'url': listing.url,
+            "type": "Feature",
+            "properties": {
+                "city": "London",  # change me
+                "country": "United Kingdom",
+                "qm": await listing.sqm_ocr(),
+                "qmprice": round(await listing.price_per_sqm(), 2),
+                "rooms": listing.bedrooms,
+                "total_price": listing.price,
+                "url": listing.url,
                 # Additional info; the above is GeoJSON format
                 # Below is all other crap we want in the UI
-                'info': await listing.dict_nicely(),
+                "info": await listing.dict_nicely(),
             },
-            'geometry': {
-                'coordinates': [
+            "geometry": {
+                "coordinates": [
                     listing.longitude,
                     listing.latitude,
                 ],
-                'type': 'Point',
-            }
+                "type": "Point",
+            },
         }
         immoweb_listings.append(immoweb_listing)
-    
-    prefix = 'var data = '
+
+    prefix = "var data = "
     serialized_data = {"type": "FeatureCollection", "features": immoweb_listings}
     result = prefix + json.dumps(serialized_data, indent=4)
-    with open(output_file_path, 'w') as f:
+    with open(output_file_path, "w") as f:
         f.write(result)
         # json.dump(serialized_data, f, indent=4)