This commit is contained in:
Andras Schmelczer 2026-03-15 21:22:28 +00:00
parent 479ef92236
commit c38d654ac7
44 changed files with 2526 additions and 701 deletions

View file

@ -6,7 +6,16 @@ from dataclasses import dataclass, field
import polars as pl
from constants import ARCGIS_PATH, CHANNELS, DATA_DIR, DELAY_BETWEEN_OUTCODES, SCRAPE_HOMECOUK, SCRAPE_OPENRENT, SCRAPE_RIGHTMOVE, SEED
from constants import (
ARCGIS_PATH,
CHANNELS,
DATA_DIR,
DELAY_BETWEEN_OUTCODES,
SCRAPE_HOMECOUK,
SCRAPE_OPENRENT,
SCRAPE_RIGHTMOVE,
SEED,
)
from homecouk import CookiesExpiredError
from homecouk import load_cookies as load_homecouk_cookies
from homecouk import make_client as make_homecouk_client
@ -64,13 +73,23 @@ def _sync_gauges() -> None:
scrape_outcodes_done.set(status.outcodes_done)
scrape_outcodes_total.set(status.outcodes_total)
# Total properties (both sources combined)
scrape_properties_total.labels(channel="buy", source="total").set(status.properties_buy)
scrape_properties_total.labels(channel="rent", source="total").set(status.properties_rent)
scrape_properties_total.labels(channel="buy", source="total").set(
status.properties_buy
)
scrape_properties_total.labels(channel="rent", source="total").set(
status.properties_rent
)
# Per-source breakdown for current channel
ch = "buy" if status.channel == "BUY" else "rent"
scrape_properties_total.labels(channel=ch, source="rightmove").set(status.rm_properties)
scrape_properties_total.labels(channel=ch, source="homecouk").set(status.hk_properties)
scrape_properties_total.labels(channel=ch, source="openrent").set(status.or_properties)
scrape_properties_total.labels(channel=ch, source="rightmove").set(
status.rm_properties
)
scrape_properties_total.labels(channel=ch, source="homecouk").set(
status.hk_properties
)
scrape_properties_total.labels(channel=ch, source="openrent").set(
status.or_properties
)
if status.started_at:
end = status.finished_at if status.finished_at else time.time()
scrape_elapsed_seconds.set(end - status.started_at)
@ -86,7 +105,9 @@ def load_outcodes() -> list[str]:
log.info("England postcodes: %d", len(england))
outcodes = (
england.select(pl.col("pcd").str.extract(r"^([A-Z]{1,2}\d[A-Z0-9]?)", 1).alias("outcode"))
england.select(
pl.col("pcd").str.extract(r"^([A-Z]{1,2}\d[A-Z0-9]?)", 1).alias("outcode")
)
.drop_nulls()
.get_column("outcode")
.unique()
@ -101,7 +122,9 @@ def build_postcode_index() -> PostcodeSpatialIndex:
"""Build spatial index from arcgis England postcodes."""
log.info("Building postcode spatial index from %s", ARCGIS_PATH)
df = pl.read_parquet(ARCGIS_PATH, columns=["pcd", "ctry", "lat", "long"])
england = df.filter(pl.col("ctry") == "E92000001").drop_nulls(subset=["lat", "long"])
england = df.filter(pl.col("ctry") == "E92000001").drop_nulls(
subset=["lat", "long"]
)
return PostcodeSpatialIndex(
england.get_column("lat").to_list(),
england.get_column("long").to_list(),
@ -114,7 +137,9 @@ def build_postcode_coords() -> dict[str, tuple[float, float]]:
Used by OpenRent scraper to resolve coordinates from postcodes."""
log.info("Building postcode coords lookup from %s", ARCGIS_PATH)
df = pl.read_parquet(ARCGIS_PATH, columns=["pcd", "ctry", "lat", "long"])
england = df.filter(pl.col("ctry") == "E92000001").drop_nulls(subset=["lat", "long"])
england = df.filter(pl.col("ctry") == "E92000001").drop_nulls(
subset=["lat", "long"]
)
coords: dict[str, tuple[float, float]] = {}
for pcd, lat, lng in zip(
england.get_column("pcd").to_list(),
@ -179,7 +204,9 @@ def run_scrape(
log.info("home.co.uk scraping ENABLED")
homecouk_enabled.set(1)
else:
log.info("home.co.uk scraping DISABLED (need FlareSolverr or HOMECOUK_CF_CLEARANCE + HOMECOUK_SESSION)")
log.info(
"home.co.uk scraping DISABLED (need FlareSolverr or HOMECOUK_CF_CLEARANCE + HOMECOUK_SESSION)"
)
homecouk_enabled.set(0)
# OpenRent: must be enabled via SCRAPE_OPENRENT + cookies available
@ -195,7 +222,9 @@ def run_scrape(
log.info("OpenRent scraping ENABLED")
openrent_enabled.set(1)
else:
log.info("OpenRent scraping DISABLED (need FlareSolverr or OPENRENT_WAF_TOKEN)")
log.info(
"OpenRent scraping DISABLED (need FlareSolverr or OPENRENT_WAF_TOKEN)"
)
openrent_enabled.set(0)
# Build postcode coords if OpenRent is active and caller didn't provide them
@ -207,7 +236,9 @@ def run_scrape(
channel_name = channel_cfg["channel"]
file_suffix = "buy" if channel_name == "BUY" else "rent"
all_properties: dict[str, dict] = {} # dedup by id
seen_dedup_keys: set[tuple] = set() # cross-source dedup by (postcode, beds, price)
seen_dedup_keys: set[tuple] = (
set()
) # cross-source dedup by (postcode, beds, price)
rm_count = 0 # Rightmove properties this channel
hk_count = 0 # home.co.uk properties this channel
hk_dedup_count = 0 # home.co.uk skipped as cross-source duplicates
@ -222,15 +253,22 @@ def run_scrape(
status.hk_properties = 0
status.or_properties = 0
log.info("=== Starting %s channel (%d outcodes) ===", channel_name, len(shuffled))
log.info(
"=== Starting %s channel (%d outcodes) ===", channel_name, len(shuffled)
)
for i, outcode in enumerate(shuffled):
with status_lock:
status.outcode = outcode
status.outcodes_done = i
log.debug("Outcode %s (%d/%d) — %d properties so far",
outcode, i + 1, len(shuffled), len(all_properties))
log.debug(
"Outcode %s (%d/%d) — %d properties so far",
outcode,
i + 1,
len(shuffled),
len(all_properties),
)
made_requests = False
@ -240,9 +278,13 @@ def run_scrape(
try:
outcode_id = resolve_outcode_id(client, outcode)
if not outcode_id:
log.debug("No Rightmove ID for outcode %s, skipping", outcode)
log.debug(
"No Rightmove ID for outcode %s, skipping", outcode
)
else:
props = search_outcode(client, outcode_id, outcode, channel_cfg, pc_index)
props = search_outcode(
client, outcode_id, outcode, channel_cfg, pc_index
)
for p in props:
pid = p["id"]
if pid not in all_properties:
@ -261,7 +303,10 @@ def run_scrape(
made_requests = True
try:
hk_props = homecouk_search_outcode(
hk_client, outcode, channel_name, pc_index,
hk_client,
outcode,
channel_name,
pc_index,
)
for p in hk_props:
pid = p["id"]
@ -276,9 +321,13 @@ def run_scrape(
seen_dedup_keys.add(key)
hk_count += 1
if hk_props:
log.info("home.co.uk %s: +%d properties", outcode, len(hk_props))
log.info(
"home.co.uk %s: +%d properties", outcode, len(hk_props)
)
except CookiesExpiredError:
log.warning("home.co.uk cookies expired — attempting refresh via FlareSolverr")
log.warning(
"home.co.uk cookies expired — attempting refresh via FlareSolverr"
)
hk_client.close()
hk_result = load_homecouk_cookies()
if hk_result:
@ -286,13 +335,17 @@ def run_scrape(
log.info("home.co.uk cookies refreshed, continuing")
cookie_refreshes_total.labels(result="success").inc()
else:
log.warning("Cookie refresh failed, disabling home.co.uk for rest of scrape")
log.warning(
"Cookie refresh failed, disabling home.co.uk for rest of scrape"
)
hk_client = None
hk_failed = True
homecouk_enabled.set(0)
cookie_refreshes_total.labels(result="failure").inc()
with status_lock:
status.errors.append("home.co.uk cookies expired and refresh failed")
status.errors.append(
"home.co.uk cookies expired and refresh failed"
)
except Exception as e:
msg = f"Error scraping home.co.uk {outcode}/{channel_name}: {e}"
log.error(msg)
@ -305,7 +358,10 @@ def run_scrape(
made_requests = True
try:
or_props = openrent_search_outcode(
or_client, outcode, pc_index, pc_coords,
or_client,
outcode,
pc_index,
pc_coords,
)
for p in or_props:
pid = p["id"]
@ -318,9 +374,13 @@ def run_scrape(
seen_dedup_keys.add(key)
or_count += 1
if or_props:
log.info("OpenRent %s: +%d properties", outcode, len(or_props))
log.info(
"OpenRent %s: +%d properties", outcode, len(or_props)
)
except WafChallengeError:
log.warning("OpenRent WAF cookies expired — attempting refresh via FlareSolverr")
log.warning(
"OpenRent WAF cookies expired — attempting refresh via FlareSolverr"
)
or_client.close()
or_result = load_openrent_cookies()
if or_result:
@ -328,13 +388,17 @@ def run_scrape(
log.info("OpenRent cookies refreshed, continuing")
cookie_refreshes_total.labels(result="success").inc()
else:
log.warning("Cookie refresh failed, disabling OpenRent for rest of scrape")
log.warning(
"Cookie refresh failed, disabling OpenRent for rest of scrape"
)
or_client = None
or_failed = True
openrent_enabled.set(0)
cookie_refreshes_total.labels(result="failure").inc()
with status_lock:
status.errors.append("OpenRent WAF cookies expired and refresh failed")
status.errors.append(
"OpenRent WAF cookies expired and refresh failed"
)
except Exception as e:
msg = f"Error scraping OpenRent {outcode}/{channel_name}: {e}"
log.error(msg)
@ -352,8 +416,14 @@ def run_scrape(
status.or_properties = or_count
_sync_gauges()
log.info("Outcode %s: total %d (rm: %d, hk: %d, or: %d)",
outcode, len(all_properties), rm_count, hk_count, or_count)
log.info(
"Outcode %s: total %d (rm: %d, hk: %d, or: %d)",
outcode,
len(all_properties),
rm_count,
hk_count,
or_count,
)
if made_requests and i < len(shuffled) - 1:
time.sleep(DELAY_BETWEEN_OUTCODES)
@ -373,7 +443,11 @@ def run_scrape(
log.info(
"=== %s channel complete: %d unique (rm: %d, hk: %d, or: %d, cross-dedup: %d) ===",
channel_name, len(deduped), rm_count, hk_count, or_count,
channel_name,
len(deduped),
rm_count,
hk_count,
or_count,
hk_dedup_count + or_dedup_count,
)
@ -382,8 +456,12 @@ def run_scrape(
status.finished_at = time.time()
_sync_gauges()
elapsed = status.finished_at - status.started_at
log.info("Scrape complete in %.0fs — buy: %d, rent: %d",
elapsed, status.properties_buy, status.properties_rent)
log.info(
"Scrape complete in %.0fs — buy: %d, rent: %d",
elapsed,
status.properties_buy,
status.properties_rent,
)
except Exception as e:
log.exception("Fatal scrape error")