353 lines
12 KiB
Python
353 lines
12 KiB
Python
#!/usr/bin/env python3
|
||
"""
|
||
Scraper ArgenProp.com.ar – Bienes Raices Argentina
|
||
Estrategia: iterar URLs del sitemap de listing (por barrio/tipo) y extraer
|
||
listing cards con datos SSR completos en atributos del <a class="card">.
|
||
DB: argenprop_ar
|
||
Checkpoint: /opt/scrapers/argenprop_ar/state.json
|
||
{"sitemap_idx": N, "url_idx": M, "total_new": X}
|
||
"""
|
||
import os
|
||
import re
|
||
import time
|
||
import json
|
||
import gzip
|
||
import logging
|
||
from datetime import datetime
|
||
|
||
import requests
|
||
from bs4 import BeautifulSoup
|
||
import psycopg2
|
||
from psycopg2.extras import execute_values
|
||
|
||
logging.basicConfig(
|
||
level=logging.INFO,
|
||
format='%(asctime)s %(levelname)s %(message)s',
|
||
datefmt='%Y-%m-%d %H:%M:%S',
|
||
)
|
||
log = logging.getLogger('argenprop')
|
||
|
||
PROXY = 'socks5h://127.0.0.1:1090'
|
||
PROXIES = {'http': PROXY, 'https': PROXY}
|
||
UA = 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 Chrome/124 Safari/537.36'
|
||
BASE = 'https://www.argenprop.com'
|
||
DELAY = 1.5
|
||
|
||
DB_HOST = '100.75.240.87'
|
||
DB_PORT = 5432
|
||
DB_NAME = 'argenprop_ar'
|
||
DB_USER = 'pgadmin'
|
||
DB_PASS = 'J5BVlq65JvedWxZVrcY96OQX'
|
||
|
||
DAILY = bool(os.environ.get('ARGENPROP_DAILY')) # True en modo daily, False en backfill
|
||
MAX_SITEMAPS = int(os.environ.get('MAX_SITEMAPS', '99')) # cuantos sub-sitemaps procesar
|
||
|
||
STATE_FILE = os.path.join(os.path.dirname(os.path.abspath(__file__)), 'state.json')
|
||
|
||
# Sub-sitemaps de listing (se descubren dinamicamente desde el índice)
|
||
SITEMAP_INDEX = 'https://www.argenprop.com/sitemap.xml'
|
||
|
||
|
||
def load_state():
|
||
if os.path.exists(STATE_FILE):
|
||
with open(STATE_FILE) as f:
|
||
st = json.load(f)
|
||
log.info('Checkpoint: sitemap_idx=%d url_idx=%d total_new=%d',
|
||
st['sitemap_idx'], st['url_idx'], st.get('total_new', 0))
|
||
return st
|
||
return {'sitemap_idx': 0, 'url_idx': 0, 'total_new': 0}
|
||
|
||
|
||
def save_state(si, ui, total_new):
|
||
with open(STATE_FILE, 'w') as f:
|
||
json.dump({'sitemap_idx': si, 'url_idx': ui, 'total_new': total_new,
|
||
'ts': datetime.utcnow().isoformat()}, f)
|
||
|
||
|
||
def get_session():
|
||
sess = requests.Session()
|
||
sess.proxies.update(PROXIES)
|
||
sess.headers.update({
|
||
'User-Agent': UA,
|
||
'Accept': 'text/html,application/xhtml+xml,*/*;q=0.8',
|
||
'Accept-Language': 'es-AR,es;q=0.9',
|
||
'Accept-Encoding': 'gzip, deflate',
|
||
'Connection': 'keep-alive',
|
||
})
|
||
return sess
|
||
|
||
|
||
def get_conn():
|
||
return psycopg2.connect(host=DB_HOST, port=DB_PORT, dbname=DB_NAME,
|
||
user=DB_USER, password=DB_PASS, connect_timeout=15)
|
||
|
||
|
||
def create_db(conn):
|
||
cur = conn.cursor()
|
||
cur.execute("""
|
||
CREATE TABLE IF NOT EXISTS listings (
|
||
listing_id TEXT PRIMARY KEY,
|
||
titulo TEXT,
|
||
precio BIGINT,
|
||
moneda TEXT DEFAULT 'USD',
|
||
tipo TEXT,
|
||
ubicacion TEXT,
|
||
url TEXT UNIQUE,
|
||
superficie_m2 DECIMAL,
|
||
ambientes INT,
|
||
scraped_at TIMESTAMP DEFAULT NOW()
|
||
)
|
||
""")
|
||
conn.commit()
|
||
cur.close()
|
||
|
||
|
||
def upsert(conn, rows):
|
||
if not rows:
|
||
return 0
|
||
deduped = list({r['listing_id']: r for r in rows}.values())
|
||
cur = conn.cursor()
|
||
ids = [r['listing_id'] for r in deduped]
|
||
cur.execute('SELECT listing_id FROM listings WHERE listing_id = ANY(%s)', (ids,))
|
||
existing = {rec[0] for rec in cur.fetchall()}
|
||
nuevos = sum(1 for r in deduped if r['listing_id'] not in existing)
|
||
now = datetime.utcnow()
|
||
execute_values(cur, """
|
||
INSERT INTO listings (listing_id, titulo, precio, moneda, tipo, ubicacion,
|
||
url, superficie_m2, ambientes, scraped_at)
|
||
VALUES %s
|
||
ON CONFLICT (listing_id) DO UPDATE SET
|
||
titulo = EXCLUDED.titulo,
|
||
precio = EXCLUDED.precio,
|
||
moneda = EXCLUDED.moneda,
|
||
tipo = EXCLUDED.tipo,
|
||
ubicacion = EXCLUDED.ubicacion,
|
||
url = EXCLUDED.url,
|
||
superficie_m2 = EXCLUDED.superficie_m2,
|
||
ambientes = EXCLUDED.ambientes,
|
||
scraped_at = EXCLUDED.scraped_at
|
||
""", [(r['listing_id'], r.get('titulo'), r.get('precio'), r.get('moneda', 'USD'),
|
||
r.get('tipo'), r.get('ubicacion'), r.get('url'),
|
||
r.get('superficie_m2'), r.get('ambientes'), now) for r in deduped])
|
||
conn.commit()
|
||
cur.close()
|
||
return nuevos
|
||
|
||
|
||
def fetch_gz(sess, url, retries=3):
|
||
"""Descarga un .xml.gz y retorna lista de <loc> URLs."""
|
||
for attempt in range(retries):
|
||
try:
|
||
r = sess.get(url, timeout=30)
|
||
if r.status_code in (404, 410):
|
||
return []
|
||
r.raise_for_status()
|
||
try:
|
||
content = gzip.decompress(r.content).decode('utf-8', errors='replace')
|
||
except Exception:
|
||
content = r.text
|
||
return re.findall(r'<loc>(.*?)</loc>', content)
|
||
except Exception as e:
|
||
log.warning('sitemap %s intento %d: %s', url, attempt+1, e)
|
||
if attempt < retries - 1:
|
||
time.sleep([5, 15][attempt % 2])
|
||
return []
|
||
|
||
|
||
def fetch_page(sess, url, retries=4):
|
||
"""Descarga una página de listado y extrae listing items."""
|
||
backoffs = [2, 5, 15, 30]
|
||
for attempt in range(retries):
|
||
try:
|
||
r = sess.get(url, timeout=35)
|
||
# HTTP 202 = AWS WAF challenge, reintentar
|
||
if r.status_code == 202:
|
||
log.debug('HTTP 202 en %s (intento %d)', url, attempt+1)
|
||
if attempt < retries - 1:
|
||
time.sleep(backoffs[attempt])
|
||
continue
|
||
return []
|
||
if r.status_code in (403, 429):
|
||
log.warning('HTTP %d en %s', r.status_code, url)
|
||
time.sleep(45)
|
||
continue
|
||
if r.status_code in (404, 410):
|
||
return []
|
||
if r.status_code != 200:
|
||
log.warning('HTTP %d en %s', r.status_code, url)
|
||
return []
|
||
break
|
||
except requests.RequestException as e:
|
||
log.warning('Red %s intento %d: %s', url, attempt+1, e)
|
||
if attempt < retries - 1:
|
||
time.sleep(backoffs[attempt])
|
||
else:
|
||
return []
|
||
|
||
soup = BeautifulSoup(r.text, 'lxml')
|
||
items = soup.find_all('div', class_='listing__item')
|
||
if not items:
|
||
return []
|
||
|
||
results = []
|
||
seen = set()
|
||
|
||
for item in items:
|
||
try:
|
||
card = item.find('a', class_='card')
|
||
if not card:
|
||
continue
|
||
|
||
listing_id = card.get('idaviso', '').strip()
|
||
if not listing_id or listing_id in seen:
|
||
continue
|
||
seen.add(listing_id)
|
||
|
||
href = card.get('href', '')
|
||
full_url = BASE + href if href.startswith('/') else href
|
||
|
||
# Precio en USD desde atributo montonormalizado
|
||
try:
|
||
precio = int(card.get('montonormalizado', 0) or 0)
|
||
except (ValueError, TypeError):
|
||
precio = None
|
||
moneda = 'USD' # argenprop lista en USD
|
||
|
||
# Dormitorios
|
||
try:
|
||
ambientes = int(card.get('dormitorios', 0) or 0) or None
|
||
except (ValueError, TypeError):
|
||
ambientes = None
|
||
|
||
# Tipo de inmueble desde la URL
|
||
tipo = 'venta'
|
||
if 'arriendo' in href.lower() or 'alquiler' in href.lower():
|
||
tipo = 'arriendo'
|
||
|
||
# Tipo de propiedad
|
||
tipo_prop = None
|
||
for t in ('casa', 'departamento', 'terreno', 'campo', 'cochera', 'oficina', 'local', 'ph'):
|
||
if t in href.lower():
|
||
tipo_prop = t
|
||
break
|
||
|
||
# Ubicación desde URL: /casa-en-venta-en-palermo-... → palermo
|
||
loc_m = re.search(r'-en-(?:venta|alquiler|arriendo)-en-([a-z\-]+?)(?:-\d|-\-|$)', href)
|
||
ubicacion = loc_m.group(1).replace('-', ' ').title() if loc_m else None
|
||
|
||
# Título y m2 del texto del card
|
||
card_text = card.get_text(separator=' ', strip=True)
|
||
titulo_m = re.search(r'((?:Casa|Departamento|Terreno|Campo|Oficina|Local|PH|Cochera)[^|]{5,80})', card_text, re.I)
|
||
titulo = titulo_m.group(1).strip()[:255] if titulo_m else card_text[:100]
|
||
|
||
m2_m = re.search(r'(\d+(?:[,\.]\d+)?)\s*m[2²]', card_text, re.I)
|
||
superficie_m2 = None
|
||
if m2_m:
|
||
try:
|
||
superficie_m2 = float(m2_m.group(1).replace(',', '.'))
|
||
except ValueError:
|
||
pass
|
||
|
||
results.append({
|
||
'listing_id': listing_id,
|
||
'titulo': titulo or 'Sin titulo',
|
||
'precio': precio or None,
|
||
'moneda': moneda,
|
||
'tipo': tipo_prop or tipo,
|
||
'ubicacion': ubicacion,
|
||
'url': full_url,
|
||
'superficie_m2': superficie_m2,
|
||
'ambientes': ambientes,
|
||
})
|
||
except Exception as e:
|
||
log.debug('Error card: %s', e)
|
||
|
||
return results
|
||
|
||
|
||
def get_listing_sitemaps(sess):
|
||
"""Retorna lista de sub-sitemaps de tipo listing."""
|
||
r = sess.get(SITEMAP_INDEX, timeout=20)
|
||
all_sms = re.findall(r'<loc>(.*?)</loc>', r.text)
|
||
# Solo los sitemaps de listing (no ficha)
|
||
listing_sms = [u for u in all_sms if 'sitemap-listing' in u]
|
||
log.info('Sitemaps de listing: %d', len(listing_sms))
|
||
return listing_sms
|
||
|
||
|
||
def main():
|
||
state = load_state()
|
||
si_start = state['sitemap_idx']
|
||
ui_start = state['url_idx']
|
||
total_new = state.get('total_new', 0)
|
||
|
||
log.info('=== Scraper ArgenProp AR — modo %s ===',
|
||
'DAILY' if DAILY else 'BACKFILL')
|
||
log.info('Desde sitemap %d url %d', si_start, ui_start)
|
||
|
||
sess = get_session()
|
||
conn = get_conn()
|
||
create_db(conn)
|
||
|
||
# Descubrir sitemaps de listing
|
||
listing_sms = get_listing_sitemaps(sess)
|
||
listing_sms = listing_sms[:MAX_SITEMAPS]
|
||
|
||
for si in range(si_start, len(listing_sms)):
|
||
sm_url = listing_sms[si]
|
||
log.info('== Sitemap %d/%d: %s ==', si+1, len(listing_sms), sm_url)
|
||
|
||
# Descargar el índice del sub-sitemap (puede tener sub-sub-sitemaps)
|
||
sm_urls = fetch_gz(sess, sm_url)
|
||
if not sm_urls:
|
||
# Intentar sin gz
|
||
sm_urls = [sm_url.replace('.gz', '')]
|
||
log.info(' sub-sitemaps/urls: %d', len(sm_urls))
|
||
|
||
# Si son sub-sitemaps (terminan en .xml.gz), expandir
|
||
final_urls = []
|
||
for u in sm_urls:
|
||
if u.endswith('.xml.gz') or u.endswith('.xml'):
|
||
inner = fetch_gz(sess, u)
|
||
final_urls.extend(inner)
|
||
time.sleep(0.5)
|
||
else:
|
||
final_urls.append(u)
|
||
|
||
log.info(' URLs totales a procesar: %d', len(final_urls))
|
||
|
||
ui_ini = ui_start if si == si_start else 0
|
||
|
||
for ui in range(ui_ini, len(final_urls)):
|
||
page_url = final_urls[ui]
|
||
rows = fetch_page(sess, page_url)
|
||
|
||
if rows:
|
||
n = upsert(conn, rows)
|
||
total_new += n
|
||
if n > 0:
|
||
log.info('[%d/%d] %s: %d nuevos (acum %d)',
|
||
ui+1, len(final_urls), page_url[-60:], n, total_new)
|
||
else:
|
||
log.debug('[%d/%d] %s: sin resultados', ui+1, len(final_urls), page_url[-60:])
|
||
|
||
save_state(si, ui + 1, total_new)
|
||
time.sleep(DELAY)
|
||
|
||
# En modo daily parar temprano si no hay nuevos (ya cubierto)
|
||
if DAILY and ui > 200 and total_new == 0:
|
||
log.info('Daily: 200 URLs sin nuevos — deteniendo')
|
||
break
|
||
|
||
# Avanzar al siguiente sitemap
|
||
save_state(si + 1, 0, total_new)
|
||
time.sleep(DELAY * 2)
|
||
|
||
conn.close()
|
||
log.info('=== Finalizado: %d nuevos ===', total_new)
|
||
print(f'Finalizado: {total_new} nuevos')
|
||
|
||
|
||
if __name__ == '__main__':
|
||
main()
|