checkpoint finder
This commit is contained in:
parent
8616837c01
commit
96dfdd7491
5 changed files with 130 additions and 20 deletions
|
|
@ -351,7 +351,7 @@ def parse_search_results(html: str) -> list[dict]:
|
|||
<div class="fw-medium text-primary fs-3">1 Bed Flat, Location, SW1Y</div>
|
||||
<ul>...<li>1 Bed</li><li>1 Bath</li><li>Furnished</li>...</ul>
|
||||
"""
|
||||
soup = BeautifulSoup(html, "html.parser")
|
||||
soup = BeautifulSoup(html, "lxml")
|
||||
properties = []
|
||||
|
||||
# Property cards: <a class="pli search-property-card">
|
||||
|
|
@ -486,7 +486,7 @@ def parse_property_detail(html: str) -> dict:
|
|||
- Tables have "Rent PCM", "Deposit", "Bills Included", etc. (NOT bedrooms)
|
||||
- Description in elements with class containing "description"
|
||||
"""
|
||||
soup = BeautifulSoup(html, "html.parser")
|
||||
soup = BeautifulSoup(html, "lxml")
|
||||
details: dict = {}
|
||||
|
||||
# --- Title from h1 ---
|
||||
|
|
@ -810,7 +810,7 @@ def search_outcode(
|
|||
if detail_html:
|
||||
detail_data = parse_property_detail(detail_html)
|
||||
# Shorter delay for detail pages (within same outcode)
|
||||
time.sleep(DELAY_BETWEEN_PAGES * 0.5)
|
||||
time.sleep(0.15)
|
||||
|
||||
transformed = transform_property(
|
||||
search_data,
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue