feat: implement pagination support for Tyomarkkinatori job listings parser

This commit is contained in:
Esa Kataja
2025-08-04 13:27:36 +03:00
parent 2086c1b6e5
commit ac1ffcb7c2
+27 -2
View File
@@ -31,7 +31,6 @@ class _TyomarkkinatoriParams(BaseModel):
async def _parse_item(item: dict) -> JobListing:
"""Parse a single job listing from the search results page."""
logger.debug(f"Parsing job listing {item['id']}", extra={"item": item})
return JobListing(
url=f"https://tyomarkkinatori.fi/tyopaikat/{item['id']}",
company=item["employer"]["businessName"]["fi"]
@@ -78,11 +77,37 @@ async def parse_tyomarkkinatori(search_term: str) -> list[JobListing]:
data = response.json()
total_pages = data["totalPages"]
logger.debug(f"Found {total_pages} pages of search results")
total_items = data["totalElements"]
logger.debug(
f"Found {total_pages} pages of search results with {total_items} items."
)
# Parse 1st page
for item in data["content"]:
job_listing = await _parse_item(item)
all_job_listings.append(job_listing)
# Parse additional pages if they exist
for page in range(2, total_pages + 1):
try:
params.paging.pageNumber = page
response = await client.post(
base_url, json=params.model_dump(mode="json")
)
if response.status_code != 200:
logger.error(
f"Failed to fetch job listings from {base_url}",
extra={"response": response},
)
raise FetchException(base_url, response.status_code)
data = response.json()
for item in data["content"]:
job_listing = await _parse_item(item)
all_job_listings.append(job_listing)
except Exception as e:
logger.error(f"Error processing page {page}: {str(e)}")
# Continue with other pages even if one fails
continue
return all_job_listings