tonio/job-research
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix: align Apec live selectors

Browse Source
This commit is contained in:
Antoine 2026-06-02 20:06:08 +02:00
parent 1c267520f0
commit c2ca1f457e
3 changed files with 25 additions and 33 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
src/job_research/apec/normalize.py
View File

@ -105,7 +105,7 @@ def normalize_apec_listing(
title = soup.select_one(".container-details-offer h1") or soup.find("h1")
details_offer_list = soup.select_one(".details-offer-list")
company = soup.select_one(".card-ents .ents-name")
company = soup.select_one(".card-ents-quote, .card-ents .ents-name")
if company is None and details_offer_list is not None:
company = details_offer_list.select_one("li:nth-of-type(1)")
@ -115,11 +115,7 @@ def normalize_apec_listing(
location = details_offer_list.select_one("li:nth-of-type(3)")
contract_type = _extract_contract_type(details_offer_list)
description_sections = [
_detail_block_text(soup, "Descriptif du poste"),
_detail_block_text(soup, "Profil recherché"),
]
description_text = "\n\n".join(section for section in description_sections if section)
description_text = _detail_block_text(soup, "Descriptif du poste")
return ApecListing(
source="apec",

42
tests/apec/test_normalize.py
View File

@ -3,27 +3,25 @@ from job_research.apec.normalize import normalize_apec_listing
def test_normalize_apec_listing_extracts_minimal_shape() -> None:
html = """
<html>
<body>
<main class="container-details-offer">
<h1>Data Engineer F/H</h1>
<div class="card-offer">
<div class="ref-offre">Ref. Apec : 178554452W</div>
<div class="details-offer-list">
<li>CLOUD TEMPLE</li>
<li>1 <span> CDI </span></li>
<li>Puteaux - 92</li>
</div>
<p>Publiée le 20/04/2026 Actualisée le 02/06/2026</p>
</div>
<div class="card-ents">
<span class="ents-name">CLOUD TEMPLE</span>
</div>
<div class="details-post">
<h4>Descriptif du poste</h4>
<p>Build pipelines</p>
</div>
<div class="details-post">
<html>
<body>
<main class="container-details-offer">
<h1>Data Engineer F/H</h1>
<div class="card-offer">
<div class="ref-offre">Ref. Apec : 178554452W</div>
<div class="details-offer-list">
<li>CLOUD TEMPLE</li>
<li>1 <span> CDI </span></li>
<li>Puteaux - 92</li>
</div>
<p>Publiée le 20/04/2026 Actualisée le 02/06/2026</p>
</div>
<div class="card-ents-quote">CLOUD TEMPLE</div>
<div class="details-post">
<h4>Descriptif du poste</h4>
<p>Build pipelines</p>
</div>
<div class="details-post">
<h4>Profil recherché</h4>
<p>Python / SQL</p>
</div>
@ -46,6 +44,6 @@ def test_normalize_apec_listing_extracts_minimal_shape() -> None:
assert listing.company == "CLOUD TEMPLE"
assert listing.location == "Puteaux - 92"
assert listing.contract_type == "CDI"
assert listing.description_text == "Build pipelines\n\nPython / SQL"
assert listing.description_text == "Build pipelines"
assert listing.published_at == "2026-04-20"
assert listing.fetched_at == "2026-06-01T10:00:00Z"

8
tests/test_apec_cli.py
View File

@ -41,9 +41,7 @@ def _apec_detail_html(
<p>Publiée le {published_at} Actualisée le {updated_at}</p>
</div>
</article>
<div class="card-ents">
<span class="ents-name">{company}</span>
</div>
<div class="card-ents-quote">{company}</div>
<div class="details-post">
<h4>Descriptif du poste</h4>
<p>{description}</p>
@ -131,7 +129,7 @@ def test_fetch_apec_reads_profile_and_writes_run_artifacts(tmp_path, monkeypatch
"company": "Example Corp",
"location": "Paris - 75",
"contract_type": "CDI",
"description_text": "Build pipelines\n\nPython / SQL",
"description_text": "Build pipelines",
"published_at": "2026-04-20",
"fetched_at": "2026-06-01T10:00:00Z",
"warnings": [],
@ -219,7 +217,7 @@ def test_fetch_apec_records_partial_failures_without_losing_successful_listings(
"company": "Example Corp",
"location": "Paris - 75",
"contract_type": "CDI",
"description_text": "Build pipelines\n\nPython / SQL",
"description_text": "Build pipelines",
"published_at": "2026-04-20",
"fetched_at": "2026-06-01T10:00:00Z",
"warnings": [],