fix(linkedin): handle None job_level when LinkedIn omits "Seniority level"

[Recursing]

Bug

linkedin/__init__.py:240 raises AttributeError: 'NoneType' object has no attribute 'lower' when parse_job_level (util.py:42) returns None — which it correctly does (per its -> str | None signature) when the LinkedIn detail page lacks a "Seniority level" criterion. Common on non-US/UK postings (NL, DK, SE).

dict.get(k, default) only uses default when the key is missing, not when the value is None. So .get("job_level", "") returns None, and None.lower() raises.

The error aborts the entire scrape_jobs call, losing all results — including from other sites (Indeed, etc.) in the same invocation.

Reproduction

from bs4 import BeautifulSoup
from jobspy.linkedin.util import parse_job_level

# Detail page with Employment type + Industries but no "Seniority level"
soup = BeautifulSoup(
    '<h3 class="description__job-criteria-subheader">Employment type</h3>',
    "html.parser",
)
assert parse_job_level(soup) is None
{"job_level": parse_job_level(soup)}.get("job_level", "").lower()
# AttributeError: 'NoneType' object has no attribute 'lower'

Impact

In a daily LinkedIn pipeline with linkedin_fetch_description=True, ~15–30% of queries abort this way. From one run:

WARNING jobspy: query 'head of innovation'/'Rotterdam'  failed: 'NoneType' object has no attribute 'lower'
WARNING jobspy: query 'circular economy'/'Netherlands'  failed: 'NoneType' object has no attribute 'lower'
WARNING jobspy: query 'head of innovation'/'Copenhagen' failed: 'NoneType' object has no attribute 'lower'

Fix

(job_details.get("job_level") or "").lower() — handles both missing key and None value, preserves the existing , "" intent. Doesn't touch parse_job_level's honest str | None signature; the call site is the bug.

Confirmed on python-jobspy==1.1.82 (latest on PyPI) and main at fda080a.

---

🤖 Written by Claude Code.