Pre-scrape: Unit Tests — SOF MEDICA S.A.

returns the primary strategy when it yields a value

falls through to the backup when the primary is empty

falls through when the primary THROWS (per-strategy try/catch)

uses the regex last-resort when every DOM strategy fails

returns {value:null} and WARNS once when all strategies fail

stays silent on total failure when opts.silent is set

wraps a single selector, passes an array through, drops falsy

cssText reads the first match under the scope

structuralText reads itemprop/meta content hooks

strips tags and collapses whitespace

returns null for empty input

extracts JobPosting objects from a bare object, an array and an @graph

ignores malformed JSON-LD blocks instead of throwing

finds the only anchor with no selector given

prefers an explicit selector when given

falls back to the first anchor when the selector misses

ignores fragment and javascript: links

returns null when the block has no anchor

mode css:<selector> when a primary selector matches

uses a fallback selector when the primary misses

mode jsonld when there is no article markup but JobPosting JSON-LD exists

mode regex:<article> when selectors miss and there is no JSON-LD

mode none when nothing matches at all

skips an invalid selector without crashing

should return response object with docs

should return empty docs when no jobs found

should throw on HTTP error

should retry a transient 503 and then succeed

should not retry a 4xx response

should accept array of jobs

should throw on HTTP error

should retry a transient 502 and then succeed

should delete a job by URL

should throw on HTTP error

should delete all jobs for a CIF

should throw on HTTP error

getCompanyByCif pads the CIF in the query string

querySOLR pads the CIF in the query string

upsertCompany pads the id field in the request body

deleteJobsByCIF pads the cif field in the request body

upsertJobs pads every job's cif field in the request body

already-padded CIFs pass through unchanged (idempotent)

should not have duplicate URLs for same CIF

should have valid CIF format for all jobs

should detect invalid CIF format

should have valid status values

should return array of companies for valid brand

should return empty array for non-existent brand

should include statusLabel in results

should fallback to CUIFirma when ANAF search fails

should encode brand name in URL

should return company data for valid CIF

should fallback to CUIScan when ANAF fails

should throw when both ANAF and CUIScan fail

should handle API-level error response

should return null when data is null

should return fresh data when API works

should use cached data when API fails

should throw when API fails and no cache available

should fetch the company via direct CIF lookup and return company data

should throw when ANAF returns no data

should throw when ANAF returns no company name

should use cached company data when available

should return company data with status active

should return inactive status when company is inactive

deletes jobs by CIF when an inactive company still has jobs in SOLR

regression: dry_run=true must never call deleteJobsByCIF, even with jobs present

strips Romanian diacritics and lowercases

collapses separators and trims dashes

converts a DD.MM.YYYY deadline to an end-of-day ISO string

returns undefined when no date is present

matches exact slugs

tolerates a trailing -2 disambiguation suffix

tolerates a small typo in the sitemap slug (edit distance <= 2)

returns null when nothing is close enough

extracts one item per article with a decoded, trimmed title

carries the deadline when present, undefined otherwise

carries the real scraped url, not a guessed slug

returns an empty array when the selector matches nothing

keeps two postings that share a title but have different URLs (dedup is title+URL, not title alone)

still drops a true duplicate match (same title AND same URL) from an overly broad fallback selector

skips an unconfigured {{PLACEHOLDER}} primary selector and uses a fallback

recovers via a fallback article selector when the class is renamed

recovers the title via a fallback heading selector when .job__title is gone

recovers the title via the anchor text when all headings are gone

falls back to JSON-LD JobPosting when there is no article markup at all

falls back to regex <article> slicing when the container selectors all miss

returns [] and does not throw when the page is unrecognisable (canary feeds off this)

should filter locations to only Romanian cities

should keep company uppercase

should normalize workmode values

should preserve expirationdate through the transform

should handle empty jobs array

should map a raw job to the job model format

should remove undefined fields

includes company name as h1

includes CIF

includes brand

includes status

includes website as markdown link

includes career page as markdown link

includes lastScraped date

omits optional fields when not present

shows job count in heading

shows 0 when no jobs

includes job title as h3

includes job URL as markdown link

includes workmode

includes location

includes tags

includes status

renders multiple jobs

handles job with no optional fields

returns a non-empty string

includes a generated timestamp

escapes # in job titles

escapes * in job titles

escapes [ ] in company name

escapes ` in tags

escapes # in location

returns the DOM, waits for the selector and closes the browser

closes the browser even when navigation fails

reports a missing playwright package as BrowserUnavailable

reports missing browser binaries as BrowserUnavailable

static mode never launches a browser

browser mode skips the static fetch

auto keeps the static HTML when it has jobs

auto renders when the static HTML has no jobs

auto falls back to the static HTML when no browser is available

auto does not swallow real rendering errors

rejects an unknown mode

ships defaulting to static

returns active when HEAD response is ok

returns expired on non-2xx response

returns error when fetch throws

issues a HEAD request

returns active when body has no expired keywords

returns expired when body has an expired keyword

returns expired on a hard 404 even without a keyword match

accepts custom keywords

returns error when fetch throws

is a non-empty array of lowercase strings

accepts a well-formed job

rejects an empty url

rejects a non-URL string

rejects a non-http(s) protocol

rejects an empty / whitespace title

rejects a title containing HTML

rejects a title over 200 chars

rejects a non-array location

rejects an empty entry

accepts a normal salary string

rejects a negative amount

rejects a non-string salary

reports multiple errors at once

rejects a non-object

keeps the good, drops the bad, and logs each drop

returns everything when all jobs are valid and logs nothing

throws on an empty array (aborts before any write)

throws on a zero count

returns the count when jobs are present

keeps only Romanian jobs, once each

maps city, remote workmode and trims the title

survives an empty or malformed payload

builds permalinks and filters by country

drops non-city locations and marks remote

parses items, decodes entities and dedupes by link

returns [] for an empty feed

returns the parsed JSON body

turns an HTML error page into a clear error

keeps active jobs and drops expired or erroring ones

returns an empty array when every job fails live validation