Files
hamkadr/src/JobsMedical.Web/Services/Scraping
soroush.asadi aaeb37e1af
CI/CD / CI · dotnet build (push) Successful in 46s
CI/CD / Deploy · hamkadr (push) Successful in 58s
Make dedupe stricter to avoid wrongly archiving distinct placeholder-facility ads
The job/talent duplicate-detector compared only the first 100–120 chars of the normalized
description. Since ~82% of jobs share the «نامشخص» placeholder facility, two genuinely different
ads that merely open the same way could collapse — wrongly archiving a valid listing (→ a 410 like
/Jobs/Details/7032). Compare a 400-char slice instead, so only near-identical full texts dedupe;
true reposts/fan-out (identical text) are still caught.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-26 03:26:25 +03:30
..