Match crawled listings to existing facilities (fuzzy) before creating new

When publishing a scraped listing we now look for a facility we already
have that is exactly or closely the same, and only create a new one when
there is no match — avoiding duplicates like «بیمارستان میلاد» vs «میلاد».

- ListingParser: extract a facility name (keyword + distinctive words) from
  the post and surface it in the parser notes.
- FacilityMatcher: Persian-aware normalization (ي/ك, ZWNJ, punctuation),
  type-word stripping for a "core" name, contains + Levenshtein similarity,
  and FindBest (same-city exact → any-city exact → same-city fuzzy → fuzzy).
- Review (manual publish): auto-select a matching facility or prefill the
  new-facility name; resolve-or-create uses fuzzy match; dropdown preselects.
- IngestionService (auto-publish): reuse FacilityMatcher against a run-wide
  facility list (grows as new ones are created) instead of exact-name only.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

src/JobsMedical.Web/Pages/Admin/Review.cshtml

@@ -51,10 +51,10 @@
                 <div class="filter-group">
                     <label>مرکز درمانی</label>
                     <select name="FacilityId">
-                        <option value="0">— انتخاب نشده —</option>
+                        <option value="0" selected="@(Model.FacilityId == 0)">— انتخاب نشده —</option>
                         @foreach (var f in Model.Facilities)
                         {
-                            <option value="@f.Id">@f.Name — @f.City?.Name</option>
+                            <option value="@f.Id" selected="@(Model.FacilityId == f.Id)">@f.Name — @f.City?.Name</option>
                         }
                     </select>
                     <input type="text" name="NewFacilityName" placeholder="یا نام مرکز جدید را وارد کن…" style="margin-top:6px;" />

src/JobsMedical.Web/Pages/Admin/Review.cshtml.cs

@@ -1,6 +1,7 @@
 using JobsMedical.Web.Data;
 using JobsMedical.Web.Models;
 using JobsMedical.Web.Services;
+using JobsMedical.Web.Services.Scraping;
 using Microsoft.AspNetCore.Authorization;
 using Microsoft.AspNetCore.Mvc;
 using Microsoft.AspNetCore.Mvc.RazorPages;
@@ -72,6 +73,25 @@ public class ReviewModel : PageModel
         if (Parsed.PayAmount is not null) { PayAmount = Parsed.PayAmount; SalaryMin = Parsed.PayAmount; }
         Description = Raw.RawText;
         Title = Parsed.RoleName is not null ? $"استخدام {Parsed.RoleName}" : "موقعیت استخدامی";
+
+        // Facility: try to match the listing's facility to one we already have; otherwise
+        // prefill the "new facility" box so publishing creates it.
+        if (!string.IsNullOrWhiteSpace(Parsed.FacilityName))
+        {
+            var cityId = await _db.Cities.Where(c => c.Name == Parsed.CityName)
+                .Select(c => (int?)c.Id).FirstOrDefaultAsync();
+            var match = FacilityMatcher.FindBest(Facilities, Parsed.FacilityName, cityId);
+            if (match is not null)
+            {
+                FacilityId = match.Id;
+                Parsed.Notes.Add($"مرکز منطبق در سیستم: «{match.Name}» — همین انتخاب شد.");
+            }
+            else
+            {
+                NewFacilityName = Parsed.FacilityName;
+                Parsed.Notes.Add($"مرکز جدید پیشنهادی: «{Parsed.FacilityName}» — هنگام انتشار ساخته می‌شود.");
+            }
+        }
         return Page();
     }
 
@@ -181,15 +201,17 @@ public class ReviewModel : PageModel
         if (string.IsNullOrWhiteSpace(NewFacilityName))
             return null;
 
-        // Reuse a same-named facility if one already exists, else create it.
         var name = NewFacilityName.Trim();
-        var existing = await _db.Facilities.FirstOrDefaultAsync(f => f.Name == name);
-        if (existing is not null) return existing.Id;
-
         var cityId = await _db.Cities.OrderByDescending(c => c.IsActive)
             .Select(c => (int?)c.Id).FirstOrDefaultAsync();
         if (cityId is null) return null;   // no cities seeded — cannot create a facility
 
+        // Reuse an existing facility that's exactly or closely the same (Persian-aware fuzzy
+        // match), so we don't create duplicates like «بیمارستان میلاد» vs «میلاد».
+        var all = await _db.Facilities.ToListAsync();
+        var match = FacilityMatcher.FindBest(all, name, cityId);
+        if (match is not null) return match.Id;
+
         var facility = new Facility
         {
             Name = name,