test(pseo): add 45 tests for health checks + pSEO Engine admin routes

Covers content/health.py (get_template_stats, get_template_freshness,
get_content_gaps, check_hreflang_orphans, check_missing_build_files,
check_broken_scenario_refs, get_all_health_issues) and all 6 routes in
admin/pseo_routes.py (dashboard, health partial, gaps partial, generate
gaps, jobs list, job status polling).

Also fixes two bugs found while writing tests:
- check_hreflang_orphans: was grouping by url_path, but EN/DE articles
  have different paths. Now extracts natural key from slug pattern
  "{template_slug}-{lang}-{nk}" and groups by nk.
- pseo_job_status.html + pseo_jobs.html: | default('') | truncate() fails
  when completed_at is None (default() only handles undefined, not None).
  Fixed to (value or '') | truncate().

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

web/src/padelnomics/admin/templates/admin/pseo_job_status.html

@@ -32,8 +32,8 @@
     </div>
     {% else %}—{% endif %}
   </td>
-  <td class="text-xs text-slate">{{ job.created_at | default('') | truncate(19, True, '') }}</td>
-  <td class="text-xs text-slate">{{ job.completed_at | default('') | truncate(19, True, '') }}</td>
+  <td class="text-xs text-slate">{{ (job.created_at or '') | truncate(19, True, '') }}</td>
+  <td class="text-xs text-slate">{{ (job.completed_at or '') | truncate(19, True, '') }}</td>
   <td>
     {% if job.error %}
       <details>

web/src/padelnomics/admin/templates/admin/pseo_jobs.html

@@ -75,8 +75,8 @@
             </div>
             {% else %}—{% endif %}
           </td>
-          <td class="text-xs text-slate">{{ job.created_at | default('') | truncate(19, True, '') }}</td>
-          <td class="text-xs text-slate">{{ job.completed_at | default('') | truncate(19, True, '') }}</td>
+          <td class="text-xs text-slate">{{ (job.created_at or '') | truncate(19, True, '') }}</td>
+          <td class="text-xs text-slate">{{ (job.completed_at or '') | truncate(19, True, '') }}</td>
           <td>
             {% if job.error %}
               <details>

web/src/padelnomics/content/health.py

@@ -235,10 +235,14 @@ async def check_hreflang_orphans(templates: list[dict]) -> list[dict]:
     For example: city-cost-de generates EN + DE. If the EN article exists but
     DE is absent, that article is an hreflang orphan.
 
+    Orphan detection is based on the slug pattern "{template_slug}-{lang}-{natural_key}".
+    Articles are grouped by natural key; if any expected language is missing, the group
+    is an orphan.
+
     Returns list of dicts:
         {
             "template_slug":      str,
-            "url_path":           str,
+            "url_path":           str,   # url_path of one present article for context
             "present_languages":  list[str],
             "missing_languages":  list[str],
         }
@@ -250,24 +254,39 @@ async def check_hreflang_orphans(templates: list[dict]) -> list[dict]:
             continue  # Single-language template — no orphans possible.
 
         rows = await fetch_all(
-            """SELECT url_path,
-                      GROUP_CONCAT(language) as langs,
-                      COUNT(DISTINCT language) as lang_count
-               FROM articles
-               WHERE template_slug = ? AND status = 'published'
-               GROUP BY url_path
-               HAVING COUNT(DISTINCT language) < ?""",
-            (t["slug"], len(expected)),
+            "SELECT slug, language, url_path FROM articles"
+            " WHERE template_slug = ? AND status = 'published'",
+            (t["slug"],),
         )
+
+        # Group by natural key extracted from slug pattern:
+        # "{template_slug}-{lang}-{natural_key}" → strip template prefix, then lang prefix.
+        slug_prefix = t["slug"] + "-"
+        by_nk: dict[str, dict] = {}  # nk → {"langs": set, "url_path": str}
         for r in rows:
-            present = set(r["langs"].split(","))
+            slug = r["slug"]
+            lang = r["language"]
+            if not slug.startswith(slug_prefix):
+                continue
+            rest = slug[len(slug_prefix):]  # "{lang}-{natural_key}"
+            lang_prefix = lang + "-"
+            if not rest.startswith(lang_prefix):
+                continue
+            nk = rest[len(lang_prefix):]
+            if nk not in by_nk:
+                by_nk[nk] = {"langs": set(), "url_path": r["url_path"]}
+            by_nk[nk]["langs"].add(lang)
+
+        for nk, info in by_nk.items():
+            present = info["langs"]
             missing = sorted(expected - present)
-            orphans.append({
-                "template_slug": t["slug"],
-                "url_path": r["url_path"],
-                "present_languages": sorted(present),
-                "missing_languages": missing,
-            })
+            if missing:
+                orphans.append({
+                    "template_slug": t["slug"],
+                    "url_path": info["url_path"],
+                    "present_languages": sorted(present),
+                    "missing_languages": missing,
+                })
 
     return orphans