fix(review): lock OCR backend init, warn on rollback fallthrough, zero-page metric

Address PR #858 review round 2:

- OcrProcessor backend resolution is now guarded by an anyio.Lock (lazy-init,
  double-checked) so a burst of concurrent first-OCR calls resolves the backend
  once instead of each fetching its own gateway M2M token.
- The document_tier1_engine=pymupdf rollback now logs a warning when it falls
  back to the fast processor (no 'structured' registered) instead of silently
  using the very engine the operator opted out of.
- classify_from_text defaults ocr_frac to 0.0 (not 1.0) for a zero-page PDF, so
  the recorded classification metric is "fast" (no OCR evidence) rather than a
  misleading "ocr"; the no_text_layer/bad_text_layer flags are gated on having
  sampled at least one page.

New tests: zero-page classify routes fast, rollback-fallback warning.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

tests/unit/test_registry_tiering.py

@@ -92,6 +92,20 @@ async def test_engine_rollback_uses_structured(monkeypatch):
     assert res.processor == "structured"
 
 
+async def test_engine_rollback_warns_when_no_structured(monkeypatch, caplog):
+    # pymupdf rollback with no structured processor registered: it falls back to
+    # the fast processor but must warn (it silently used what the user opted out
+    # of otherwise).
+    monkeypatch.setattr(reg_mod, "get_settings", lambda: _Settings(engine="pymupdf"))
+    r = _registry((_Fake("fast", "fast"), 20))
+    with caplog.at_level(
+        "WARNING", logger="nextcloud_mcp_server.document_processors.registry"
+    ):
+        res = await r.process(b"%PDF-1.7", "application/pdf")
+    assert res.processor == "fast"
+    assert any("no 'structured' processor" in rec.message for rec in caplog.records)
+
+
 async def test_records_classification(monkeypatch):
     monkeypatch.setattr(reg_mod, "get_settings", lambda: _Settings())
     rec = MagicMock()