feat: Ollama und Modell vor dem Start prüfen, kurze Meldung statt Traceback

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

tests/conftest.py

@@ -1,4 +1,4 @@
-"""Shared fixtures: yt-dlp is always replaced by fakes (no network)."""
+"""Shared fixtures: yt-dlp and Ollama are always replaced by fakes (no network)."""
 import json
 import shutil
 import subprocess
@@ -7,6 +7,17 @@ from pathlib import Path
 import pytest
 
 
+@pytest.fixture(autouse=True)
+def ollama_available(request, monkeypatch):
+    """Skip the Ollama server check; tests marked real_ollama_check test it."""
+    if "real_ollama_check" in request.keywords:
+        return
+    import glossar_vorschlag
+    import rede_korrigieren
+    for module in (rede_korrigieren, glossar_vorschlag):
+        monkeypatch.setattr(module, "check_ollama", lambda model: None)
+
+
 @pytest.fixture
 def fake_yt_dlp():
     """Factory for a subprocess.run stand-in.