fix: clean up CF_ORCH_URL/GPU_SERVER_URL env leak in test_config.py

Found during freeze-branch integration testing (PR #12 + #14 + #15
combined): app/config.py's write-back (os.environ["CF_ORCH_URL"] =
GPU_SERVER_URL) is a module-level side effect monkeypatch never tracks,
so it wasn't being reverted between tests. Once any test_config.py test
set a truthy GPU_SERVER_URL/CF_LICENSE_KEY, the resulting CF_ORCH_URL
wrote leaked into every later test in the same pytest session.

This silently broke all five "skips_embeddings_without_ollama_url"
tests across the shelve_*.py suite (pdf, docx, odt, ods, xlsx) — their
get_llm_config() check no longer saw a clean "nothing configured" state,
so they attempted embedding instead of skipping it. Neither PR #12 nor
#14 alone exercised this combination in the same test session, so it
only surfaced once merged together.

Fix: explicitly pop CF_ORCH_URL/GPU_SERVER_URL/CF_LICENSE_KEY from
os.environ in the autouse teardown fixture before reloading app.config,
rather than relying on monkeypatch to catch a write it never made itself.

tests/test_config.py

@@ -9,6 +9,7 @@ environment restored so later tests see normal module state.
 from __future__ import annotations
 
 import importlib
+import os
 
 import pytest
 
@@ -18,6 +19,13 @@ import app.config as config
 @pytest.fixture(autouse=True)
 def _restore_config_module():
     yield
+    # config.py's write-back (`os.environ["CF_ORCH_URL"] = GPU_SERVER_URL`) is a
+    # module-level side effect monkeypatch never sees, so it can't auto-revert
+    # it — without this explicit cleanup, a value set here leaks into every
+    # other test in the session (e.g. shelve tests' get_llm_config() checks).
+    os.environ.pop("CF_ORCH_URL", None)
+    os.environ.pop("GPU_SERVER_URL", None)
+    os.environ.pop("CF_LICENSE_KEY", None)
     importlib.reload(config)