feat: air-gapped deployment guide, no implicit downloads

- Ship tiktoken's cl100k_base inside the package and build the encoding
  from it, so token counting never downloads anything.
- Default EMBEDDINGS_CACHE_DIR to <data home>/models instead of FastEmbed's
  temp dir, and read tokenizer.json and repo metadata from that cache, so
  a model downloads once and survives reboots.
- TTS_PROVIDER=none and STT_PROVIDER=none switch the speech features off:
  the endpoints return 404, audio files fail to ingest with a clear
  message, /api/config reports tts_available/stt_available, and the UI
  hides the Speak and microphone buttons.
- Drop the Google Fonts Roboto import from the web UI.
- prefetch-models fills the cache the app reads; verify-offline checks the
  packaged encoding.
- Docs: new Air-Gapped Deployment guide, settings and cache notes.

tests/scripts/test_verify_offline.py

@@ -8,8 +8,11 @@ from docsgpt.scripts import verify_offline
 
 
 def _fake_tiktoken(monkeypatch):
+    """The check must exercise the app's own loader, which reads the packaged encoding."""
+    encoding = types.SimpleNamespace(name="cl100k_base", encode=lambda text: [1, 2])
+    monkeypatch.setattr("docsgpt.utils.get_encoding", lambda: encoding)
     module = types.ModuleType("tiktoken")
-    module.get_encoding = lambda name: types.SimpleNamespace(encode=lambda text: [1, 2])
+    module.get_encoding = lambda name: (_ for _ in ()).throw(AssertionError("downloaded via tiktoken"))
     monkeypatch.setitem(sys.modules, "tiktoken", module)