Enhance LLM and SwarmUI integration with improved configuration options

- Updated `env.example` and `gpu-rent.vars.example` to include new variables for LLM runtime and SwarmUI options.
- Refactored CLI commands to support interactive selection of LLM runtime and workload type (SwarmUI, LLM, or both).
- Improved access link generation to handle cases where SwarmUI is disabled, providing clearer user feedback.
- Enhanced provisioning logic to conditionally bootstrap SwarmUI based on user configuration, allowing for LLM-only setups.
- Updated documentation across multiple files to reflect changes in LLM integration, CLI usage, and configuration management.

src/gpu_rent/provision.py

@@ -440,33 +440,68 @@ def provision_vm(
     server_id: str | None = None,
     update: bool = True,
 ) -> None:
+    from gpu_rent.llm_runtime import normalize_runtime
+    from gpu_rent.state import load_state, save_state
+
+    swarm = bool(getattr(cfg, "enable_swarmui", True))
+    rt = normalize_runtime(cfg.llm_runtime)
+    if not swarm and rt == "none":
+        raise CloudError(
+            "llm-only: нужен LLM_RUNTIME=ollama|llamacpp (или --ollama / --llamacpp)"
+        )
+
     restart = bool(update)
-    try:
-        if seed_extensions(cfg, host, log, update=update):
-            restart = True
-    except GpuRentError as exc:
-        log(f"extensions: {exc}")
-        raise
-    try:
-        if seed_autocomplete(cfg, host, log):
-            restart = True
-    except GpuRentError as exc:
-        log(f"autocomplete: {exc}")
-    seed_civitai(cfg, host, log)
-    push_tree(cfg, host, cfg.local_models_dir, f"{DATA}/Models", log, models=True)
-    push_tree(cfg, host, cfg.local_wildcards_dir, f"{DATA}/Data/Wildcards", log, models=False)
-    push_tree(cfg, host, cfg.local_workflows_dir, f"{DATA}/CustomWorkflows", log, models=False)
-    if cfg.pull_output:
-        pull_tree(cfg, host, f"{DATA}/Output", cfg.local_output_dir, log)
+    if swarm:
+        try:
+            if seed_extensions(cfg, host, log, update=update):
+                restart = True
+        except GpuRentError as exc:
+            log(f"extensions: {exc}")
+            raise
+        try:
+            if seed_autocomplete(cfg, host, log):
+                restart = True
+        except GpuRentError as exc:
+            log(f"autocomplete: {exc}")
+        seed_civitai(cfg, host, log)
+        push_tree(cfg, host, cfg.local_models_dir, f"{DATA}/Models", log, models=True)
+        push_tree(
+            cfg, host, cfg.local_wildcards_dir, f"{DATA}/Data/Wildcards", log, models=False
+        )
+        push_tree(
+            cfg,
+            host,
+            cfg.local_workflows_dir,
+            f"{DATA}/CustomWorkflows",
+            log,
+            models=False,
+        )
+        if cfg.pull_output:
+            pull_tree(cfg, host, f"{DATA}/Output", cfg.local_output_dir, log)
+    else:
+        log("SwarmUI: skip (llm-only)")
+        run_ssh(
+            cfg,
+            host,
+            "sudo -n systemctl stop swarmui 2>/dev/null; "
+            "sudo -n systemctl disable swarmui 2>/dev/null || true; "
+            "echo llm-only | sudo -n tee /mnt/swarm_data/.gpu-rent-llm-only >/dev/null",
+            check=False,
+        )
 
     try:
         probe_gpu(cfg, host, log)
     except Exception as exc:
         log(f"GPU probe: {exc}")
 
-    ensure_swarmui_running(cfg, host, log, restart=restart)
-
-    from gpu_rent.state import load_state, save_state
+    if swarm:
+        ensure_swarmui_running(cfg, host, log, restart=restart)
+        run_ssh(
+            cfg,
+            host,
+            "sudo -n rm -f /mnt/swarm_data/.gpu-rent-llm-only",
+            check=False,
+        )
 
     try:
         provision_llm(cfg, host, log)
@@ -474,7 +509,6 @@ def provision_vm(
         st = load_state()
         st.notes = dict(st.notes or {})
         st.notes["llm_error"] = str(exc)[:500]
-        # Do not claim success — leave previous notes.llm_runtime or clear to none.
         st.notes["llm_runtime"] = "none"
         save_state(st)
         raise CloudError(f"LLM runtime: {exc}") from exc
@@ -498,10 +532,8 @@ def provision_vm(
                 "⚠ idle-killer НЕ вооружён — GPU может тарифицироваться без авто-stop. "
                 "См. status / docs/setup.md"
             )
-    log("SwarmUI слушает 127.0.0.1:7801 — gpu-rent tunnel")
-    from gpu_rent.llm_runtime import normalize_runtime
-
-    rt = normalize_runtime(cfg.llm_runtime)
+    if swarm:
+        log("SwarmUI слушает 127.0.0.1:7801 — gpu-rent tunnel")
     if rt == "ollama":
         log(f"Ollama API → localhost:{cfg.ollama_local_port} (туннель)")
     elif rt == "llamacpp":