fix(mosaico_integration): address notebook review feedback

- Setup cell: replace the stale git-clone + b3867be pin instructions
  with `pip install 'mosaicolabs>=0.4.0,<0.5' matplotlib pandas`,
  matching the Dockerfile + README bump in 4012fdf.
- Fleet comparison cell: pick the two sequences that are actually
  stationary instead of taking `lidar_sequences[0]` and `[1]`. The
  previous alphabetical selection silently paired robot-01 (noise)
  with robot-02 (noise under rotation), so the "noise vs drift"
  comparison promised in the markdown collapsed to noise vs noise.
  Now we intersect `lidar_sequences` with the stationary result
  from §6 so the plot compares robot-01 (noise) against robot-03
  (drift) as documented.

Author: Michał Fąferek

demos/mosaico_integration/notebooks/mosaico_demo.ipynb

@@ -43,13 +43,10 @@
     "- `localhost:18080` - medkit gateway (REST + SSE)\n",
     "- `localhost:16726` - mosaicod (Apache Arrow Flight)\n",
     "\n",
-    "The `LaserScanAdapter` we need merged into mosaico-labs/mosaico via PR #368 on 2026-04-13 (commit `b3867be`). The subsequent `mosaicolabs==0.3.2` wheel on PyPI is missing the `futures` subpackage, so until a fixed release ships install from the merged-PR commit directly:\n",
+    "Install the SDK plus the two plotting deps used below:\n",
     "\n",
     "```bash\n",
-    "pip install matplotlib pandas\n",
-    "git clone https://github.com/mosaico-labs/mosaico.git /tmp/mosaico\n",
-    "cd /tmp/mosaico && git checkout b3867be\n",
-    "pip install /tmp/mosaico/mosaico-sdk-py\n",
+    "pip install 'mosaicolabs>=0.4.0,<0.5' matplotlib pandas\n",
     "```\n",
     "\n",
     "If you skipped installing locally and only have the docker stack, you can still execute the cells inside the bridge container with `docker compose exec bridge python -c '...'`."
@@ -279,7 +276,7 @@
     "\n",
     "Mosaico's `QueryOntologyCatalog` supports compound expressions with implicit AND, so we can push all six conditions into one server-side query.\n",
     "\n",
-    "In the fleet variant this query is not cosmetic: the fleet injector rotates robot-02 (IMU `drift_rate=0.3 rad/s`) while triggering the same LIDAR noise as robot-01, so robot-02's `angular_velocity.z` sits around 0.3 rad/s. The `between(-0.1, 0.1)` predicate excludes that sequence, leaving only the two stationary LIDAR faults for downstream content comparison (§7). Step 2 doing real work, not just showing the SDK surface."
+    "In the fleet variant this query is not cosmetic: the fleet injector rotates robot-02 (IMU `drift_rate=0.3 rad/s`) while triggering the same LIDAR noise as robot-01, so robot-02's `angular_velocity.z` sits around 0.3 rad/s. The `between(-0.1, 0.1)` predicate excludes that sequence, leaving only the two stationary LIDAR faults for downstream content comparison (\u00a77). Step 2 doing real work, not just showing the SDK surface."
    ]
   },
   {
@@ -325,7 +322,13 @@
    "source": [
     "from pathlib import Path\n",
     "\n",
-    "if len(lidar_sequences) >= 2:\n",
+    "# Use the stationary sequences from \u00a76 - the IMU query already excluded robot-02\n",
+    "# (which was rotating during its fault). The two remaining stationary LIDAR\n",
+    "# sequences are robot-01 (noise injection) and robot-03 (drift injection).\n",
+    "stationary_names = {item.sequence.name for item in stationary}\n",
+    "stationary_lidar = [name for name in lidar_sequences if name in stationary_names]\n",
+    "\n",
+    "if len(stationary_lidar) >= 2:\n",
     "    def _pull_scan_stats(seq_name):\n",
     "        _sh = client.sequence_handler(seq_name)\n",
     "        _topic = _sh.get_topic_handler(\"/sensors/scan\")\n",
@@ -342,7 +345,9 @@
     "        df[\"t_s\"] = (df[\"timestamp_ns\"] - df[\"timestamp_ns\"].iloc[0]) / 1e9\n",
     "        return df\n",
     "\n",
-    "    seq_a, seq_b = lidar_sequences[0], lidar_sequences[1]\n",
+    "    # Sort so robot-01 is consistently on the left, robot-03 on the right.\n",
+    "    stationary_lidar = sorted(stationary_lidar)\n",
+    "    seq_a, seq_b = stationary_lidar[0], stationary_lidar[1]\n",
     "    df_a = _pull_scan_stats(seq_a)\n",
     "    df_b = _pull_scan_stats(seq_b)\n",
     "\n",
@@ -378,7 +383,11 @@
     "    plt.show()\n",
     "    print(f\"Left: {seq_a}\\nRight: {seq_b}\\nSaved to {_out_path}\")\n",
     "else:\n",
-    "    print(f\"Only {len(lidar_sequences)} LIDAR_SIM sequence(s). Run the fleet variant for comparison.\")"
+    "    print(\n",
+    "        f\"Only {len(stationary_lidar)} stationary LIDAR_SIM sequence(s) \"\n",
+    "        f\"(of {len(lidar_sequences)} total). Run the fleet variant and inject \"\n",
+    "        f\"faults on robot-01 and robot-03 for the comparison.\"\n",
+    "    )"
    ],
    "metadata": {},
    "execution_count": null,
@@ -407,4 +416,4 @@
  },
  "nbformat": 4,
  "nbformat_minor": 4
-}
+}