From 2804db61738fafab5ba183b6bdc4a77abf25df16 Mon Sep 17 00:00:00 2001
From: Matin Nuhamunada <matinnu@biosustain.dtu.dk>
Date: Mon, 22 Apr 2024 23:22:47 +0200
Subject: [PATCH] fix: correct antismash version on example

---
 .examples/notebooks/custom_tutorial.ipynb | 88 ++++++++++++++++++++++-
 1 file changed, 87 insertions(+), 1 deletion(-)

diff --git a/.examples/notebooks/custom_tutorial.ipynb b/.examples/notebooks/custom_tutorial.ipynb
index f6ec332b..9c91b52b 100644
--- a/.examples/notebooks/custom_tutorial.ipynb
+++ b/.examples/notebooks/custom_tutorial.ipynb
@@ -219,7 +219,7 @@
    "outputs": [],
    "source": [
     "# Define the paths to the input files\n",
-    "antismash_regions_file = report_directory / \"tables/df_regions_antismash_7.0.0.csv\"\n",
+    "antismash_regions_file = report_directory / f\"tables/df_regions_antismash_{antismash_version}.csv\"\n",
     "display(Markdown(f\">`{antismash_regions_file}`\"))"
    ]
   },
@@ -1143,6 +1143,92 @@
     "display(HTML(filename=str(outfile)))"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "id": "88f5552d-f998-4b99-a86b-87b902710c7f",
+   "metadata": {},
+   "source": [
+    "### Adding adjacency edge\n",
+    "If you have a high quality (complete) genomes, it might be interesting to see the position of each BGC regions relative to each other in a genome. Unfortunately, there are limited physics-based layout algorithm in python, so it is better to use Cytoscape or Gephi to visualize the network"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "ce910590-50af-4505-b565-e6c555558963",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "antismash_table = report_dir / f\"tables/df_regions_antismash_{antismash_version}.csv\"\n",
+    "df_antismash = pd.read_csv(antismash_table)\n",
+    "\n",
+    "# Create a new graph\n",
+    "filtered_graph = nx.Graph()\n",
+    "\n",
+    "# Iterate over the nodes in the original graph\n",
+    "for node, data in G.nodes(data=True):\n",
+    "    # If the node meets the condition, add it to the new graph\n",
+    "    if data.get('node_trace') == 'BGC':\n",
+    "        filtered_graph.add_node(node, **data)\n",
+    "\n",
+    "# Now add only the edges that connect the nodes in the new graph\n",
+    "for u, v, data in G.edges(data=True):\n",
+    "    if filtered_graph.has_node(u) and filtered_graph.has_node(v):\n",
+    "        filtered_graph.add_edge(u, v, **data)\n",
+    "\n",
+    "for i in df_antismash.index:\n",
+    "    current_bgc = df_antismash.loc[i].to_dict()\n",
+    "    next_index = i + 1\n",
+    "    if next_index < len(df_antismash):\n",
+    "        neighbor_bgc = df_antismash.loc[next_index].to_dict()\n",
+    "        if current_bgc[\"accession\"] == neighbor_bgc[\"accession\"]:\n",
+    "            distance = neighbor_bgc[\"start_pos\"] - current_bgc[\"end_pos\"]\n",
+    "            assert distance > 0\n",
+    "            if not filtered_graph.has_edge(current_bgc[\"bgc_id\"], neighbor_bgc[\"bgc_id\"]):\n",
+    "                filtered_graph.add_edge(current_bgc[\"bgc_id\"], neighbor_bgc[\"bgc_id\"], distance_bp=distance, relation_type=\"genomic_adjacency\")\n",
+    "\n",
+    "display(Markdown(get_graph_stats(filtered_graph, \"integrated_graph_with_adjacency\")))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "4f0ee577-3e01-4ee7-8b9d-b73420933edb",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Get self-loops\n",
+    "self_loops = nx.selfloop_edges(filtered_graph)\n",
+    "\n",
+    "# Check if there are any self-loops\n",
+    "if self_loops is not None:\n",
+    "    # Remove self-loops\n",
+    "    filtered_graph.remove_edges_from(self_loops)\n",
+    "\n",
+    "# Iterate over the nodes of the graph, getting the node and its attributes\n",
+    "for n, data in filtered_graph.nodes(data=True):\n",
+    "    # Create a list of keys to remove after iterating over the dictionary\n",
+    "    keys_to_remove = []\n",
+    "    # Iterate over the items in the attributes dictionary\n",
+    "    for k, v in data.items():\n",
+    "        # Check if the value is not of a type compatible with GraphML\n",
+    "        if isinstance(v, (list, tuple)):\n",
+    "            data[k] = \", \".join([str(i) for i in v])\n",
+    "        elif v is None:\n",
+    "            # Add the key to the list of keys to remove\n",
+    "            keys_to_remove.append(k)\n",
+    "        elif not isinstance(v, (int, float, str, bool, np.int64)):\n",
+    "            print(f\"Node {n} has attribute {k} of incompatible type {type(v)}\")\n",
+    "    # Remove the keys with None values\n",
+    "    for key in keys_to_remove:\n",
+    "        del filtered_graph.nodes[n][key]\n",
+    "\n",
+    "outfile = Path(f\"assets/data/bigscape_{bigscape_cutoff}_as{antismash_version}_with_genomic_position.graphml\")\n",
+    "outfile.parent.mkdir(parents=True, exist_ok=True)\n",
+    "nx.write_graphml(filtered_graph, outfile)\n",
+    "display(Markdown(f\"[Download Graph]({str(outfile)})\"+'{:target=\"_blank\" .md-button}'))"
+   ]
+  },
   {
    "cell_type": "markdown",
    "id": "d7cc0c3e-8b47-47b7-bfbc-31398bcefd1d",