Boltz: Add smiles lookup from compound name from PubChem API

Boltz1.ipynb

@@ -50,6 +50,7 @@
         "import re\n",
         "import hashlib\n",
         "import random\n",
+        "import requests\n",
         "from string import ascii_uppercase\n",
         "\n",
         "# Function to add a hash to the jobname\n",
@@ -63,6 +64,8 @@
         "#@markdown  - Use `:` to specify multiple ligands as smile strings\n",
         "ligand_input_ccd = 'SAH'  #@param {type:\"string\"}\n",
         "#@markdown - Use `:` to specify multiple ligands as CCD codes (three-letter codes)\n",
+        "ligand_input_common_name = ''  #@param {type:\"string\"}\n",
+        "#@markdown - Use `:` to specify multiple ligands with their common name (e.g. Aspirin; SMILES fetched from [PubChem](https://pubchem.ncbi.nlm.nih.gov) API)\n",
         "dna_input = ''  #@param {type:\"string\"}\n",
         "#@markdown - Use `:` to specify multiple DNA sequences\n",
         "jobname = 'test'  #@param {type:\"string\"}\n",
@@ -71,6 +74,7 @@
         "query_sequence = \"\".join(query_sequence.split())\n",
         "ligand_input = \"\".join(ligand_input.split())\n",
         "ligand_input_ccd = \"\".join(ligand_input_ccd.split())\n",
+        "ligand_input_common_name = \"\".join(ligand_input_common_name.split())\n",
         "dna_input = \"\".join(dna_input.split())\n",
         "basejobname = \"\".join(jobname.split())\n",
         "basejobname = re.sub(r'\\W+', '', basejobname)\n",
@@ -95,8 +99,46 @@
         "protein_sequences = query_sequence.strip().split(':') if query_sequence.strip() else []\n",
         "ligand_sequences = ligand_input.strip().split(':') if ligand_input.strip() else []\n",
         "ligand_sequences_ccd = ligand_input_ccd.strip().split(':') if ligand_input_ccd.strip() else []\n",
+        "ligand_sequences_common_name = ligand_input_common_name.strip().split(':') if ligand_input_common_name.strip() else []\n",
         "dna_sequences = dna_input.strip().split(':') if dna_input.strip() else []\n",
         "\n",
+        "def get_smiles(compound_name):\n",
+        "    autocomplete_url = f\"https://pubchem.ncbi.nlm.nih.gov/rest/autocomplete/compound/{compound_name}/json?limit=1\"\n",
+        "    autocomplete_response = requests.get(autocomplete_url)\n",
+        "    if autocomplete_response.status_code != 200:\n",
+        "        return None\n",
+        "\n",
+        "    autocomplete_data = autocomplete_response.json()\n",
+        "    if autocomplete_data.get(\"status\", {}).get(\"code\") != 0 or autocomplete_data.get(\"total\", 0) == 0:\n",
+        "        return None\n",
+        "\n",
+        "    suggested_compound = autocomplete_data.get(\"dictionary_terms\", {}).get(\"compound\", [])\n",
+        "    if not suggested_compound:\n",
+        "        return None\n",
+        "    suggested_compound_name = suggested_compound[0]\n",
+        "\n",
+        "    smiles_url = f\"https://pubchem.ncbi.nlm.nih.gov/rest/pug/compound/name/{suggested_compound_name}/property/CanonicalSMILES/JSON\"\n",
+        "    smiles_response = requests.get(smiles_url)\n",
+        "    if smiles_response.status_code != 200:\n",
+        "        return None\n",
+        "\n",
+        "    smiles_data = smiles_response.json()\n",
+        "    properties = smiles_data.get(\"PropertyTable\", {}).get(\"Properties\", [])\n",
+        "    if len(properties) == 0:\n",
+        "        return None\n",
+        "\n",
+        "    return properties[0].get(\"CanonicalSMILES\")\n",
+        "\n",
+        "smiles_cache = {}\n",
+        "for name in ligand_sequences_common_name:\n",
+        "    if name not in smiles_cache:\n",
+        "        smiles_cache[name] = get_smiles(name)\n",
+        "        if smiles_cache[name] is not None:\n",
+        "          print(f\"Mapped compound {name} to {smiles_cache[name]}\")\n",
+        "\n",
+        "    if smiles_cache[name] is not None:\n",
+        "        ligand_sequences.append(smiles_cache[name])\n",
+        "\n",
         "# Initialize chain labels starting from 'A'\n",
         "chain_labels = iter(ascii_uppercase)\n",
         "\n",