02. Ligand Poses (Blind Docking).ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "0e64d2e3-ade5-4094-95c5-074b82e5e25d",
   "metadata": {},
   "source": [
    "# Ligand Poses (Blind Docking)\n",
    "The [Ligand Poses challenge](https://polarishub.io/competitions/asap-discovery/antiviral-ligand-poses-2025) of the [ASAP Discovery competition](https://polarishub.io/competitions/asap-discovery/antiviral-admet-2025) is a little more complex than the other challenges. \n",
    "\n",
    "The goal in ligand pose challenge is to predict the 3D binding pose of the ligand, given the ligand's SMILES (or CXSMILES, to be more precise) and the protein's sequence. This is also known as _blind docking_.\n",
    "\n",
    "Let's take a look!\n",
    "\n",
    "## Load the competition\n",
    "Make sure you're logged in! If you haven't logged in yet, simply run `polaris login` in a terminal of your choice. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "12392ca1-d954-4c97-ab84-c0e2bb265c19",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\u001b[32m2025-01-13 09:00:13.013\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mpolaris._artifact\u001b[0m:\u001b[36m_validate_version\u001b[0m:\u001b[36m90\u001b[0m - \u001b[1mThe version of Polaris that was used to create the artifact (0.11.0) is different from the currently installed version of Polaris (0.10.1.dev5+gffca9da).\u001b[0m\n",
      "\u001b[32m2025-01-13 09:00:13.017\u001b[0m | \u001b[33m\u001b[1mWARNING \u001b[0m | \u001b[36mpolaris.dataset._base\u001b[0m:\u001b[36mzarr_root\u001b[0m:\u001b[36m179\u001b[0m - \u001b[33m\u001b[1mYou're loading data from a remote location. If the dataset is small enough, consider caching the dataset first using CompetitionSpecification.cache() for more performant data access.\u001b[0m\n"
     ]
    }
   ],
   "source": [
    "import polaris as po\n",
    "\n",
    "competition = po.load_competition(\"asap-discovery/antiviral-ligand-poses-2025\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ecdd0f20-4fbd-455c-8d25-c9331d47bf4d",
   "metadata": {},
   "source": [
    "Let's again cache it as suggested!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "48fdb335-01a1-4210-81df-025187aeaf8d",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\u001b[32m2025-01-13 09:00:13.931\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mpolaris.dataset._base\u001b[0m:\u001b[36m_cache_zarr\u001b[0m:\u001b[36m374\u001b[0m - \u001b[1mCopying Zarr archive to /Users/cas.wognum/Library/Caches/polaris/datasets/c49e97bc-c2a7-4f11-abad-97edb05c8e3b/data.zarr. This may take a while.\u001b[0m\n",
      "\u001b[32m2025-01-13 09:00:44.482\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mpolaris.hub.storage\u001b[0m:\u001b[36mcopy_to_destination\u001b[0m:\u001b[36m177\u001b[0m - \u001b[1mCopied 200 (159.41 MiB), skipped 0, of 476 keys. 42.02% completed.\u001b[0m\n",
      "\u001b[32m2025-01-13 09:01:16.729\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mpolaris.hub.storage\u001b[0m:\u001b[36mcopy_to_destination\u001b[0m:\u001b[36m177\u001b[0m - \u001b[1mCopied 400 (240.17 MiB), skipped 0, of 476 keys. 84.03% completed.\u001b[0m\n",
      "\u001b[32m2025-01-13 09:01:26.985\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mpolaris.hub.storage\u001b[0m:\u001b[36mcopy_to_destination\u001b[0m:\u001b[36m177\u001b[0m - \u001b[1mCopied 476 (318.38 MiB), skipped 0, of 476 keys. 100.00% completed.\u001b[0m\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'/Users/cas.wognum/Library/Caches/polaris/datasets/c49e97bc-c2a7-4f11-abad-97edb05c8e3b'"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "competition.cache()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "8fc40a53-1f4d-4365-8c2a-23ba023533ca",
   "metadata": {},
   "outputs": [],
   "source": [
    "train, test = competition.get_train_test_split()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "103b1990-6db1-488d-b518-219ff2f3dca7",
   "metadata": {},
   "outputs": [],
   "source": [
    "X, y = train[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "7f48af30-de0a-48c2-aa5e-44c6be13f9ea",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'Chain B Sequence': 'SGFRKMAFPSGKVEGCMVQVTCGTTTLNGLWLDDVVYCPRHVICTSEDMLNPNYEDLLIRKSNHNFLVQAGNVQLRVIGHSMQNCVLKLKVDTANPKTPKYKFVRIQPGQTFSVLACYNGSPSGVYQCAMRPNFTIKGSFLNGSCGSVGFNIDYDCVSFCYMHHMELPTGVHAGTDLEGNFYGPFVDRQTAQAAGTDTTITVNVLAWLYAAVINGDRWFLNRFTTTLNDFNLVAMKYNYEPLTQDHVDILGPLSAQTGIAVLDMCASLKELLQNGMNGRTILGSALLEDEFTPFDVVRQCSGVT',\n",
       " 'CXSMILES': 'c1ccc(SCC[NH+]2CCOCC2)cc1',\n",
       " 'Chain A Sequence': 'SGFRKMAFPSGKVEGCMVQVTCGTTTLNGLWLDDVVYCPRHVICTSEDMLNPNYEDLLIRKSNHNFLVQAGNVQLRVIGHSMQNCVLKLKVDTANPKTPKYKFVRIQPGQTFSVLACYNGSPSGVYQCAMRPNFTIKGSFLNGSCGSVGFNIDYDCVSFCYMHHMELPTGVHAGTDLEGNFYGPFVDRQTAQAAGTDTTITVNVLAWLYAAVINGDRWFLNRFTTTLNDFNLVAMKYNYEPLTQDHVDILGPLSAQTGIAVLDMCASLKELLQNGMNGRTILGSALLEDEFTPFDVVRQCSGVT',\n",
       " 'Protein Label': 'SARS-CoV-2 Mpro'}"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "8264f466-a77c-42a1-a1ed-78c24901b663",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<rdkit.Chem.rdchem.Mol at 0x163d5e160>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5917f420-7e31-413c-86bb-de6ba4c66a96",
   "metadata": {},
   "source": [
    "## Extra data\n",
    "\n",
    "The dataset of the competition provides additional data you could use for the development of your model. You're free to use other publicly available data to develop your method."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "29c89d01-00a7-40bb-ac05-5902255b2f12",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['CXSMILES', 'Chain A Sequence', 'Chain B Sequence', 'Complex Structure', 'Ligand Pose', 'Protein Label', 'Protein Structure'])"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Get a single data point\n",
    "datapoint = competition[0]\n",
    "\n",
    "# See the data that's available\n",
    "datapoint.keys()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "89c71ffc-bc64-47b0-b72a-e8957e1ba639",
   "metadata": {},
   "source": [
    "### The data structures\n",
    "\n",
    "Polaris hosts ML-ready datasets. You don't have to worry about loading domain-specific file formats, but can just start hacking right away.\n",
    "\n",
    "For example, to get the 3D coordinates of the complex, you can simply do:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "bf451da6-ff6c-4c87-8d79-9eaff07bc0c9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ -2.395,   4.215, -17.097],\n",
       "       [ -1.977,   5.504, -16.496],\n",
       "       [ -2.341,   5.543, -15.01 ],\n",
       "       ...,\n",
       "       [ 21.061,  -8.835,   7.04 ],\n",
       "       [ 22.269, -10.73 ,   6.01 ],\n",
       "       [ 28.088, -10.911,   5.35 ]], dtype=float32)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "datapoint[\"Complex Structure\"].coord"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4e98635e-100d-43e5-b4e5-48b8a0c22b83",
   "metadata": {},
   "source": [
    "Want to find the corresponding atom labels? You can do:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "da2cae15-6d64-4924-82ac-798a992f3d9d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['N', 'CA', 'C', ..., 'H', 'H', 'H'], dtype='<U6')"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "datapoint[\"Complex Structure\"].atom_name"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "69bee0cb-cd9a-42cc-b389-f6903efc3390",
   "metadata": {},
   "source": [
    "What about the chain ids?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "27ec63ba-f1ed-4529-b587-dba4308e70e7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['A', 'A', 'A', ..., '', '', ''], dtype='<U4')"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "datapoint[\"Complex Structure\"].chain_id"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d6ecb956-7454-4725-8d1b-90d77a31f7a2",
   "metadata": {},
   "source": [
    "Polaris provides easy to use data structures for ML. In the case of this challenge, we specifically have two data types that are worth mentioning:\n",
    "- The [Biotite](https://www.biotite-python.org/) `AtomArray`\n",
    "- The [RDKit](https://www.rdkit.org/) `Chem.Mol`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "2da56def-f8e0-4065-a27a-fbfb006a9bba",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(biotite.structure.AtomArray, biotite.structure.AtomArray)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "type(datapoint[\"Complex Structure\"]), type(datapoint[\"Protein Structure\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "c6bdd0bd-3c19-4acb-9024-d1931b56a8f7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "rdkit.Chem.rdchem.Mol"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "type(datapoint[\"Ligand Pose\"])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "122cfd56-8938-4dec-b3e5-c53a0e581a84",
   "metadata": {},
   "source": [
    "While this is convenient for many ML applications, you may prefer to use more traditional formats (e.g. as input to traditional docking tools). No problem!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "c79f2b08-9189-4c33-945b-ba679e2ad68c",
   "metadata": {},
   "outputs": [],
   "source": [
    "import fastpdb\n",
    "import datamol as dm\n",
    "\n",
    "from tempfile import TemporaryDirectory\n",
    "from pathlib import Path\n",
    "\n",
    "\n",
    "with TemporaryDirectory() as tmpdir: \n",
    "\n",
    "    tmpdir = Path(tmpdir)\n",
    "    \n",
    "    # The RDKit Mol object supports various IO formats. \n",
    "    # Here we use the Datamol library (which wraps RDKit) to effortlessly write to SDF\n",
    "    dm.to_sdf(datapoint[\"Ligand Pose\"], tmpdir / \"mol.sdf\")\n",
    "\n",
    "    # Similarly, the Biotite AtomArray also support various IO formats\n",
    "    # Here we use fastpdb, a fast drop-in replacement for Biotite IO.\n",
    "    out_file = fastpdb.PDBFile()\n",
    "    out_file.set_structure(datapoint[\"Complex Structure\"])\n",
    "    out_file.write(tmpdir / \"complex.pdb\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b947c513-aced-4b1c-85ec-89135c9fb80a",
   "metadata": {},
   "source": [
    "### Raw data dump\n",
    "We've decided to sacrifice the completeness of the scientific data to improve its ease of use. For those that are interested, you can also access the raw data dump that this dataset has been created from."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "4f61acee-c81a-4b17-abe6-4f25cac29a1d",
   "metadata": {},
   "outputs": [],
   "source": [
    "import fsspec\n",
    "import zipfile\n",
    "\n",
    "with fsspec.open(\"https://fs.polarishub.io/2025-01-asap-discovery/raw_data_package.zip\") as fd:\n",
    "    with zipfile.ZipFile(fd, 'r') as zip_ref:\n",
    "        zip_ref.extractall(\"./raw_data_package/\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "2a87870f-9f36-456d-a3f3-cb31d41746cb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "PosixPath('raw_data_package/ligand-posing/SARS-CoV-2_Mpro-x0398_0A_CONFIDENTIAL')"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "path = Path(\"./raw_data_package\") / \"ligand-posing\"\n",
    "subdir = list(path.iterdir())[0]\n",
    "subdir"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "94ccce1f-3c59-4135-a276-16263ec8df65",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[PosixPath('raw_data_package/ligand-posing/SARS-CoV-2_Mpro-x0398_0A_CONFIDENTIAL/protein.fasta'),\n",
       " PosixPath('raw_data_package/ligand-posing/SARS-CoV-2_Mpro-x0398_0A_CONFIDENTIAL/ligand.smi'),\n",
       " PosixPath('raw_data_package/ligand-posing/SARS-CoV-2_Mpro-x0398_0A_CONFIDENTIAL/ligand.sdf'),\n",
       " PosixPath('raw_data_package/ligand-posing/SARS-CoV-2_Mpro-x0398_0A_CONFIDENTIAL/complex.pdb'),\n",
       " PosixPath('raw_data_package/ligand-posing/SARS-CoV-2_Mpro-x0398_0A_CONFIDENTIAL/protein.pdb')]"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "list(subdir.iterdir())"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7f4d72cd-faef-4285-a7e1-fef4bf1284d8",
   "metadata": {},
   "source": [
    "# Build a model\n",
    "Building a ligand pose prediction model is non trivial (e.g. see [Boltz-1](https://github.com/jwohlwend/boltz)) and outside the scope of this notebook.\n",
    "\n",
    "# Prepare your inputs\n",
    "Since building a ligand pose prediction model is out of scope, we'll simply generate a conformer using RDKit for the sake of simplicitly. This just serves as a way to show how to prepare your inputs for submission."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "545094e0-1d59-46ba-9af6-3d2051b4cf6d",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████████| 195/195 [00:11<00:00, 17.25it/s]\n"
     ]
    }
   ],
   "source": [
    "import datamol as dm\n",
    "import numpy as np\n",
    "\n",
    "from tqdm import tqdm\n",
    "\n",
    "y_pred = []\n",
    "\n",
    "for x in tqdm(test): \n",
    "\n",
    "    # Generate a random conformer\n",
    "    mol = dm.to_mol(x[\"CXSMILES\"])\n",
    "    mol = dm.conformers.generate(mol, n_confs=1)\n",
    "\n",
    "    # That's our prediction\n",
    "    y_pred.append(mol)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "eb804646-f8a7-4e3d-a90b-08d6dcacba61",
   "metadata": {},
   "source": [
    "## Alignment\n",
    "\n",
    "We will evaluate your predictions using RMSD between the predicted and actual pose. This raises a problem: You could have the perfect pose, but rotated, scaled, or translated. To allow for a fair comparison, everyone is therefore required to align their pose predictions ahead of submission with a reference structure that we provide.\n",
    "\n",
    "Let's start by downloading the reference structures!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "9c29d582-8168-44aa-83ed-ff8139cd4270",
   "metadata": {},
   "outputs": [],
   "source": [
    "import zipfile\n",
    "\n",
    "with fsspec.open(\"https://fs.polarishub.io/2025-01-asap-discovery/ligand_poses_reference_structures.zip\") as fd:\n",
    "    with zipfile.ZipFile(fd, 'r') as zip_ref:\n",
    "        zip_ref.extractall(\"./reference_structures/\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "d882feb5-10f6-478b-b4e9-537f27dd24ac",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[PosixPath('reference_structures/MERS-CoV-Mpro'),\n",
       " PosixPath('reference_structures/__MACOSX'),\n",
       " PosixPath('reference_structures/SARS-CoV-2-Mpro')]"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "path = Path(\"./reference_structures\")\n",
    "list(path.iterdir())"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c6fdb8d5-51a8-4d4d-b86c-4262058473a6",
   "metadata": {},
   "source": [
    "We provide more information than just the protein structure"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "92f1755d-e969-4b29-bd01-068b6c39ae0e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[PosixPath('reference_structures/SARS-CoV-2-Mpro/.DS_Store'),\n",
       " PosixPath('reference_structures/SARS-CoV-2-Mpro/protein.fasta'),\n",
       " PosixPath('reference_structures/SARS-CoV-2-Mpro/ligand.smi'),\n",
       " PosixPath('reference_structures/SARS-CoV-2-Mpro/ligand.sdf'),\n",
       " PosixPath('reference_structures/SARS-CoV-2-Mpro/complex.pdb'),\n",
       " PosixPath('reference_structures/SARS-CoV-2-Mpro/protein.pdb')]"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "list((path / \"SARS-CoV-2-Mpro\").iterdir())"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3baf93a1-88cb-4380-a171-5b6bdaf40f0d",
   "metadata": {},
   "source": [
    "#### Dummy example\n",
    "To show how to do the alignment, let's work through a dummy example by aligning the reference structure with a corrupted version of itself.\n",
    "\n",
    "We will use PyMol. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "f8af3bcc-4362-42fd-95a5-0e79ad52ab5d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "30.15460205078125"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pymol\n",
    "\n",
    "# Let's reset! Just to be sure.\n",
    "pymol.cmd.delete(\"all\")\n",
    "\n",
    "# Load the reference and mobile structure\n",
    "ref_path = path / \"SARS-CoV-2-Mpro\" / \"complex.pdb\"\n",
    "pymol.cmd.load(ref_path, \"mobile\")\n",
    "pymol.cmd.load(ref_path, \"reference\")\n",
    "\n",
    "# Corrupt the mobile protein\n",
    "pymol.cmd.rotate(\"x\", 90, selection=\"mobile\")\n",
    "\n",
    "# Compute the RMSD\n",
    "pymol.cmd.rms_cur(\"mobile\", \"reference\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "16512e54-60b4-4294-9325-7d80605d8859",
   "metadata": {},
   "source": [
    "As expected, the alignment between the proteins is horrible!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "96721783-152a-4b9a-9b74-f62461d27951",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " Match: read scoring matrix.\n",
      " Match: assigning 603 x 603 pairwise scores.\n",
      " MatchAlign: aligning residues (603 vs 603)...\n",
      " MatchAlign: score 3245.000\n",
      " ExecutiveAlign: 603 atoms aligned.\n",
      " Executive: RMSD =    0.000 (603 to 603 atoms)\n"
     ]
    }
   ],
   "source": [
    "# Now let's align our proteins again using PyMol\n",
    "pymol.cmd.align(\n",
    "    \"polymer and name CA and mobile\",\n",
    "    \"polymer and name CA and reference\",\n",
    "    quiet=0,\n",
    ");"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4b177983-72c5-4755-b5b5-86e348788813",
   "metadata": {},
   "source": [
    "And the RMSD is 0 again!\n",
    "\n",
    "**Please make sure to align your predictions with the provided reference structure!**"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "40bcdda6-d780-442d-8f70-e859f5347af9",
   "metadata": {},
   "source": [
    "## Serialize\n",
    "To submit to Polaris, you need to serialize the molecules to a bytestring. Please stick to the exact below example."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "e3d8f1d7-31cd-4949-9a68-4d996e33a6c0",
   "metadata": {},
   "outputs": [],
   "source": [
    "import base64\n",
    "from rdkit import Chem\n",
    "\n",
    "\n",
    "def serialize_rdkit_mol(mol: Chem.Mol): \n",
    "    props = Chem.PropertyPickleOptions.AllProps\n",
    "    mol_bytes = mol.ToBinary(props)\n",
    "    return base64.b64encode(mol_bytes).decode('ascii')  \n",
    "\n",
    "\n",
    "y_pred_serialized = [serialize_rdkit_mol(mol) for mol in y_pred]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c69f421e-911f-4229-a833-58f738101f3b",
   "metadata": {},
   "source": [
    "#### Sanity check\n",
    "Let's deserialize the molecule again to make sure we've done this correctly"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "0b5e064c-7deb-406a-b0de-a12cef4d330a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAcIAAACWCAIAAADCEh9HAAAABmJLR0QA/wD/AP+gvaeTAAAgAElEQVR4nO3deVyN2R8H8M9d6t72jVZllCLM2GZsGYXsWaOMscSQnTGGBjOawe+nMVuWn5G1jG1KDBJNSWSylUKFVLK0SZv2ut3z++MxjTE01b11c/u+/+h13fuc83yvMZ/Os53DY4yBEEJIQ/EVXQAhhLzdKEYJIUQmFKOEECITilFCCJEJxSghhMiEYpQ0dyEhId99911CQoKiCyHk9Xh0wxNpngoKCvz9/ffv3x8VFcUYE4vFISEhAwYMUHRdhLyKYpQ0L1VVVWfPnvXz8wsKCqqoqACgo6MjlUqLiorEYvGePXumTJmi6BoJ+Rs6qCfNRUJCwhdffGFubj5mzJjAwMCqqipHR0c/P7/09PT8/PwlS5aUl5dPnTr1iy++kEqlii6WkL/QaJQoWF5e3tGjR318fG7cuMG9Y2tr6+LiMnPmzLZt2wKoqqpydnaeOXNmTk7OokWLqqqqJk6c6Ofnp66uLst+JRJJeHh4r169dHV15fA1SEvGCFGE8vLykydPTpo0SUVFhfunqKen5+7uHhkZ+cqWvr6+APh8vpeXV0hIiI6ODoDevXtnZmY2bNfx8fGenp5t2rQBYGFhcfnyZZm/DWnRKEZJU/Pz83N3d9fX1+fSU0VFZezYsceOHauoqHhTE29vbz6fD2DmzJmxsbHt2rUDYGZmFhMTU/f9pqWl/fe//+3SpUvNGIJLcC0trfPnz8vhi5GWimKUNKlNmzbVpFinTp28vLyys7Pr0vD06dNaWloA7Ozs7t27x12y19TUPHHiRO0N8/Pz/fz8HB0deTxezbB32rRpoaGhVVVVbm5uAEQi0bFjx+Tx/UhLRDFKmlTXrl0B2Nra3r59u75tb968aWFhAcDKyiouLm769OkABAKBl5fXPzcuKyvjThqoqqpy6SkWi52cnPz9/SsrK7lt9uzZc/Hixfnz53Mj00OHDsv69UiLRDFKmk5RUZFIJBIIBI8fP67jIPQV6enp77//PjeiDA0N9fLy4g72Z8+ezYVjdXV1ZGSku7u7trY2l558Pt/Ozs7Hx+f58+cvd3XhwgU+n6+hoRESEuLl5aWiotKr1ylvb/l8U9KiUIySphMYGAigf//+R44c4fP5ixYtakAnZWVlkydPBiAUCrdt2xYQEMBdsu/Xr9/ChQtNTU1rThr07t178+bNWVlZr+1HIpHMnj0bgKqqakBAwP/+l8jjMR6PvW5oS0htKEZJ05k1axaAjRs3fvzxxwB+/PHHhvUjlUo9PT25rFyyZElUVFTr1q01NTW5d9q2bevh4XHv3r269LNs2TLuzMCRI3/s2MH4fAYwDw8mlTasNNISUYySJiKVSrmh4s2bNw0MDAAkJSXJ0uGhQ4fEYvGwYcOqqqo+++wzAF27dr1y5Up9+/Hy8rKz+1ggqP7hB3boEFNRYQCbP5+SlNQVPcVEmkh0dHRGRoaFhcXz589zc3M7dOhgbW0tS4cfffTRuXPn/P39hULhH3/8AWDDhg29e/eubz8eHh4uLr9Ipfzly5GUhOPHoaYGXV0sWIDCQgC4dAmXL8tSKVFyQkUXQFqK06dPAxgzZgz3wsnJSfY++/XrB+DZs2fXr18XiUQODg4N62fJEp6+PmbOxNdfY+FCxMSgY0cYG0MoxNatuHsXQiH69pW9XqKcaDRKmkhQUBCAUaNG1byQV89nz56VSqUODg41p0cbYOpUBAZCLEZBATp0AI+HTp1QUIDr1+VVJlFaFKOkKWRmZt64cUNDQ8PKyio+Pl5HR6d///7y6vzMmTMARowYIWM/Y8YgMhL79oH/5/8W336LlStBE6GQ2lGMkqYQFBTEGHN0dAwJCQEwbNiwmkfpZSSVSsPCwiCPGAXw/vt4uS5TU4wdC19f2TsmyoxilDQF7nzoqFGjal7Iq+dr1649ffrU0tLSxsZGXn2+bPFilJU1RsdEeVCMkkZXUVERHh7O4/EcHBwiIiL4fP7w4cPl1Tl3RC/HXK7h6oqsLAgEOHMGo0fLvXuiPChGSaM7f/58UVFRjx49EhMTy8vL+/TpY2hoKK/Og4ODIacj+pfl5WHRIrRrh9JSGBvDwEC+3ROlQjFKGl3NHU5yP6LPycm5ceOGmpqavb29vPrkhISguhr9+0O2uaFJi0AxShodl54jR47kRo5yuWOUExwcLJVKBw4cKONM+P905gwAyHuMS5QTxShpXFeuXHnw4IGhoaGpqalIJDI3N3/33Xfl1bm8bnV6hVSK338HgJEj5dsxUU70FBNpXPv37+fxeEKh0MzMLCUl5enTpzXTJ8uourqau9VJjhesONHRyM5Gu3bo2FG+HRPlRKNR0rimTJnC5/MzMjJcXFxKS0tlv7hUXl5eVlYG4MqVK7m5uTY2Nu3bt5dHpX/hjuhpKErqiGKUNK7+/fufOHFCV1f36NGjdnZ2jx8/bnBXMTExS5cuNTc39/Pzw59H9CMbIe3i4lY5OIQ6OT2Xe89EOSl6iinSIiQlJXXo0AGAqalpdHR0vdomJyd7enpaWlrW/KOdM2cOY6x79+4AQkJC5FtqTk4On88Xi8UlJSXy7ZkoKzo3SpqCtbV1VFSUs7NzRESEvb39gQMHxo0bV3uTwsLCEydO/PLLL+fOnWOMATAzM3N2dnZzc8vNzf3444/T0tJUVFReXulTLmomOpH71X+itBSd46QFqaio4Fbi5PF4np6er91GIpGEhoZOmzatJsXU1NQmTZp08uTJq1evLlmypHXr1tz7QqEQQKdOnR4/fix7bdI/Z2meMmUKgM2bN8veJ2khKEZJU6tZdH7WrFkvr00fHx/v4eFhZGTEpWTNUnT37t3z9vbu1q1bze/+Tp06eXp6Xrt2jXuzbdu2d+/ebXA9CQkJHh4ePXv2rK6ulkgkcpmZn7QoFKNEAQIDAzU0NADY2dk9ffqUMVZVVVUToFxK3rlzx9/f38nJiRt1AtDX13d3d4+MjKzpJz8/387OjvuovsuHPH/+fOfOnR988EFNOkdFRUVFRQGwtLSU8xcmSo1ilChGXFxczaLziYmJjLF169YtW7YsJiaGWyG5Zg5mkUj0yvryLyspKeFuv9fU1AwNDa3LrqOjo93d3bW0tLj+dXR03N3dY2Jijh071qpVKz6fP2/ePDl/W6LUKEaJwryy6HxCQoKnp+c777xTMzzs2bOnt7f3s2fPau+noqKCW3JZJBIFBga+abPCwkIfH58ePXq83L+Pj098fPzq1atNTEy4N1VUVMLDw+X9XYkyoxglilRcXDxhwgTuolNNullbW69bt+7Bgwd170cqlX766acABALBnj17XvmUG37WDG91dXXd3d3j4uJCQ0MnTZokEAi49zt06LB+/Xo6K0rqi2KUKFh1dTV385OKisq0adNCQ0OlDV3a2MvLi0vk77//nnvn6dOnNXdE8Xi8wYMHHzlyhLumVHMqViQSTZo0SZb9khaOxxir7X4oQppEYWEhj8fT1taWsZ9t27YtXbpUKpV6eHhwqdqlS5dnz565ubnNmDEjPj5+586dNTei2trazpgxY/bs2QY0nyiRAcUoUTYHDhyYNWtWVVXV/Pnzt23b9uDBg4qKiv379+/duzcnJweAWCwePXq0u7u7o6OjooslyoBilCihoKAgFxeXsrIybv3RS5cuce/36NFjzpw5U6ZMkX3YS0gNilGinC5evDhixAgtLa3s7GwtLa1x48ZNnz6dhp+kMVCMEqXVvn37lJSUpUuXbtiwoeYyPSFyRzFKlFNcXFz37t1bt26dnp6u8vLa84TIG803SpTTwYMHAbi6ulKGksZGo1GihKRSadu2bZ88eXL58uU+ffoouhyi5Gg0SpRQRETEkydPrKysevfurehaiPKjGCVKiDuinzJliryWzyOkFhSjRNmUlyM93dLU1JybgJmQxkbnRomyOXYMzs54/312/ToNRUlToNEoUTYHDgDAlCmUoaSJ0GiUKJXnz2FkhMpKPHoEMzNFV0NaBhqNEqUSEIDycgwaRBlKmg7FKFEqBw8CwMcfK7oO0pLQQT1RHhkZsLCAigqysqCjo+hqSIshVHQBhMhNQQGGDoWuLmUoaVJ0UE+UwZw5SE1Fp05Yvx4uLoquhrQwFKNEGVy+jBUrACAjA/fuKboa0sJQjBJl0Lo1bG1x9Kii6yAtEsUoeYslJ2P9ety6BQCrV+Onn1BcrOiaSMtDl5jI2yc7O9vfP8Df3/3SJVUAeXkAoK4ODw98/TVcXRVcHmlpaDRK3ibHjh0bNmyYmZnZkiWLq6rOaGtjxgyMH//i0zFj0KaNQusjLRLdN0reGpGRkfb29mKxWCqVjhgxws1t5dChfdXUACAjA6qqSEwEn48uXaCrq+haSUtCB/XkrfHDDz8wxkxNTa9fv66np/fyR6am8PeHqysmTkRAgKIKJC0UHdSTt8O5c+dOnDihpaUVGRn5SoZyjIwAIDu7qQv7V0VFWLQIpaUAcOYM4uMVXRCRN4pR8haQSCTLli0DsHbtWhMTk9duY2wMAFlZTVlXnZSV4bffsH49AMTG4tEjRRdE5I1ilLwFtm/ffvv2bSsrq8WLF79pm2Y7GgVgZ4eYGCQkKLoO0jgoRklzl5eXt27dOgCbN28WiURv2kxXF2Ixnj9/cfjcrDx4gPBwTJqk6DpI46AYJc3dmjVrcnNzHR0dR40aVfuWhoYA8PRpU1RVd7dv48YNVFfDxgbHjim6GtIIKEZJs5aQkLB7926hUOjt7f2vG9vb/25ntyMnJ6kJCqs7f39UV2PBAvz6Kz1kpZwoRkmztmjRIolEsmTJks6dO//rxgUF2/74Y35Gxp0mKKwW5eXw8wN3Q3ZcHDw8sG0b1q3DjRvYvh0ODoqtjsgf3TdKmq+AgICIiAhDQ8OvvvqqLtsbGRkByFb0ZaaiIixYAMbg5oZDhxAYCGNjmJjA2RlOTjh1SrHVEfmj0ShppsrKylauXAlgw4YNunV7LKmZxCiAYcPg54fcXKirIzUVcXGwsgKAlBRFV0YaAcUoaaa+++67tLS07t27z5o1q45NuBi9du1aqaKv1guF8PSEhwdEImhro7QUGhrg8ZCaiupqxZZG5I9ilDRHT5482bRpEwBvb2+BQFDHVsbGxgCCgoK+HTwYQ4ciOhoAystRUtJolb6RgwMqKxEdDVNTAMjMhJERKiqQnt70tZDGRTFKmqOVK1eWlJRMnjx5wIABdW9VWVkJQF1dfaJIhNBQiESYPRuzZ2PxYri4NP0NpZs2ITz8xeNVyclo3557IWniMkhjoxglzU5UVNSRI0fU1NS8vLzq1TAkJITH4xkbG3fZuxcBAUhMhKkpDhzA3r0YOBA7djRSwa/Q0sKcOZBIYGyMEyfQv/9JG5uuaWkbevZcq6mpnZy8t2nKIE2GYpQ0L1Kp9NNPP2WMffHFF23btq17w2fPnh0/fpwxZlpWxuvQASkpiI5GzWB24EBcv94oFf+DWPxinecePVBVBT+/wqSkW0lJCa1aqRYXF6WmpjZNGaTJUIyS5mXPnj3Xr183Nzf//PPP69Vw48aNxcXFnTt3/nLUKPD5sLGBhsZfB/LFxdDUlH+5b8BNyC8SISMDjx9bAUhOTrayevGiycogTYNilDQj2dnZCxYsAPDdd9+pq6vXq+GOHTt4PJ6vr++wXbvw7BlGjMCYMdi5ExUVkEqxbdtfs+Q3voICANDTQ34+gPYAkpOT27dvDyCFbnpSOhSjpBmpqKiQSCQAYmNjq+tzZ9C3335bWlo6bty4999/HwC0tCAWo0cPuLvDzQ1TpmDwYIwc2Uhl/1N+PgDo6nIvDEUi7YKCAn19fQDJycm05ISSoaeYSHPx7NmzyZMn6+joFBcXf/vtt9euXTt48OCbZhd9WVZWlo+PD4/HW7169aufjRmDMWMapdxa1YxGuReffPLbggWGxsbG3J2tpaWlGhoaTV8VaSQ0GiXNQlZW1pAhQy5fvmxnZ3fu3DlTU9Pz58937dr17Nmz/9p206ZNpaWl48ePfzEUbQa40WhNjH7wwcDOnTtfuXIlOzvb1NSUMlTJUIySZiE4OLioqAiAjo6Ovb19XFzc8OHDc3JyRo4cuXTp0qqqqjc1rBmKrl27tgnr/Rdcev55UA9u0ZNz584BGDx4sAILI42BYpQ0C7NmzYqJiVm4cOHQoUMBtG7dOjg42NvbWygUbtmyZciQIRkZGa9tyJ0VHT9+fNeuXZu25NowtqVHj5VGRlH6+oH29qf19dNBMarEGCEKFRYW5uzsPHHixB07dvzz04iICFNTUwCtW7c+c+bMK59mZmaqq6vzeLy4uLgmKbaunJ2dARw9evS9994DcPPmzYKCAoFAoKqqWlRUpOjqiJzRaJQo0pMnT1atWrVnz56AgIDXziha+wG+l5dXaWnphAkTmtVQFEBBQQEAXV3dmhcPo6I66On17t1bswlvXyVNg2KUKNLFixfHjx+vo6MDoH///q/d5pUDfEdHR+4APysra+fOnTwer46zkTalf8boe2fPJuTmBjfhTVekyVCMEkWqrKxUVVX91814PN7SpUvDwsJMTU0vXrzYtWvXM2fObNy4sayszNnZubkNRQHk5+cD0NbWLioqEggEWlpaCAsDY5pv+FVB3mo8RncCE8WJj49fsWJFcHAwj8ery/ZPnz6dNm3a77//zufz+Xx+dXV1bGxsM4xRAwODvLy8+/fvW1tbGxgYPEtIgIkJ1NWRl4c6/NogbxcajRJF6tKly6BBg4YOHerm5sbNdV87Q0PDs2fPcsvbaWhoDBgw4N133238MuuHMVZYWMjj8aRSKQBdXV1uKIoBAyhDlRI9xUQUafv27b169Vq+fHllZaVYLK5LEx6P5+zsvHLlSu6St76+vr29/dy5c4cNG1b3CZ4bVUpKipWVVUlurtqlSxMnTmzVqhXOnQMAutVJWSn4TgHSguXk5IhEIoFA8OjRo3o1nDdvHoCPPvqIMXblypX+/ftraGjY2NiMGjXqyJEjFRUVjVNvXXl6emoCEQYG7O7dF2+dOcPmzWPx8QqtizQWilGiMN9//z2AUaNG1atVWlqaqqqqQCC4c+cO945UKt2yZQt3v5SqqmrHjh2HDx++detWhdyhKZVKuQnxLhw/zhhjYWFs4kTm6spWrmRlZU1fD2kCFKNEYTp16gTgxIkT9WrFrXA3ffr0V94vLCycMWMGtxwTAKFQ2L59+9GjR0ul0srKygsXLoSGhpaWlsqv/NcLDw8H0KZNG4lEwtLTWf/+rLiYMcZ+/pl9+WVj750oBF1iIq/BPRIOoLIS5eWNsouIiIjExEQTE5OR9bmVMiUlZf/+/QKBYM2aNa98pK2t7evrGx4ebm9vr6GhIZFIkpOTi4qKqqqqhg0b9scffyQlJTXBc/d+fn4AZs6cKRAIEBmJ0aPBTUQyaxZCQxt770Qh6BITeQ0rK+zejfHjceoUHj3CsmXy30VgYDyAOXPmCIX1+Ef4zTffSCSS2bNn29jYvHYDW1vbiIgIX1/frVu33rlzZ+HChcePH+/Xr9+qVavkU3etiouLAwMDeTzejBkzAKCqCjXfTiCgtZWVFY1GyWvY2mLLFhQXN1b/z55h165F1tbls2cvrHurpKSkw4cPq6iovGZe0b9zc3OLjIzcsGGDs7Nzamrqax8zbQz+/v7FxcUffvghd3oUvXvj3Dlwt2b//jv69GmaMkgTo9Eo+ZuUFOjrQ1UVy5bB0xP9+jXKXvz8UFEBGxuRublh3Vt5enpKJJL58+e3a9fuXzdWV1f/7LPPABgZGaU31drwvr6+ANzc3F782doaY8Zg1CgYGKC8HD//3DRlkCZGTzERlJXhjz8QFoawMMTEwMcHR44gPBwTJqB7d2hqyv+gvnNnJCbit98wdmxdmyQkJLz33nuqqqr3799v06ZN3feVn58/ePDgH374wczMLCMjw8DAYMuWLVu2bFFTU2tI6W/24MEDKysrdXX1zMxMLS2tv30mkaA+5y7I24X+07Zcd+7ciYzE0aO2kZF/XUcyMPjrWP7HH9GnDzw85LzfCxeQmAgTk/qtjbR27VqpVDp37tx6ZSgAPT29U6dOHTp0KCIiYvDgwdOmTbt582ZsbOzx48fNzc3rV3qt9u3bxxibOHHiqxkKUIYqOQXfKUCaVnFxcWho6JIlS7gl4O3spgOMz2c9ezIPDxYayiorGWNs+fIX22/fzo4eZT/+yO7fl8fu9+5l7u7n2rvbI2Lt2nq0u3XrFp/PF4vFT548kbGE+Ph47sSlkZFRZOQlGXurUV1dzf2Vnj9/Xl59krcFxWhL8Z///GfAgAEqKio1v0GNjIzmz//s0CH27FltDX18GMAsLVlGhmwV7N3LrK0ZwIAKodq1sHsFBQV1bDp69GgAn3/+uWwVvJCbmzt06FAVFfWePfM2b5ZLlyw0NBTAO++8U11dLZ8eyduDjjVahPDwcG9v75ycHIFA0LNnT0dHRycnp379+vH5td2qUVWFFSswdy769UNUFIYOxcWLL5YVaoA474hu9+/78njOjGlJykY5dslBCDBIT09PVXUsj9fTxiZAW1tb5yV6enra2to5OTlBQUEaGhorVqxo4L7/Tl9fPzg4+Ntvk9as0YuJwd272LwZL/1+qbfc3Nx169YBcHNzq/2vlCglitEW4eTJkzk5OYMHDz527Ji2tnYdW33zDTZvxpkzOHUK48YhPh4jRyIsDLKsazmJ8dVRDcDY2DgvR1xdzU3NmQPkZGVdfFMrHR2d4cOHGxrW47J+7QQCwerVtubmcHfHzz8jIQEBAWjVCnw+JBJUVaGW608ZGRmJiYmpqakJCQnci9TUVD6fLxQKJ02aJK8KyVuErtQrP6lUamFhkZ6efvXq1bt3744ePVqvbkPK588xcCBu3EDXrjhwACNG4MkTjB3Ljh6trtc98wDy87Ft4sHl4TO4DD2r58I/NHe6W/9x44oBPHqk0rZtsavrvefPnxf+XUFBwd27d2/dumVhYZGQkCD3FThiYzF+PMaMwfnzcHfH4sW4fBnHj2PTJgCorERyMu7cwd27SEzE3bswMPg6NPSbVzrR0tIyNTVdvHjxwoX1uA2WKA2KUeUXHh4+ePBgKyurvXv32tvbd+jQ4e7du3Vsm5ODDz/EvXsYOBBbtmD4cNa27TZLy2v79+//14mWq6tx+TKCgnDiBPT18fDhvsL0XdMmLzI2HrE7IO9xeifgEuAKABiootLOwOB/Ojo63HG9np4ed1zfs2dPFxeXPn36REdHL126lJtpVL6ePoWeHoYMgZoa9uzBw4c4fhwZGYiORkoKJJK/bdyv36HU1OW2trYdO3a0tbW1tbXt0KGDfK/4k7ePok/OkkY3e/ZsAF999RU3v9yqVavq1TwlhZmYMCur8lmzVl25cktDQwPAihUr3rR9fj47fJh99BHT1eWuJzGA6emVA176+mNKSkrGjWNmZs/s7YdoasZaWloaGBgIBGOA1z/tPmXKFMbYzZs3VVRU+Hx+VFSUrH8db+DgwKKj2eTJLCqKrVjBunV7UbmJCXN0ZEuWMB8fFhrKsrIaaf/kLUYxquQqKysNDAwA3Lx5kzu3ePPmzfp2cuuW5J13egCYN29eWFiYSCQCsGnTppe3SUtL8/HxcXJyGjRodU16WlqyJUtYaCj74IO+wPzly39hjN2/z8aPZ1VVbM6cF20vXWL/+195enp6YmLilStXQkJC/P39fXx8Nm3aVDP/Ezc3/rvvvlvJ3ZMlbw4OjDG2YAFbv56tWMEuX2Zxcay8vDF2RZQNxaiSO3nyJID33nsvODgYQMeOHRvWz+XLl7lx6FdffXX48GE+n8/j8Xbt2nX16tXVq1e/vJKHlZXtwIHsxx9ZcvLfajA2Ni4pKWnwFyktLW3fvj2A//73vw3upBZcjOblMSsr9uahNiGvQTGq5GbMqO7VK3Pr1vDp06cDWL9+fYO7OnXqFHdl6ccff9yyZQsAbljK0dHRcXFxOXDgQF5e3isNP/jgAwA//fSTbF+FnT9/nsfjiUSixMREGbt6hUTy1xMH/v7Mx0e+3RMlRzGqzEpKmKYm4/HYnTusV6/FPB4vKSlJlg737dvH4/FWrlzJGOvcubOenp6xsbG7u/vJkyfL33AAfOLECdmHojW4WT/s7e2lUqnsvdVYs+bFuVFCGoBiVJkdOcIA1rcvCwxkABs+vB5zv98vv18ufU0yXrt2jTGWm5srEomEQmFmZmYtnUilUm4oullOTwvl5uYaGRkB2L17t1w6ZIwVFr64GnZJbo+GkpaFnrhQZkeOAMDkyTh8GACGDKltTiMG9kP2D+NTx09InRBcGPyfrP9kVGX8czMuFg8fPlxRUTFkyJCaRTte68SJE9evXzcxMZkzZ44M3+Mv+vr6P/30E4Dly5dnZLymvAbYtg0FBRg0CHZ2cumPtDyKznHSWAoLmVjM+Hx27x5TU2N8Pnv8uLbtj+Ufc0tzkzJphbTidulttzS3++X3K6Wvvyzeu3dvAIcOHaqlQ6lU+v777wPYunWrLF/kn7hH7F1cXGTvqqSEGRoygIWFyd4ZaaHoYVClFRiI8nIMGoSrV1FWBnt71D7D3Lmic7MMZvHAU+WpdlHrAuB22W3rBGtNvqa+UF9PoFfzE49x7do1saZY2EYYFhamr6+vp6enr6+vo6Pzcoe//fZbdHS0iYnJJ598It+vtm3btoiICH9//ylTpoyt+5Slr7NzJ54+Ra9etIY8aTiKUaV19iwATJ7816F97SRMIuAJXn6nlJWq8FSKpcXFlcWP8OivD3wBhvJB5T+c/uHqt1dr3ubz+VykcqkaHR0NYM2aNWpqag8fPmzVqpWGLE/jv8TCwmL9+vWffvrpggULHBwcXonvuquoqPj9909tbOavWfOeXAojLZSih8NEznJyWFwcY4xJJGzHDvboETM0ZEIhe/r0X0zMlbkAAAiHSURBVBr6PvP9Iv2Lmj+6pbmlVqQyxp5Lnj+seBhbGhv+PDwwP3Dn0102Dnbgwemw08IfFw4ePLhHjx7t2rX7Z5YZGBiIxeLo6Ojjx49raWktWLBAjl+zurq6b9++ABYtWtTgTnbu3Amga9eucr3sT1ocilFlc/o009ZmaWmMMTZhAsvOZuXldboGXSWtmv9o/tjksa6prruf7Z73cF5aRdo/N7t4kQHsgw8e/HNizerq6pycnKSkpCtXrpw5c8bJyQlAt27dbt68KRaLeTze2bNn5fAN/3Tr1i3uCdFLDbrELpFIrK2tARw5ckSOVZEWiKYmUTbBwQgMREEBAgPh7Iyff4b8ppcDAHd37NqFL77Axo3/smVJSUm3bt2Sk5O//vprFRWVNWvWmJmZxcfH6+rqylhDfn6+WCxWU1NbtWqVl5eXSCTS19c3NTW1tLQ0MTExNTXlflpaWpqbm6u8YSbRgwcPTp061crK6t69ewKB4LXbEFIXdG5UCXXogJwcHD8u/57LyxEQAAAff/zvG2toaPj6+trb22/YsOHSpUsnT568evXqihUrdu3aJWMZa9euPXr06N69ey0sLABUVFRkZmZmZmbGxMS8siWPxzM2NjY1NTUzM2vTpo2JiYm5ubmpqamBgcE333wDwNbWljKUyIhGo8omOBjx8ViwACNGQFsb+/bJczQaEAAXF/TsiejoujZZtmyZt7d3p06dDh061Ldv37KystOnT4+s12p2f5eZmWllZVVRURETEzNhwoQHDx4cOHCgb9++GRkZmZmZ3M/U1FTuxaNHjySvTHX3Jz6fL5VKBQJBZGQkd5qVkIah0ahy0tTE0qX46CM5d/vLLwAwbVo9mmzcuPH3339PTEz09/dfv379559/PmfOnNu3b+vr6zesho0bN5aVlbm6ukZHRz948MDW1vajjz7i8/mWlpb/3FgikWRnZz9+/DgzM/Px48cZGRnp6elHjx4tLy8XiUQODg5nzpyZOnVqXFzca5bzJKSOFH1ylsjZ3bssIuLF602bWHIy++wzVlUlh56fPWOqqkwoZLU+//kaly9fFggEQqEwKipqwIABAKZNm9awGjIyMtTU1Ph8/o0bN7jcPHz4cH074R7zF4lEMTExXbt2BTB37tyG1UMIoyv1yk0qZb17M4B98okceouPZwMGsFGjGtKWmy20Y8eOiYmJ3EIggYGBDehnyZIlAFxcXHbv3g3A1ta2YStxzpo1C0D37t3j4uLEYjGAoKCgBvRDCKMYVXrXrzMNDQawdevq1zA/n/36K+MWXf/sM8atDx8by4KDG1JGeXk5Nyfp8uXLuUn2WrdunZ2dXa9OMjMz1dTUeDxebGwsd69S7U+j1qK4uJibvfTLL7/ctGlTh3btrg8Y8C8rTRPyBhSjyi8oiAmFjMdj+/b9+8YpKczbmzk6MlVVBrCxYxljrEMHNmUKY4wdO8a+/76BZdy4cYO7zTMiImLo0KEAXF1d69XD0qVLAUyaNGnPnj0ArK2tJRJJA6th7MKFC3w+X0VFJfXKlRJHRwawiRMb3BtpyShGW4Tt2xnAVFRYWNhrJv0sKSk5derU/Pnzx40Lr1n/Q0WFDRzItm1jjDEHB7ZsGTt9WqYYZYx9+eWXACwtLRMSErhLOr/++msd22ZmZqqrq788FD148GDDS2GMMeb11VchPXsya2uWkMC0tRnAfvlFxj5JC0Qx2lJ4eLDevX8zNDSKjY3l3klPT9+2bduIESPU/lyU3dp6QqtWbNIk5ufHkpKYnx9zcWEREczBgRUWsn792KFDMsVoVVVVz549ASxevHjHjh0AWrVqlVW3VeKWLVsGwNnZed++fdxQtEr2C2fl5ey99xjA5s9ne/YwgOnosIcPZe2WtDAUoy2FVMo+/ngqADMzs8ePHzPGfv31Vy49+Xx+r169vv766+vXr9+6xby8mKMjEwpfDEuXLn2xTtHBg6xXL5lilDEWFxenqqrKPRg6cuRId3f3usyKn5WVxQ1Fb9y4YWNjA+DAgQMy1VEjPp6JxYzHY6dPM2dnBrBBg1iDLluRFotitAWprKx0dHQE0KlTp7y8vPz8fFdXV19f3/v37wcEBMycOXPUqJiag3qxmA0fzrZuZQ8evIhRxtigQbLGKGNs3bp1AN55552CgoI6Nqmqqtq3b9+nn37q6+srt6FoDS+vFysp373LDA0Zn//XLWOE1AE9xdSyPH/+/MMPP7x165a9vf2OHTtCQkKCgoIuXrxYWVkJoGvXJVlZm4cOxejRGD4cNTekJyXBxgYAcnMhkcDISKYaJBJJnz59YmJitLW1XV1ddd9MXV395YbV1dWdO3e+d+/eL7/8MnXqVJmKeJlUikGDcOECxo+HuztUVTFokNw6Jy0AxWiL8/Dhw379+nFrKHHvCIVCOzu7kSNHjhw5qkuXzk1QQ1BQ0Pjx49/0mGYNVVVVHR2dmlRNTU1NSUmxtra+c+eOnB+ET01Ft24YNw579iAvD+fPQyjE0KHQ1pbnXoiSohhtia5fvz558uS8vLwhQ4Y4OTmNHj1aT0+viWtITk4+ePCgqalpwZuVlpb+s+Hnn3/+3Xffyb+gR49gYYHUVLi5YdkylJZi+3acPg2Z56MiSo9itOVijPF4PEVXUZvKysqCgoLCwkIuVWNiYjp37swtxNRYli/HsGEYOhQAfHxQWYnFixtxd0Qp0NQkLVczz1AAqqqqhoaGhn9OUTVkyJBG3+XDh2jf/sVra2ucPt3oeyRvP1pgmZCXmJsjLe3F6wcPYGGhyGLIW4JGo4S8ZN48zJ0LFRWUlGDPHgQFKbog8hagc6OE/F1aGoKDIRRi/Hi0bq3oashbgGKUEEJkQudGCSFEJhSjhBAiE4pRQgiRCcUoIYTIhGKUEEJkQjFKCCEy+T8S9vIhTnCzawAAA256VFh0cmRraXRQS0wgcmRraXQgMjAyNC4wOS40AAB4nH2RbWxTVRjHzz29vbft3bqutOvabu2Fvg42HQESP6z3nExZSMYCxhjUjO3Iy7yTieGDBrLBmEbNaIQsZpFoFt6MbGQswkI3sd7nDOSDryhvisAkvIVAYBCGHxxBO5ouvARP8s/vnOc8+T//kzNm7B1FmeXICKPsKsvo2Yw2ChYUz9CEpSxFieaoT9IkPLTJ3hQiNcMn65IanqyLEmKTk7CcdTTlzrk+OUtTjlNGFvpolFy0KaPHoz3h8H/hn/acqXjq3Ek+nP9BJxYeD/70+bmCgoQ4FrCKTTo2iUw069gsIbPMZIuOLVZktYmCTUFKHsrLR/l2HdsLWIFDx45CZnUi57Q4drpUl1t1F+m4yMM8xTou9jKvT8c+P/OX6LiklJUGdOwKoqCK1OmiMH0GmhFSQ2EdhyMsEtVxNMZicR0rglqYzwJuFg8xC2aqFQW8rFrOpJQExWa1YMkkmmULNtvzCx0FZqdVDbqkIk+xN+CWfP6S0oBXCkeisXjI87eQ+VyUFSobOGunPv9b3C3cN2D1EXKppoK32p8zHB2VmiqE+fpEDObW7DJWli/n85vScGrwJxiQG3jr6Xvwx3E3n7XXwyfWHIWAbQNfsMnBR5sc/M7la3CwbwAmjiyAyJgN5rcMQ3d10uhY0phOlx81kru9pD2+mSRfaIYDPUPkla8w/X7RSWhIRMnpika67vwQpHpnkobP3qV7+A3jcHTEoK42ak/VE3F4u3Hx10467fk+svgOh/SazTR8e4t2cImLv9z0AVU6F8LXL/0MkSqZjCkV5G25xxiX1pP6TxRyzJ+GruETWvmZcXJBO6PdO9tIkkodDb0xJ9H/oUxLtrXQf2oHtX3lAWrf30P/DL72rfbxb2SpkaLnu6thnTSi/cB6adm/SVh2S4TdWzmdqL8KvWsvQuWOffS9mxZ+u8vKfW920tVwDtY6tkN60TL6zswOaP5uVdXsiXa60PkMmOs+MpzBU+TFK7+TRN/7cM6W0u6OHCBbqnphp+1VI52spJ82X4dDicOQqptHf/m8hve0jUNtm4sm+1v59e6NkNc1Surav+Awpx9a/NvIYOd+/o03xtvvf0n++nErHx96nS+vvEVmFazgm/y1/Br30KL/AFMKJLAiQX4yAAAEZHpUWHRNT0wgcmRraXQgMjAyNC4wOS40AAB4nH1Wy27gNgy85yv0A2vwKYmHHjbJdlEUmwBt2n8o0GP/Hx3SiaSgQJ2YsJUxNSSHVB5aXr89//rXP21d+vzw0Br9z29EtD+ViB5+tHxoj9++//LSnt6+Pn6sPL3+8fL2ezNuNvANfj5jv769/vhY4fba5BIV4Whf9BpkY+CJLovOJPtbaU+NLzcxk/ZFgBzmkkgeMUw3UoHMVdE+EuldvJfP7irz8GlAYll5zvDc3i184okvGzwHb6gnFPtHiM6EWo/izFcM1zk2tCdULjUm93rqEWYJ7UK99w0d7envdBY9tM/Eik8zup+6IBULO28GU50zAwy35rkBXWOanGTjjstztaBwOclyDSsIYUOZbiyxdq10GllHiMgWkTDy2p5+/v4Trw+4CsFMHY/AB7OgUNhEzemoNrZ+QXHJBM4r+t4tRnq20GEHMkuGlCLjo5CmKFVxwKVHbtkKOV1R32LbgS0eOobYkS92bI/UojaKlF+zC7ve9dagc/sOpF1KfcABEiuiXqV1iEz8QKJgQJqzTji4ZDJriQx8pp8+J5RtF/c5rIiOmIHSZxrIGH/e0CjBOqSJSBDz8Og3UagljuiF4BQZn7MEhwK7gFImrAekeSAZISE77BAyQgqW0MgV7MIHT5HaHCwN7PQS0ollCMJ0ntWUrJGg7mDnAIKYcNJVJ57n1lZlx742JlIQUPEsmRLaz49GFS8pocAM734NppHNhWRBe5/C6cUSJSZKJBI1ufqUoCQ54xnv2udAxK1D2LNrNR++UD3yLu8dNQb2R4df7hTYCdWEQNzPxMfd04z5oIqgJCxFn1+r2CkmpdurODm4YqZAn3eREBPTQUC5oiI26DGlGoGpVoNqRPRjf5U7KumQY0q0D9RR82H2wcdEUb23J5k9xyfKg6vaE42fw6/aWRbebryChFXuUVUEjwoOjEI5iqDZUDWBgjtVe6CcEJBgCdudkdUM1ItJ5YYicpQfUBWieRIe6TX7FHnsd5rMh7dM7Zx0jDWtgjmqQG7lFRNHJfMhpnTKAFp/SaiPUIosHU6TsPSqmG92HgJVMIMMbKKr8hSQjgED6Bwdx9EB5ZurQ1EpFAQ4kN+eYQ1mP9rl28vzp4PuPvoeX1+e99Enee/zLRd0H2KW9z6p8sf3aYSX1veJw7jHPlUMr3MfHIY79uHAeZ8nAJfh/wx6LrMocnJMs1hicHOZRZSTJwwvrpxkYXjRhV4tDS/GnJRheJHGgLxZLd4Yk1LmHIdcK4uzJGcY2WkVeE6NLs6SqYWRxVmSM4wszpKcYWRxluQMI4uzJGcYWZwlMw0jizO0aWn0HA/WqvePKVCcdfdk5hlGtxoyzzDZrWeN8M5pdDFHk3IaXcw1mcPoYq7JHEYXc03mMLqYazJPc/aJlVm7G996WMQtk52V27otHZ8FsZIyHwQtkx1b+tbvYLdEbbx7Wd9ke53NlO8f/4Pi+eFfCWcGhSo0Xk4AAAKGelRYdFNNSUxFUyByZGtpdCAyMDI0LjA5LjQAAHicNVO7btxAEPuVlHeALMz7EcGAgWtS2VWqwJXawCnT+OPDWTlqvGfucjgcztvzg085z/P2+H0/T/n1eHl5l9vjlR8nf3ycf/j+uD2/3V9vJ58f56nnfKz3x9z88S6PV9wD8sHfPm+yi4rI9qR7kmVtT7RbB5McvLuJ2fYkgNJ8IM5O02MOopGDeYjnYOEqJQdOylXtw+nWDk7eLbnywMG7xQay6KnLe6drH2BSY/Kh1Oi2gUIo8Er3jtZVTbwuSYK6NYSlzuuyhvkSmWWSPEJ8DquIGa3e8BsCBiPWUF7KjSy0NrRAJKzonJmCB2pmIdlAoOaH7GRCvKRF+LwwCLMD7VQO2VwMt0HwaQIp/18mgC1WzRRrYK2Vc7lC2HVBrE122K5o3VEY49EYYqex3nZzXmKlmC9TwF8OhGPYxoGujvUoyFgwL8doUBuy0juuShTRCayrIGMbu1y+mogOIMaOwUFEs/ToNNCwAUEhQwXdhbQUCJlm91jEjBJAQC5TUZ24GgCorDbbW3BzFBAS42M3jMKYfE+mrwYwie5jOUWkgKC+1iQJhq8BOre0boFJ1nSEKMBDm0xkXnzuuD0OwVSPyRgjkAoNaGexpcp6IY7oIa8Y0FWFMPljUmKYBwPq7lzDC6ifVGICYYJWoT/hy/ytGG14LhWrdXwrPUjdSquCb4VUmBtxxW45AnIFufkaNEwRIIV9WulnUllIM9G8UYG+tReQz5dew6JNP4U6jj7Jr60wVwUgpthpIJ6tPJ5gyXsFDOmHBwbXrHptrUQNUhlWU8ZDTQfhVI8RkMx+3/7+/N5Qtyn+U5//AA4C8KuvQN0VAAAAAElFTkSuQmCC",
      "text/plain": [
       "<rdkit.Chem.rdchem.Mol at 0x163d447c0>"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "deserialized = Chem.Mol(base64.b64decode(y_pred_serialized[0].encode(\"ascii\")))\n",
    "deserialized"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "c4621850-73e4-4b69-b2a0-0a5cca306a78",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<rdkit.Chem.rdchem.Mol at 0x1645a0660>"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "0fcda166-1747-47a2-a50a-b8e6b71d5a9b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dm.same_mol(y_pred[0], deserialized)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2fb2929f-961f-42ac-b693-b0600fa198a1",
   "metadata": {},
   "source": [
    "# Submit to Polaris"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "2b225605-ff8a-42fb-8ff2-5f929ac036ae",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "✅ SUCCESS: \u001b[1mYour competition predictions have been successfully uploaded to the Hub for evaluation.\u001b[0m\n",
      " \n"
     ]
    }
   ],
   "source": [
    "competition.submit_predictions(\n",
    "    predictions=y_pred_serialized,\n",
    "    prediction_name=\"my-second-predictions\",\n",
    "    prediction_owner=\"cwognum\",\n",
    "    report_url=\"https://www.example.com\", \n",
    "    # The below metadata is optional, but recommended.\n",
    "    github_url=\"https://github.com/polaris-hub/polaris\",\n",
    "    description=\"Just testing the Polaris API here!\",\n",
    "    tags=[\"tutorial\"],\n",
    "    user_attributes={\"Framework\": \"RDKit\", \"Method\": \"Dummy Baseline\"}\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3b52c5dc-9921-420c-9dfc-ac0408d38a50",
   "metadata": {},
   "source": [
    "For the ASAP competition, we will only evaluate your latest submission. \n",
    "\n",
    "The results will only be disclosed after the competition ends."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1f77ead1-89ba-4b49-b522-0b39afa48ae1",
   "metadata": {},
   "source": [
    "The End."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}