{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "4760ef5ca3ad7b8a",
   "metadata": {},
   "source": [
    "# Fewshot Tutorial"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5591e5ea0b8550e6",
   "metadata": {},
   "source": [
    "This tutorial demonstrates how to finetune the scLinguist model on a fewshot dataset and then use it to predict protein expression from RNA data."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c20285aad68c6452",
   "metadata": {},
   "source": [
    "Import necessary packages and define paths for checkpoints and save directory."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "initial_id",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2025-08-31T04:26:08.061560Z",
     "start_time": "2025-08-31T04:26:04.537193Z"
    },
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import torch\n",
    "from pathlib import Path\n",
    "from torch.utils.data import DataLoader\n",
    "import sys\n",
    "sys.path.append('../../')\n",
    "from scLinguist.data_loaders.data_loader import scMultiDataset\n",
    "from scLinguist.model.configuration_hyena import HyenaConfig\n",
    "from scLinguist.model.model import scTrans\n",
    "import importlib, sys\n",
    "sys.modules['model'] = importlib.import_module('scLinguist.model')\n",
    "\n",
    "ENCODER_CKPT = Path(\"../../pretrained_model/encoder.ckpt\")\n",
    "DECODER_CKPT = Path(\"../../pretrained_model/decoder.ckpt\")\n",
    "FINETUNE_CKPT = Path(\"../../pretrained_model/finetune.ckpt\")\n",
    "SAVE_DIR = Path(\"../../docs/tutorials/fewshot_output\")\n",
    "SAVE_DIR.mkdir(exist_ok=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "162900a151b5b992",
   "metadata": {},
   "source": [
    "Configure dataloaders for fewshot and test datasets.\n",
    "\n",
    "First, inspect the data structure of the datasets to ensure they are compatible with the model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "bff7d74c1ed6cbae",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2025-08-31T03:17:10.791866Z",
     "start_time": "2025-08-31T03:17:10.734023Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "AnnData object with n_obs × n_vars = 20 × 19202\n",
       "    obs: 'orig.ident', 'nCount_RNA', 'nFeature_RNA', 'nCount_ADT', 'nFeature_ADT', 'lane', 'donor', 'celltype.l1', 'celltype.l2', 'RNA.weight', 'cell_type'"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import scanpy as sc\n",
    "rna_train = sc.read_h5ad('../../data/fewshot_sample_rna.h5ad')\n",
    "rna_train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "406faeec6e69a11f",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2025-08-31T03:17:11.957905Z",
     "start_time": "2025-08-31T03:17:11.947240Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "matrix([[ 0.,  0.,  0., ..., 15.,  1., 27.],\n",
       "        [ 0.,  0.,  0., ...,  0.,  0.,  2.],\n",
       "        [ 0.,  0.,  0., ...,  2.,  0.,  7.],\n",
       "        ...,\n",
       "        [ 0.,  0.,  0., ...,  2.,  0.,  1.],\n",
       "        [ 0.,  0.,  0., ..., 31.,  0., 69.],\n",
       "        [ 0.,  0.,  0., ...,  2.,  0.,  8.]], dtype=float32)"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rna_train.X.todense()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "4e26d6ed540d442d",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2025-08-31T03:17:15.306723Z",
     "start_time": "2025-08-31T03:17:15.297132Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['ENSG00000186092', 'ENSG00000284733', 'ENSG00000284662',\n",
       "       'ENSG00000187634', 'ENSG00000188976', 'ENSG00000187961',\n",
       "       'ENSG00000187583', 'ENSG00000187642', 'ENSG00000188290',\n",
       "       'ENSG00000187608',\n",
       "       ...\n",
       "       'ENSG00000198712', 'ENSG00000228253', 'ENSG00000198899',\n",
       "       'ENSG00000198938', 'ENSG00000198840', 'ENSG00000212907',\n",
       "       'ENSG00000198886', 'ENSG00000198786', 'ENSG00000198695',\n",
       "       'ENSG00000198727'],\n",
       "      dtype='object', length=19202)"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rna_train.var_names"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "1c06bdf4dddb39e3",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2025-08-31T03:17:17.433460Z",
     "start_time": "2025-08-31T03:17:17.312128Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "AnnData object with n_obs × n_vars = 10546 × 19202\n",
       "    obs: 'orig.ident', 'nCount_RNA', 'nFeature_RNA', 'nCount_ADT', 'nFeature_ADT', 'lane', 'donor', 'celltype.l1', 'celltype.l2', 'RNA.weight', 'cell_type'"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import scanpy as sc\n",
    "rna_test = sc.read_h5ad('../../data/test_sample_rna.h5ad')\n",
    "rna_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "4c53fdea8be7bd78",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2025-08-31T03:17:18.880122Z",
     "start_time": "2025-08-31T03:17:18.566271Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "matrix([[ 0.,  0.,  0., ...,  5.,  0.,  3.],\n",
       "        [ 0.,  0.,  0., ...,  0.,  0.,  7.],\n",
       "        [ 0.,  0.,  0., ...,  1.,  0.,  6.],\n",
       "        ...,\n",
       "        [ 0.,  0.,  0., ...,  1.,  0.,  2.],\n",
       "        [ 0.,  0.,  0., ...,  2.,  0.,  2.],\n",
       "        [ 0.,  0.,  0., ...,  4.,  0., 12.]], dtype=float32)"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rna_test.X.todense()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "2c6747360b131ca2",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2025-08-31T03:17:20.270653Z",
     "start_time": "2025-08-31T03:17:20.261668Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['ENSG00000186092', 'ENSG00000284733', 'ENSG00000284662',\n",
       "       'ENSG00000187634', 'ENSG00000188976', 'ENSG00000187961',\n",
       "       'ENSG00000187583', 'ENSG00000187642', 'ENSG00000188290',\n",
       "       'ENSG00000187608',\n",
       "       ...\n",
       "       'ENSG00000198712', 'ENSG00000228253', 'ENSG00000198899',\n",
       "       'ENSG00000198938', 'ENSG00000198840', 'ENSG00000212907',\n",
       "       'ENSG00000198886', 'ENSG00000198786', 'ENSG00000198695',\n",
       "       'ENSG00000198727'],\n",
       "      dtype='object', length=19202)"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rna_test.var_names"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "9b84f4dda663c60c",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2025-08-31T03:17:23.161588Z",
     "start_time": "2025-08-31T03:17:22.415663Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "AnnData object with n_obs × n_vars = 16994 × 6427\n",
       "    obs: 'orig.ident', 'nCount_RNA', 'nFeature_RNA', 'nCount_ADT', 'nFeature_ADT', 'lane', 'donor', 'celltype.l1', 'celltype.l2', 'RNA.weight', 'cell_type'"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import scanpy as sc\n",
    "adt_train = sc.read_h5ad('../../data/train_sample_adt.h5ad')\n",
    "adt_train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "497bd4321c403102",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2025-08-31T03:17:34.433262Z",
     "start_time": "2025-08-31T03:17:34.211569Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "matrix([[1.100e+02, 1.400e+01, 4.900e+01, ..., 1.100e+01, 2.120e+02,\n",
       "         2.800e+01],\n",
       "        [1.200e+02, 1.000e+00, 5.890e+02, ..., 2.000e+00, 3.600e+01,\n",
       "         2.400e+01],\n",
       "        [6.450e+02, 5.000e+00, 1.256e+03, ..., 1.000e+00, 7.200e+01,\n",
       "         1.320e+02],\n",
       "        ...,\n",
       "        [2.330e+02, 2.700e+01, 8.420e+02, ..., 3.000e+00, 7.700e+01,\n",
       "         4.600e+01],\n",
       "        [3.120e+02, 1.500e+01, 1.079e+03, ..., 4.000e+00, 4.800e+01,\n",
       "         8.000e+01],\n",
       "        [1.960e+02, 2.000e+00, 1.910e+02, ..., 0.000e+00, 3.400e+01,\n",
       "         1.900e+01]])"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import numpy as np\n",
    "mask = ~np.isnan(adt_train.X[0].toarray())\n",
    "adt_train[:, mask].X.todense()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "23db37c02a2cc70e",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2025-08-31T03:20:13.452987Z",
     "start_time": "2025-08-31T03:20:13.444177Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['SP110', 'GTPBA', 'SNX2', 'FRG1', 'TT21A', 'RHG18', 'AR', 'DOCK1',\n",
       "       'RAB1A', 'MUC1.HMFG2',\n",
       "       ...\n",
       "       'CYTSA', 'LFNG', 'PFKFB4', 'LIPB1', 'ZN225', 'TRI69', 'CCL14', 'ZN541',\n",
       "       'TAP1', 'SCG3'],\n",
       "      dtype='object', length=6427)"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "adt_train.var_names"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "19c888ff6fe5f34a",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2025-08-31T03:17:37.257945Z",
     "start_time": "2025-08-31T03:17:36.797743Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "AnnData object with n_obs × n_vars = 10546 × 6427\n",
       "    obs: 'orig.ident', 'nCount_RNA', 'nFeature_RNA', 'nCount_ADT', 'nFeature_ADT', 'lane', 'donor', 'celltype.l1', 'celltype.l2', 'RNA.weight', 'cell_type'"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import scanpy as sc\n",
    "adt_test = sc.read_h5ad('../../data/test_sample_adt.h5ad')\n",
    "adt_test"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3dff4dd5528c93a8",
   "metadata": {},
   "source": [
    "Then, create the dataloaders for both fewshot and test datasets. The `scMultiDataset` class is used to load the RNA and protein data from the specified paths."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "80ddcdbf5b659c88",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2025-08-31T03:18:07.226968Z",
     "start_time": "2025-08-31T03:18:04.587677Z"
    }
   },
   "outputs": [],
   "source": [
    "BATCH_SIZE = 4\n",
    "fewshot_data = scMultiDataset(\n",
    "    data_dir_1=\"../../data/fewshot_sample_rna.h5ad\",\n",
    "    data_dir_2=\"../../data/fewshot_sample_adt.h5ad\",\n",
    ")\n",
    "test_data = scMultiDataset(\n",
    "    data_dir_1=\"../../data/test_sample_rna.h5ad\",\n",
    "    data_dir_2=\"../../data/test_sample_adt.h5ad\",\n",
    ")\n",
    "fewshot_dataloader = DataLoader(\n",
    "    fewshot_data,\n",
    "    batch_size=BATCH_SIZE,\n",
    "    shuffle=True,\n",
    "    num_workers=8,\n",
    "    pin_memory=True\n",
    ")\n",
    "test_dataloader = DataLoader(\n",
    "    test_data,\n",
    "    batch_size=BATCH_SIZE,\n",
    "    shuffle=False,\n",
    "    drop_last=False,\n",
    "    num_workers=0,\n",
    "    pin_memory=True,\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "37a4e2fd1165741d",
   "metadata": {},
   "source": [
    "Last, configure the model with the appropriate encoder and decoder checkpoints, and set the mode to \"RNA-protein\". The `HyenaConfig` class is used to define the model configuration parameters such as `d_model`, `emb_dim`, `max_seq_len`, `vocab_len`, and `n_layer`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "97bf208af1652d7d",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2025-08-31T03:18:11.653358Z",
     "start_time": "2025-08-31T03:18:07.266110Z"
    }
   },
   "outputs": [],
   "source": [
    "enc_cfg = HyenaConfig(\n",
    "    d_model        = 128,\n",
    "    emb_dim        = 5,\n",
    "    max_seq_len    = 19202,\n",
    "    vocab_len      = 19202,\n",
    "    n_layer        = 1,\n",
    "    output_hidden_states=False,\n",
    ")\n",
    "dec_cfg = HyenaConfig(\n",
    "    d_model        = 128,\n",
    "    emb_dim        = 5,\n",
    "    max_seq_len    = 6427,\n",
    "    vocab_len      = 6427,\n",
    "    n_layer        = 1,\n",
    "    output_hidden_states=False,\n",
    ")\n",
    "model = scTrans.load_from_checkpoint(checkpoint_path=FINETUNE_CKPT)\n",
    "model.encoder_ckpt_path = ENCODER_CKPT\n",
    "model.decoder_ckpt_path = DECODER_CKPT\n",
    "model.mode = \"RNA-protein\""
   ]
  },
  {
   "cell_type": "markdown",
   "id": "15572e727148bd7e",
   "metadata": {},
   "source": [
    "Start training the model using PyTorch Lightning. The `ModelCheckpoint` callback is used to save the best model based on validation loss, and the `EarlyStopping` callback is used to stop training if the validation loss does not improve for a specified number of epochs."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "d21252083680bc85",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2025-08-31T03:20:13.397572Z",
     "start_time": "2025-08-31T03:19:15.028049Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "GPU available: True, used: True\n",
      "TPU available: False, using: 0 TPU cores\n",
      "IPU available: False, using: 0 IPUs\n",
      "LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0,1]\n",
      "\n",
      "  | Name       | Type             | Params\n",
      "------------------------------------------------\n",
      "0 | encoder    | scHeyna_enc      | 313 K \n",
      "1 | decoder    | scHeyna_dec      | 249 K \n",
      "2 | translator | MLPTranslator    | 284 M \n",
      "3 | cos_gene   | CosineSimilarity | 0     \n",
      "4 | cos_cell   | CosineSimilarity | 0     \n",
      "------------------------------------------------\n",
      "285 M     Trainable params\n",
      "0         Non-trainable params\n",
      "285 M     Total params\n",
      "1,141.275 Total estimated model params size (MB)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "cc14adc5d8ba4306b1d39b88c45ece7a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation sanity check: 0it [00:00, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "8311eb35fab24d01b8299e214daba15d",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Training: 0it [00:00, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "c6dcb6ff6357417da461fd36fca2af04",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validating: 0it [00:00, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "da76a1e996b64ba6aad6e9fcac2ce634",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validating: 0it [00:00, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a565eabac1a5413f8a5f5a2cc0ab32b0",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validating: 0it [00:00, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2e3aa5ec7fff4066ab255ff26ab7f589",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validating: 0it [00:00, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f70afd3c542d4d27afb1a233c32cefba",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validating: 0it [00:00, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "73b2bf0fb4cf46d38f8eca5399a3a8e3",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validating: 0it [00:00, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import pytorch_lightning as pl\n",
    "from pytorch_lightning.callbacks import ModelCheckpoint, EarlyStopping\n",
    "\n",
    "ckpt_cb = ModelCheckpoint(\n",
    "    dirpath      = SAVE_DIR/\"ckpt\",\n",
    "    monitor      = \"valid_loss\",\n",
    "    mode         = \"min\",\n",
    "    save_top_k   = 1,\n",
    "    filename     = \"best-{epoch}-{valid_loss:.4f}\",\n",
    ")\n",
    "early_cb = EarlyStopping(monitor=\"valid_loss\", mode=\"min\", patience=3)\n",
    "\n",
    "trainer = pl.Trainer(\n",
    "    accelerator       = \"gpu\",\n",
    "    devices           = [1],\n",
    "    max_epochs        = 6,\n",
    "    log_every_n_steps = 50,\n",
    "    callbacks         = [ckpt_cb, early_cb],\n",
    ")\n",
    "\n",
    "trainer.fit(model, fewshot_dataloader, fewshot_dataloader)\n",
    "best_ckpt = ckpt_cb.best_model_path"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "874f62a4f446a010",
   "metadata": {},
   "source": [
    "Inference with the trained model on the test dataset. Use RNA data to predict proteins in ../../docs/tutorials/protein_names.txt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b597c4c88b552386",
   "metadata": {
    "jupyter": {
     "is_executing": true
    }
   },
   "outputs": [],
   "source": [
    "import scanpy as sc\n",
    "import torch\n",
    "\n",
    "# only use 10 cells for example\n",
    "fewshot_adata = sc.read_h5ad(\"../../data/test_sample_rna.h5ad\")[:10]\n",
    "sc.pp.normalize_total(fewshot_adata, target_sum=10000)\n",
    "sc.pp.log1p(fewshot_adata)\n",
    "fewshot_rna_tensor = torch.tensor(fewshot_adata.X.todense(), dtype=torch.float32).cuda()\n",
    "\n",
    "model.eval().cuda()\n",
    "\n",
    "with torch.no_grad():\n",
    "    _, _, protein_pred = model(fewshot_rna_tensor)\n",
    "\n",
    "# predict given proteins\n",
    "target_proteins = [line.strip() for line in open(\"../../docs/tutorials/protein_names.txt\")]\n",
    "\n",
    "import pandas as pd\n",
    "prot_map = pd.read_csv(\"../../docs/tutorials/protein_index_map.csv\")\n",
    "name_to_idx = dict(zip(prot_map[\"name\"], prot_map[\"index\"]))\n",
    "\n",
    "idx = [name_to_idx[p] for p in target_proteins if p in name_to_idx]\n",
    "\n",
    "pred_df = pd.DataFrame(\n",
    "    protein_pred[:, idx].cpu().numpy(),\n",
    "    columns = target_proteins,\n",
    "    index   = fewshot_adata.obs_names,\n",
    ")\n",
    "pred_df.to_csv(SAVE_DIR/\"predicted_protein_expression.csv\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "py3.8",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "3.8.16"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}