torch-molecule 0.1.0.dev2__py3-none-any.whl

torch_molecule/encoder/attrmask/modeling_attrmask.py

@@ -0,0 +1,304 @@
+import os
+import numpy as np
+from tqdm import tqdm
+from typing import Optional, Union, Dict, Any, Tuple, List, Callable, Literal, Type
+import warnings
+from dataclasses import dataclass, field
+
+import torch
+from torch_geometric.loader import DataLoader
+from torch_geometric.data import Data
+
+from .model import GNN
+from ..constant import GNN_ENCODER_MODELS, GNN_ENCODER_READOUTS, GNN_ENCODER_PARAMS
+from ...base import BaseMolecularEncoder
+from ...utils import graph_from_smiles
+
+ALLOWABLE_ENCODER_MODELS = GNN_ENCODER_MODELS
+ALLOWABLE_ENCODER_READOUTS = GNN_ENCODER_READOUTS
+
+@dataclass
+class AttrMaskMolecularEncoder(BaseMolecularEncoder):
+    """This encoder implements a GNN-based model for molecular representation learning
+    using the attribute masking pretraining strategy.
+
+    References
+    ----------
+    - Paper: Strategies for Pre-training Graph Neural Networks (ICLR 2020) https://arxiv.org/abs/1905.12265
+    - Code: https://github.com/snap-stanford/pretrain-gnns/tree/master/chem
+
+    Parameters
+    ----------
+    mask_num : int, default=0
+        Number of atom features to mask during pretraining. If set to 0, masking is determined by `mask_rate`.
+    mask_rate : float, default=0.15
+        Proportion of atoms to mask randomly. Ignored if `mask_num` is set.
+    num_layer : int, default=5
+        Number of GNN layers.
+    hidden_size : int, default=300
+        Dimension of hidden node features.
+    drop_ratio : float, default=0.5
+        Dropout probability.
+    norm_layer : str, default="batch_norm"
+        Type of normalization layer to use. One of ["batch_norm", "layer_norm", "instance_norm", "graph_norm", "size_norm", "pair_norm"].
+    encoder_type : str, default="gin-virtual"
+        Type of GNN architecture to use. One of ["gin-virtual", "gcn-virtual", "gin", "gcn"].
+    readout : str, default="sum"
+        Method for aggregating node features to obtain graph-level representations. One of ["sum", "mean", "max"].
+    batch_size : int, default=128
+        Number of samples per batch for training.
+    epochs : int, default=500
+        Maximum number of training epochs.
+    learning_rate : float, default=0.001
+        Learning rate for optimizer.
+    grad_clip_value : float, optional
+        Maximum norm of gradients for gradient clipping.
+    weight_decay : float, default=0.0
+        L2 regularization strength.
+    use_lr_scheduler : bool, default=False
+        Whether to use a learning rate scheduler.
+    scheduler_factor : float, default=0.5
+        Factor by which to reduce the learning rate when plateau is detected.
+    scheduler_patience : int, default=5
+        Number of epochs with no improvement after which learning rate will be reduced.
+    verbose : bool, default=False
+        Whether to print progress information during training.
+    model_name : str, default="AttrMaskMolecularEncoder"
+        Name of the encoder model.
+    """
+    # Task related parameters
+    mask_num: int = 0
+    mask_rate: float = 0.15
+
+    # Model parameters
+    num_layer: int = 5
+    hidden_size: int = 300
+    drop_ratio: float = 0.5
+    norm_layer: str = "batch_norm"
+    
+    encoder_type: str = "gin-virtual"
+    readout: str = "sum"
+    
+    # Training parameters
+    batch_size: int = 128
+    epochs: int = 500
+    learning_rate: float = 0.001
+    grad_clip_value: Optional[float] = None
+    weight_decay: float = 0.0
+    
+    # Scheduler parameters
+    use_lr_scheduler: bool = False
+    scheduler_factor: float = 0.5
+    scheduler_patience: int = 5
+    
+    # Other parameters
+    verbose: bool = False
+    model_name: str = "AttrMaskMolecularEncoder"
+    
+    # Non-init fields
+    fitting_loss: List[float] = field(default_factory=list, init=False)
+    fitting_epoch: int = field(default=0, init=False)
+    model_class: Type[GNN] = field(default=GNN, init=False)
+    
+    def __post_init__(self):
+        """Initialize the model after dataclass initialization."""
+        super().__post_init__()
+        if self.encoder_type not in ALLOWABLE_ENCODER_MODELS:
+            raise ValueError(f"Invalid encoder_model: {self.encoder_type}. Currently only {ALLOWABLE_ENCODER_MODELS} are supported.")
+        if self.readout not in ALLOWABLE_ENCODER_READOUTS:
+            raise ValueError(f"Invalid encoder_readout: {self.readout}. Currently only {ALLOWABLE_ENCODER_READOUTS} are supported.")
+
+    @staticmethod
+    def _get_param_names() -> List[str]:
+        """Get parameter names for the estimator.
+
+        Returns
+        -------
+        List[str]
+            List of parameter names that can be used for model configuration.
+        """
+        return ["mask_num", "mask_rate"] + GNN_ENCODER_PARAMS.copy()
+    
+    def _get_model_params(self, checkpoint: Optional[Dict] = None) -> Dict[str, Any]:
+        params = {
+            "num_layer": self.num_layer,
+            "hidden_size": self.hidden_size,
+            "drop_ratio": self.drop_ratio,
+            "norm_layer": self.norm_layer,
+            "readout": self.readout,
+            "encoder_type": self.encoder_type,
+            "mask_num": self.mask_num,
+            "mask_rate": self.mask_rate
+        }
+        
+        if checkpoint is not None:
+            if "hyperparameters" not in checkpoint:
+                raise ValueError("Checkpoint missing 'hyperparameters' key")
+            hyperparameters = checkpoint["hyperparameters"]
+            params = {k: hyperparameters.get(k, v) for k, v in params.items()}
+            
+        return params
+        
+    def _convert_to_pytorch_data(self, X):
+        """Convert numpy arrays to PyTorch Geometric data format.
+        """
+        if self.verbose:
+            iterator = tqdm(enumerate(X), desc="Converting molecules to graphs", total=len(X))
+        else:
+            iterator = enumerate(X)
+
+        pyg_graph_list = []
+        for idx, smiles_or_mol in iterator:
+            graph = graph_from_smiles(smiles_or_mol, None)
+            g = Data()
+            g.num_nodes = graph["num_nodes"]
+            g.edge_index = torch.from_numpy(graph["edge_index"])
+
+            del graph["num_nodes"]
+            del graph["edge_index"]
+
+            if graph["edge_feat"] is not None:
+                g.edge_attr = torch.from_numpy(graph["edge_feat"])
+                del graph["edge_feat"]
+
+            if graph["node_feat"] is not None:
+                g.x = torch.from_numpy(graph["node_feat"])
+                del graph["node_feat"]
+
+            pyg_graph_list.append(g)
+
+        return pyg_graph_list
+    
+    def _setup_optimizers(self) -> Tuple[torch.optim.Optimizer, Optional[Any]]:
+        """Setup optimization components including optimizer and learning rate scheduler.
+        """
+
+        optimizer = torch.optim.Adam(self.model.parameters(), lr=self.learning_rate, weight_decay=self.weight_decay)
+        scheduler = None
+        if self.use_lr_scheduler:
+            scheduler = torch.optim.lr_scheduler.ReduceLROnPlateau(
+                optimizer,
+                mode="min",
+                factor=self.scheduler_factor,
+                patience=self.scheduler_patience,
+                min_lr=1e-6,
+                cooldown=0,
+                eps=1e-8,
+            )
+
+        return optimizer, scheduler
+    
+    def fit(
+        self,
+        X_train: List[str],
+    ) -> "AttrMaskMolecularEncoder":
+        """Fit the model to the training data with optional validation set.
+
+        Parameters
+        ----------
+        X_train : List[str]
+            Training set input molecular structures as SMILES strings
+        Returns
+        -------
+        self : AttrMaskMolecularEncoder
+            Fitted estimator
+        """
+        
+        self._initialize_model(self.model_class)
+        self.model.initialize_parameters()
+        optimizer, scheduler = self._setup_optimizers()
+        
+        # Prepare datasets and loaders
+        X_train, _ = self._validate_inputs(X_train, return_rdkit_mol=True)
+        train_dataset = self._convert_to_pytorch_data(X_train)
+        train_loader = DataLoader(
+            train_dataset,
+            batch_size=self.batch_size,
+            shuffle=True,
+            num_workers=0
+        )
+        self.fitting_loss = []
+
+        for epoch in range(self.epochs):
+            # Training phase
+            train_losses = self._train_epoch(train_loader, optimizer, epoch)
+            self.fitting_loss.append(np.mean(train_losses))
+            if scheduler:
+                scheduler.step(np.mean(train_losses))
+
+        self.fitting_epoch = epoch
+        self.is_fitted_ = True
+        return self
+
+    def _train_epoch(self, train_loader, optimizer, epoch):
+        """Training logic for one epoch.
+
+        Args:
+            train_loader: DataLoader containing training data
+            optimizer: Optimizer instance for model parameter updates
+
+        Returns:
+            list: List of loss values for each training step
+        """
+        self.model.train()
+        losses = []
+
+        iterator = (
+            tqdm(train_loader, desc="Training", leave=False)
+            if self.verbose
+            else train_loader
+        )
+
+        for batch in iterator:
+            batch = batch.to(self.device)
+            optimizer.zero_grad()
+            loss = self.model.compute_loss(batch)
+            loss.backward()
+            if self.grad_clip_value is not None:
+                torch.nn.utils.clip_grad_norm_(self.model.parameters(), self.grad_clip_value)
+            optimizer.step()
+            losses.append(loss.item())
+
+            if self.verbose:
+                iterator.set_postfix({"Epoch": f"{epoch}", "Loss": f"{loss.item():.4f}"})
+
+        return losses
+
+    def encode(self, X: List[str], return_type: Literal["np", "pt"] = "pt") -> Union[np.ndarray, torch.Tensor]:
+        """Encode molecules into vector representations.
+
+        Parameters
+        ----------
+        X : List[str]
+            List of SMILES strings
+        return_type : Literal["np", "pt"], default="pt"
+            Return type of the representations
+
+        Returns
+        -------
+        representations : ndarray or torch.Tensor
+            Molecular representations
+        """
+        self._check_is_fitted()
+
+        # Convert to PyTorch Geometric format and create loader
+        X, _ = self._validate_inputs(X, return_rdkit_mol=True)
+        dataset = self._convert_to_pytorch_data(X)
+        loader = DataLoader(dataset, batch_size=self.batch_size, shuffle=False)
+
+        if self.model is None:
+            raise RuntimeError("Model not initialized")
+
+        # Generate encodings
+        self.model = self.model.to(self.device)
+        self.model.eval()
+        encodings = []
+        with torch.no_grad():
+            for batch in tqdm(loader, disable=not self.verbose):
+                batch = batch.to(self.device)
+                out = self.model(batch)
+                encodings.append(out["graph"].cpu())
+
+        # Concatenate and convert to requested format
+        encodings = torch.cat(encodings, dim=0)
+        return encodings if return_type == "pt" else encodings.numpy()

torch_molecule/encoder/constant.py

@@ -0,0 +1,13 @@
+GNN_ENCODER_MODELS = ["gin-virtual", "gcn-virtual", "gin", "gcn"]
+GNN_ENCODER_READOUTS = ["sum", "mean", "max"]
+
+GNN_ENCODER_PARAMS = [
+    # Model Hyperparameters
+    "encoder_type", "readout", "num_layer", "hidden_size", "drop_ratio", "norm_layer",
+    # Training Parameters  
+    "batch_size", "epochs", "learning_rate", "weight_decay", "grad_clip_value",
+    # Scheduler Parameters
+    "use_lr_scheduler", "scheduler_factor", "scheduler_patience",
+    # Other Parameters
+    "fitting_epoch", "fitting_loss", "device", "verbose", "model_name"
+]

torch_molecule/encoder/contextpred/__init__.py

@@ -0,0 +1,3 @@
+from .modeling_contextpred import ContextPredMolecularEncoder
+
+__all__ = ['ContextPredMolecularEncoder']

torch_molecule/encoder/contextpred/model.py

@@ -0,0 +1,207 @@
+import torch
+import torch.nn as nn
+from torch_geometric.nn import global_add_pool, global_mean_pool, global_max_pool
+from torch_geometric.data import Data
+
+from ...nn import GNN_node, GNN_node_Virtualnode
+from ...utils import init_weights
+
+from .utils import ExtractSubstructureContextPair
+
+class_criterion = torch.nn.CrossEntropyLoss()
+
+def cycle_index(num, shift):
+    arr = torch.arange(num) + shift
+    arr[-shift:] = torch.arange(shift)
+    return arr
+
+class GNN(nn.Module):
+    def __init__(
+        self,
+        num_layer,
+        hidden_size,
+        drop_ratio=0.5,
+        norm_layer="batch_norm",
+        encoder_type="gin-virtual",
+        readout="max",
+        mode="cbow",
+        context_size=3,
+        neg_samples=1
+    ):
+        super(GNN, self).__init__()
+        gnn_name = encoder_type.split("-")[0]
+        self.num_layer = num_layer
+        self.hidden_size = hidden_size
+        self.mode = mode
+        self.neg_samples = neg_samples
+        self.context_size = context_size
+
+        encoder_params_substruct = {
+            "num_layer": num_layer,
+            "hidden_size": hidden_size,
+            "JK": "last", 
+            "drop_ratio": drop_ratio,
+            "residual": True,
+            "gnn_name": gnn_name,
+            "norm_layer": norm_layer
+        }
+        
+        encoder_params_context = {
+            "num_layer": context_size,
+            "hidden_size": hidden_size,
+            "JK": "last", 
+            "drop_ratio": drop_ratio,
+            "residual": True,
+            "gnn_name": gnn_name,
+            "norm_layer": norm_layer
+        }
+        
+        # Choose encoder type based on encoder_model
+        encoder_class = GNN_node_Virtualnode if "virtual" in encoder_type else GNN_node
+        self.graph_encoder_substuct = encoder_class(**encoder_params_substruct)
+        pooling_funcs = {
+            "sum": global_add_pool,
+            "mean": global_mean_pool,
+            "max": global_max_pool
+        }
+        self.pool = pooling_funcs.get(readout)
+        if self.pool is None:
+            raise ValueError(f"Invalid graph pooling type {readout}.")
+
+        self.graph_encoder_context = encoder_class(**encoder_params_context)
+    
+    def initialize_parameters(self, seed=None):
+        """
+        Randomly initialize all model parameters using the init_weights function.
+        
+        Args:
+            seed (int, optional): Random seed for reproducibility. Defaults to None.
+        """
+        if seed is not None:
+            torch.manual_seed(seed)
+        
+        # Initialize the main components
+        init_weights(self.graph_encoder_substuct)
+        init_weights(self.graph_encoder_context)
+        
+        # Reset all parameters using PyTorch Geometric's reset function
+        def reset_parameters(module):
+            if hasattr(module, 'reset_parameters'):
+                module.reset_parameters()
+            elif hasattr(module, 'weight') and hasattr(module.weight, 'data'):
+                init_weights(module)
+        
+        self.apply(reset_parameters)
+
+    def compute_loss(self, batched_data):
+        device = batched_data.x.device
+        extract_context = ExtractSubstructureContextPair(self.num_layer, self.num_layer - 1, self.num_layer + self.context_size - 1)
+        keys = ["center_substruct_idx", "edge_attr_substruct", "edge_index_substruct", "x_substruct", "overlap_context_substruct_idx", "edge_attr_context", "edge_index_context", "x_context"]
+        for key in keys:
+            batched_data[key] = []
+
+        #used for pooling the substructure
+        batched_data.batch_substruct = [] 
+        #used for pooling the context
+        batched_data.batch_context = [] 
+        batched_data.batch_overlapped_context = []
+        batched_data.overlapped_context_size = []
+        
+        cumsum_main = 0
+        cumsum_substruct = 0
+        cumsum_context = 0
+
+        i = 0
+        
+        for j in range(len(batched_data)):
+            data = batched_data[j]
+            data = extract_context(data)
+
+            if hasattr(data, "x_context"):
+                num_nodes = data.num_nodes
+                num_nodes_substruct = len(data.x_substruct)
+                num_nodes_context = len(data.x_context)
+
+                batched_data.batch_substruct.append(torch.full((len(data.substruct_node_idxes), ), i, dtype=torch.long).to(device))
+                batched_data.batch_context.append(torch.full((len(data.context_node_idxes), ), i, dtype=torch.long).to(device))
+                batched_data.batch_overlapped_context.append(torch.full((len(data.overlap_context_substruct_idx), ), i, dtype=torch.long).to(device))
+                batched_data.overlapped_context_size.append(len(data.overlap_context_substruct_idx))
+
+                ###batching for the substructure graph
+                for key in ["center_substruct_idx", "edge_attr_substruct", "edge_index_substruct", "x_substruct"]:
+                    item = data[key]
+                    item = item + cumsum_substruct if key in ["edge_index", "edge_index_substruct", "edge_index_context", "overlap_context_substruct_idx", "center_substruct_idx"] else item
+                    batched_data[key].append(item.to(device))
+                
+
+                ###batching for the context graph
+                for key in ["overlap_context_substruct_idx", "edge_attr_context", "edge_index_context", "x_context"]:
+                    item = data[key]
+                    item = item + cumsum_context if key in ["edge_index", "edge_index_substruct", "edge_index_context", "overlap_context_substruct_idx", "center_substruct_idx"] else item
+                    batched_data[key].append(item.to(device))
+
+                cumsum_main += num_nodes
+                cumsum_substruct += num_nodes_substruct   
+                cumsum_context += num_nodes_context
+                i += 1
+
+        for key in keys:
+            if key in ["edge_index", "edge_index_substruct", "edge_index_context"]:
+                dim = -1
+            else:
+                dim = 0
+            batched_data[key] = torch.cat(batched_data[key], dim=dim)
+
+        batched_data.batch_substruct = torch.cat(batched_data.batch_substruct, dim=-1).to(device)
+        batched_data.batch_context = torch.cat(batched_data.batch_context, dim=-1).to(device)
+        batched_data.batch_overlapped_context = torch.cat(batched_data.batch_overlapped_context, dim=-1).to(device)
+        batched_data.overlapped_context_size = torch.LongTensor(batched_data.overlapped_context_size).to(device)
+
+        # generate predictions
+        substruct_data = Data(x=batched_data.x_substruct, edge_index=batched_data.edge_index_substruct, edge_attr=batched_data.edge_attr_substruct, batch=batched_data.batch_substruct)
+        context_data = Data(x=batched_data.x_context, edge_index=batched_data.edge_index_context, edge_attr=batched_data.edge_attr_context, batch=batched_data.batch_context)
+        
+        substruct_node_rep, _ = self.graph_encoder_substuct(substruct_data)
+        #substruct_h_rep = self.pool(substruct_node_rep, batched_data.batch_substruct)
+        
+        context_node_rep, _ = self.graph_encoder_context(context_data)
+        context_h_rep = self.pool(context_node_rep, batched_data.batch_context)
+        
+        # contexts are represented by 
+        if self.mode == "cbow":
+            # positive context representation
+            context_rep = self.pool(context_node_rep[batched_data.overlap_context_substruct_idx], batched_data.batch_overlapped_context)
+            # negative contexts are obtained by shifting the indicies of context embeddings
+            neg_context_rep = torch.cat([context_rep[cycle_index(len(context_rep), i+1)] for i in range(self.neg_samples)], dim = 0)
+            
+            pred_pos = torch.sum(substruct_node_rep[batched_data.center_substruct_idx] * context_rep, dim = 1)
+            pred_neg = torch.sum(substruct_node_rep[batched_data.center_substruct_idx].repeat((self.neg_samples, 1))*neg_context_rep, dim = 1)
+
+        elif self.mode == "skipgram":
+
+            expanded_substruct_rep = torch.cat([substruct_node_rep[batched_data.center_substruct_idx][i].repeat((batched_data.overlapped_context_size[i],1)) for i in range(len(substruct_node_rep[batched_data.center_substruct_idx]))], dim = 0)
+            pred_pos = torch.sum(expanded_substruct_rep * context_h_rep, dim = 1)
+
+            #shift indices of substructures to create negative examples
+            shifted_expanded_substruct_rep = []
+            for i in range(self.neg_samples):
+                shifted_substruct_rep = substruct_node_rep[batched_data.center_substruct_idx][cycle_index(len(substruct_node_rep[batched_data.center_substruct_idx]), i+1)]
+                shifted_expanded_substruct_rep.append(torch.cat([shifted_substruct_rep[i].repeat((batched_data.overlapped_context_size[i],1)) for i in range(len(shifted_substruct_rep))], dim = 0))
+
+            shifted_expanded_substruct_rep = torch.cat(shifted_expanded_substruct_rep, dim = 0)
+            pred_neg = torch.sum(shifted_expanded_substruct_rep * context_h_rep.repeat((self.neg_samples, 1)), dim = 1)
+
+        else:
+            raise ValueError("Invalid mode!")
+        
+        loss_pos = class_criterion(pred_pos.double(), torch.ones(len(pred_pos)).to(pred_pos.device).double())
+        loss_neg = class_criterion(pred_neg.double(), torch.zeros(len(pred_neg)).to(pred_neg.device).double())
+        
+        loss_class = loss_pos + self.neg_samples*loss_neg
+
+        return loss_class
+
+    def forward(self, batched_data):
+        h_node, _ = self.graph_encoder_substuct(batched_data)
+        h_rep = self.pool(h_node, batched_data.batch)
+        return {"graph": h_rep, "node": h_node}