# import os # import pandas as pd # import numpy as np # from sklearn.model_selection import train_test_split # from sklearn.feature_extraction.text import TfidfVectorizer # from sklearn.preprocessing import LabelEncoder # # from torch.utils.data import DataLoader, TensorDataset # from torch.utils.data import DataLoader, TensorDataset, WeightedRandomSampler # import torch # # Other imports remain the same # def create_weighted_sampler(labels): # class_counts = np.bincount(labels) # class_weights = 1. / class_counts # sample_weights = class_weights[labels] # return sample_weights # Return sample weights, not the sampler # # Function to read Finsen data # def read_finsen_data(file_path): # df = pd.read_csv(file_path) # texts = df['Content'].tolist() # labels = LabelEncoder().fit_transform(df['Category']) # return texts, labels # # Function to create train and validation DataLoaders # def get_train_valid_loader(batch_size=128, valid_size=0.1, get_val_temp=0, random_seed=42, shuffle=True, num_workers=4, pin_memory=False): # # texts, labels = read_finsen_data('./data/FinSen/finsen.csv') # texts, labels = read_finsen_data('./adafocal-main/data/FinSen/finsen.csv') # # Vectorize and split the dataset # vectorizer = TfidfVectorizer(stop_words='english', max_features=2000) # features = vectorizer.fit_transform(texts).toarray() # X_train, X_val, y_train, y_val = train_test_split(features, labels, test_size=valid_size, random_state=random_seed) # # Create weighted sampler for the training set to handle class imbalance # train_weights = create_weighted_sampler(y_train) # # train_sampler = WeightedRandomSampler(train_weights, len(train_weights)) # train_sampler = WeightedRandomSampler(weights=train_weights, num_samples=len(train_weights), replacement=True) # # Convert to tensors and create DataLoaders # train_dataset = TensorDataset(torch.tensor(X_train, dtype=torch.float32), torch.tensor(y_train, dtype=torch.long)) # val_dataset = TensorDataset(torch.tensor(X_val, dtype=torch.float32), torch.tensor(y_val, dtype=torch.long)) # # Use sampler for train_loader # train_loader = DataLoader(train_dataset, batch_size=batch_size, sampler=train_sampler, num_workers=num_workers, pin_memory=pin_memory) # # Validation loader doesn't need weighted sampling # val_loader = DataLoader(val_dataset, batch_size=batch_size, shuffle=shuffle, num_workers=num_workers, pin_memory=pin_memory) # return train_loader, val_loader # # # Convert to tensors and create DataLoaders # # train_dataset = TensorDataset(torch.tensor(X_train, dtype=torch.float32), torch.tensor(y_train, dtype=torch.long)) # # val_dataset = TensorDataset(torch.tensor(X_val, dtype=torch.float32), torch.tensor(y_val, dtype=torch.long)) # # train_loader = DataLoader(train_dataset, batch_size=batch_size, shuffle=shuffle, num_workers=num_workers, pin_memory=pin_memory) # # val_loader = DataLoader(val_dataset, batch_size=batch_size, shuffle=shuffle, num_workers=num_workers, pin_memory=pin_memory) # # return train_loader, val_loader # # Function to create a test DataLoader # def get_test_loader(batch_size=128, shuffle=True, num_workers=4, pin_memory=False): # # texts, labels = read_finsen_data('./data/FinSen/finsen.csv') # Ensure this path is correct # texts, labels = read_finsen_data('./adafocal-main/data/FinSen/finsen.csv') # # Vectorize and split the dataset for test data # vectorizer = TfidfVectorizer(stop_words='english', max_features=2000) # features = vectorizer.fit_transform(texts).toarray() # _, X_test, _, y_test = train_test_split(features, labels, test_size=0.2, random_state=42, shuffle=True) # # Convert to tensors and create DataLoader # test_dataset = TensorDataset(torch.tensor(X_test, dtype=torch.float32), torch.tensor(y_test, dtype=torch.long)) # test_loader = DataLoader(test_dataset, batch_size=batch_size, shuffle=shuffle, num_workers=num_workers, pin_memory=pin_memory) # return test_loader import os import pandas as pd import numpy as np from sklearn.model_selection import train_test_split from sklearn.feature_extraction.text import TfidfVectorizer from sklearn.preprocessing import LabelEncoder from torch.utils.data import DataLoader, TensorDataset, WeightedRandomSampler import torch def create_weighted_sampler(labels): class_counts = np.bincount(labels) class_weights = 1. / class_counts sample_weights = class_weights[labels] weighted_sampler = WeightedRandomSampler(weights=sample_weights, num_samples=len(sample_weights), replacement=True) return weighted_sampler def read_finsen_data(file_path): df = pd.read_csv(file_path) texts = df['Content'].tolist() labels = df['Category'].apply(lambda x: x.strip()).tolist() # Ensure labels are stripped of whitespace label_encoder = LabelEncoder() labels = label_encoder.fit_transform(labels) return texts, labels, label_encoder.classes_ def prepare_data_loaders(texts, labels, batch_size, valid_size, random_seed, shuffle, num_workers, pin_memory): vectorizer = TfidfVectorizer(stop_words='english', max_features=2000) features = vectorizer.fit_transform(texts).toarray() X_train, X_val_test, y_train, y_val_test = train_test_split(features, labels, test_size=valid_size + 0.1, random_state=random_seed) # Adjust valid_size + test_size X_val, X_test, y_val, y_test = train_test_split(X_val_test, y_val_test, test_size=0.5, random_state=random_seed) # Split remaining data equally for val and test train_sampler = create_weighted_sampler(y_train) train_dataset = TensorDataset(torch.tensor(X_train, dtype=torch.float32), torch.tensor(y_train, dtype=torch.long)) val_dataset = TensorDataset(torch.tensor(X_val, dtype=torch.float32), torch.tensor(y_val, dtype=torch.long)) test_dataset = TensorDataset(torch.tensor(X_test, dtype=torch.float32), torch.tensor(y_test, dtype=torch.long)) train_loader = DataLoader(train_dataset, batch_size=batch_size, sampler=train_sampler, num_workers=num_workers, pin_memory=pin_memory) val_loader = DataLoader(val_dataset, batch_size=batch_size, shuffle=shuffle, num_workers=num_workers, pin_memory=pin_memory) test_loader = DataLoader(test_dataset, batch_size=batch_size, shuffle=shuffle, num_workers=num_workers, pin_memory=pin_memory) return train_loader, val_loader, test_loader def get_data_loaders(finsen_data_path, batch_size=128, valid_size=0.2, random_seed=42, shuffle=True, num_workers=4, pin_memory=False): texts, labels, _ = read_finsen_data(finsen_data_path) return prepare_data_loaders(texts, labels, batch_size, valid_size, random_seed, shuffle, num_workers, pin_memory) # Now, train_loader, val_loader, and test_loader can be used in training and evaluation.