Upload ProDiff/Experiments/trajectory_a40_temporal_optimized_TKY_temporal_len3_ddpm_20250724-101534/code_snapshot/data_util.py with huggingface_hub

ProDiff/Experiments/trajectory_a40_temporal_optimized_TKY_temporal_len3_ddpm_20250724-101534/code_snapshot/data_util.py

@@ -0,0 +1,90 @@
+import torch
+import numpy as np
+import pandas as pd
+from torch.utils.data import Dataset
+from tqdm import tqdm
+import h5py
+
+class TrajectoryDataset(Dataset):
+    """Dataset for loading trajectory data from HDF5 files."""
+    def __init__(self, file_paths, traj_length):
+        self.samples = []
+        self.load_samples(file_paths, traj_length)
+    
+    def load_samples(self, file_paths, traj_length):
+        for file_path in tqdm(file_paths, desc="Loading files", unit="file"):
+            with h5py.File(file_path, 'r') as h5_file:
+                for user_id in h5_file.keys(): # Iterate over users in the HDF5 file
+                    user_group = h5_file[user_id]
+                    latitudes = user_group['latitudes'][:]
+                    longitudes = user_group['longitudes'][:]
+                    hours = user_group['hours'][:]
+                    
+                    # Create samples by sliding a window of traj_length over the user's trajectory
+                    if len(latitudes) > traj_length:
+                        for j in range(len(latitudes) - traj_length + 1):
+                            self.samples.append((hours[j:j+traj_length], latitudes[j:j+traj_length], longitudes[j:j+traj_length]))
+                    elif len(latitudes) == traj_length:
+                        self.samples.append((hours[:], latitudes[:], longitudes[:]))
+    
+    def __len__(self):
+        return len(self.samples)
+
+    def __getitem__(self, idx):
+        hours, latitudes, longitudes = self.samples[idx]
+        return torch.tensor(hours, dtype=torch.float32), torch.tensor(latitudes, dtype=torch.float32), torch.tensor(longitudes, dtype=torch.float32)
+
+
+class PatternDataset:
+    """Dataset for loading trajectory patterns, possibly for prototype learning."""
+    def __init__(self, file_paths):
+        self.trajectories = []
+        self.load_samples(file_paths)
+
+    def load_samples(self, file_paths):
+        for file_path in tqdm(file_paths, desc="Loading files", unit="file"):
+            with pd.HDFStore(file_path, 'r') as store: # Using pandas HDFStore
+                data = store['data']
+                for i in range(len(data)):
+                    abs_time_list = np.array(data['ABS_TIME'][i])
+                    lat_list = np.array(data['LAT'][i])
+                    lng_list = np.array(data['LNG'][i])
+                    trajectory = list(zip(abs_time_list, lat_list, lng_list))
+                    self.trajectories.append(trajectory)
+
+    def get_all_trajectories(self):
+        return self.trajectories
+
+    def pad_trajectories(self):
+        max_length = max(len(traj) for traj in self.trajectories)
+        padded_samples = []
+
+        for traj in self.trajectories:
+            if len(traj) < max_length:
+                # Pad shorter trajectories with their last point
+                last_point = traj[-1]
+                padding = [last_point] * (max_length - len(traj))
+                padded_traj = traj + padding
+            else:
+                padded_traj = traj
+            padded_samples.append(padded_traj)
+
+        return padded_samples
+
+
+class MinMaxScaler:
+    """Min-Max Scaler for trajectory data."""
+    def __init__(self):
+        self.min_val = None
+        self.max_val = None
+
+    def fit(self, data):
+        self.min_val = data.amin(dim=(0, 1), keepdim=True)
+        self.max_val = data.amax(dim=(0, 1), keepdim=True)
+
+    def transform(self, data):
+        return (data - self.min_val) / (self.max_val - self.min_val)
+
+    def inverse_transform(self, data):
+        return data * (self.max_val - self.min_val) + self.min_val
+