opensportslib 0.1.1.dev1__tar.gz → 0.1.1.dev2__tar.gz

{opensportslib-0.1.1.dev1/opensportslib.egg-info → opensportslib-0.1.1.dev2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: opensportslib
-Version: 0.1.1.dev1
+Version: 0.1.1.dev2
 Summary: OpenSportsLib is the professional library, designed for advanced video understanding in sports. It provides state-of-the-art tools for action recognition, spotting, retrieval, and captioning, making it ideal for researchers, analysts, and developers working with sports video data.
 Author: Jeet Vora
 Requires-Python: >=3.12

{opensportslib-0.1.1.dev1 → opensportslib-0.1.1.dev2}/opensportslib/core/utils/data.py

@@ -52,7 +52,16 @@ def tracking_collate_fn(batch):
     Custom collate function for tracking data.
     Uses PyG Batch.from_data_list for efficient C++ batching.
     """
-    from torch_geometric.data import Batch
+    try:
+        from torch_geometric.data import Batch
+    except ImportError as exc:
+        raise ImportError(
+            "torch-geometric is required for tracking_collate_fn. "
+            "Install with: pip install \"opensportslib[py-geometric]\" "
+            "-f https://pytorch-geometric.com/whl/torch-2.10.0+cu128.html "
+            "or (editable): pip install -e \".[py-geometric]\" "
+            "-f https://pytorch-geometric.com/whl/torch-2.10.0+cu128.html"
+        ) from exc
     
     batch_size = len(batch)
     seq_len = batch[0]['seq_len']
@@ -82,4 +91,4 @@ def mixup_data(x, y, alpha=0.2):
     index = torch.randperm(x.size(0)).to(x.device)
     mixed_x = lam * x + (1 - lam) * x[index]
     return mixed_x, y, y[index], lam
-    
+    

{opensportslib-0.1.1.dev1 → opensportslib-0.1.1.dev2}/opensportslib/datasets/classification_dataset.py

@@ -496,7 +496,16 @@ class TrackingDataset(ClassificationDataset):
         a copy of the feature array is made before augmentation and 
         normalization so the cached data is never mutated.
         """
-        from torch_geometric.data import Data
+        try:
+            from torch_geometric.data import Data
+        except ImportError as exc:
+            raise ImportError(
+                "torch-geometric is required for tracking_parquet datasets. "
+                "Install with: pip install \"opensportslib[py-geometric]\" "
+                "-f https://pytorch-geometric.com/whl/torch-2.10.0+cu128.html "
+                "or (editable): pip install -e \".[py-geometric]\" "
+                "-f https://pytorch-geometric.com/whl/torch-2.10.0+cu128.html"
+            ) from exc
 
         from opensportslib.datasets.utils.tracking import normalize_features
         
@@ -533,7 +542,16 @@ class TrackingDataset(ClassificationDataset):
 
     def _getitem_on_the_fly(self, idx):
         """load, parse, and process a single sample from disk."""
-        from torch_geometric.data import Data
+        try:
+            from torch_geometric.data import Data
+        except ImportError as exc:
+            raise ImportError(
+                "torch-geometric is required for tracking_parquet datasets. "
+                "Install with: pip install \"opensportslib[py-geometric]\" "
+                "-f https://pytorch-geometric.com/whl/torch-2.10.0+cu128.html "
+                "or (editable): pip install -e \".[py-geometric]\" "
+                "-f https://pytorch-geometric.com/whl/torch-2.10.0+cu128.html"
+            ) from exc
         
         from opensportslib.datasets.utils.tracking import (
             compute_deltas,
@@ -600,4 +618,4 @@ class TrackingDataset(ClassificationDataset):
         if label is not None:
             out["label"] = label
         return out
-        
+        

{opensportslib-0.1.1.dev1 → opensportslib-0.1.1.dev2}/opensportslib/models/backbones/builder.py

@@ -38,6 +38,23 @@ import os
 from opensportslib.models.utils.shift import make_temporal_shift
 
 
+_PYG_IMPORT_ERROR_MSG = (
+    "torch-geometric is required for graph/tracking models. "
+    "Install with: pip install \"opensportslib[py-geometric]\" "
+    "-f https://pytorch-geometric.com/whl/torch-2.10.0+cu128.html "
+    "or (editable): pip install -e \".[py-geometric]\" "
+    "-f https://pytorch-geometric.com/whl/torch-2.10.0+cu128.html"
+)
+
+
+def _import_pyg_nn():
+    try:
+        import torch_geometric.nn as pyg_nn
+    except ImportError as exc:
+        raise ImportError(_PYG_IMPORT_ERROR_MSG) from exc
+    return pyg_nn
+
+
 def _use_pretrained_weights(default=True):
     """Control pretrained backbone loading from env for offline/local runs."""
     flag = os.environ.get("OSL_PRETRAINED_WEIGHTS", "1").strip().lower()
@@ -350,7 +367,8 @@ class GraphEncoder(nn.Module):
     def __init__(self, input_dim, hidden_dim, num_layers, conv_type='gin', dropout=0.1):
         super().__init__()
 
-        from torch_geometric.nn import DeepGCNLayer
+        pyg_nn = _import_pyg_nn()
+        DeepGCNLayer = pyg_nn.DeepGCNLayer
         self.conv_type = conv_type
         self.feat_dim = hidden_dim
 
@@ -371,10 +389,14 @@ class GraphEncoder(nn.Module):
             self.gcn_layers.append(layer)
         
     def _build_conv_layer(self, conv_type, hidden_dim, dropout):
-        from torch_geometric.nn import (
-            GATv2Conv, EdgeConv, SAGEConv, 
-            GINConv, GENConv, GraphConv, MultiAggregation
-        )
+        pyg_nn = _import_pyg_nn()
+        GATv2Conv = pyg_nn.GATv2Conv
+        EdgeConv = pyg_nn.EdgeConv
+        SAGEConv = pyg_nn.SAGEConv
+        GINConv = pyg_nn.GINConv
+        GENConv = pyg_nn.GENConv
+        GraphConv = pyg_nn.GraphConv
+        MultiAggregation = pyg_nn.MultiAggregation
 
         if conv_type == 'graphconv':
             return GraphConv(
@@ -434,7 +456,8 @@ class GraphEncoder(nn.Module):
             raise ValueError(f"Unknown conv type: {conv_type}")
             
     def forward(self, x, edge_index, batch):
-        from torch_geometric.nn import global_mean_pool
+        pyg_nn = _import_pyg_nn()
+        global_mean_pool = pyg_nn.global_mean_pool
         
         x = self.node_encoder(x)
 
@@ -448,7 +471,8 @@ class GraphEncoder(nn.Module):
 
     def _compute_edge_dynamic(self, x, batch, k=8):
         """Compute dynamic KNN edges for EdgeConv layers."""
-        from torch_geometric.nn import knn_graph
+        pyg_nn = _import_pyg_nn()
+        knn_graph = pyg_nn.knn_graph
         edge_index = knn_graph(x, k=k, batch=batch, loop=False)
         return edge_index
 

{opensportslib-0.1.1.dev1 → opensportslib-0.1.1.dev2}/opensportslib/models/base/tracking.py

@@ -1,6 +1,5 @@
 import torch
 import torch.nn as nn
-from torch_geometric.nn import global_mean_pool
 
 from opensportslib.models.backbones.builder import build_backbone
 from opensportslib.models.neck.builder import build_neck
@@ -70,4 +69,4 @@ class TrackingModel(nn.Module):
         # classification
         logits = self.head(x)  # (B, num_classes)
         
-        return logits
+        return logits

{opensportslib-0.1.1.dev1 → opensportslib-0.1.1.dev2/opensportslib.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: opensportslib
-Version: 0.1.1.dev1
+Version: 0.1.1.dev2
 Summary: OpenSportsLib is the professional library, designed for advanced video understanding in sports. It provides state-of-the-art tools for action recognition, spotting, retrieval, and captioning, making it ideal for researchers, analysts, and developers working with sports video data.
 Author: Jeet Vora
 Requires-Python: >=3.12

{opensportslib-0.1.1.dev1 → opensportslib-0.1.1.dev2}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "opensportslib"
-version = "0.1.1.dev1"
+version = "0.1.1.dev2"
 description = "OpenSportsLib is the professional library, designed for advanced video understanding in sports. It provides state-of-the-art tools for action recognition, spotting, retrieval, and captioning, making it ideal for researchers, analysts, and developers working with sports video data."
 readme = "README.md"
 requires-python = ">=3.12"