Update verified WDS index (2 shards)

README.md

@@ -56,7 +56,10 @@ annotations. Metadata never requires remote pickle execution.
 training loader, preserving camera/depth corrections. Use it alongside the
 SpatialEncoder checkout and its dependencies (WebDataset, torch, OpenCV,
 Pillow, numpy, h5py, protobuf, etc.). Linux is required for its memory-backed
-seekable asset handles. No original source directories are needed.
+seekable asset handles. HDF5 additionally needs one sequence-sized temporary
+file per active worker, deleted after decoding; set `SPATIAL_WDS_TMPDIR` to a
+local SSD or sufficiently large tmpfs. No original source directories or
+whole-shard downloads are needed.
 
 ```python
 import webdataset as wds

manifest.json

@@ -1 +1 @@
-{"schema": "spatialencoder-wds-native-v1", "release": "wds_v20260913", "repo": "xxxspatialencoderwds2/data_2", "status": "uploading", "shards": [{"schema": "spatialencoder-wds-native-v1", "dataset": "HSSD", "split": "train", "shard_index": 1, "file": "hssd-train-000001.tar", "bytes": 1044736000, "sha256": "0aed8da5b8f4cb14c140e92f1ee7ed9e41db1569768eb23ae1603802a3cac4a2", "samples": 34, "frames": 204, "repo": "xxxspatialencoderwds2/data_2", "path": "wds_v20260913/HSSD/hssd-train-000001.tar", "commit": "6f5ad487ae9c6c36e5e56d37aff246af2f422b86", "verified_at": 1789348674.290689, "url": "https://huggingface.co/datasets/xxxspatialencoderwds2/data_2/resolve/6f5ad487ae9c6c36e5e56d37aff246af2f422b86/wds_v20260913/HSSD/hssd-train-000001.tar"}], "updated_at": 1789348674.358436}
+{"schema": "spatialencoder-wds-native-v1", "release": "wds_v20260913", "repo": "xxxspatialencoderwds2/data_2", "status": "uploading", "shards": [{"schema": "spatialencoder-wds-native-v1", "dataset": "CA-1M", "split": "train", "shard_index": 0, "file": "ca_1m-train-000000.tar", "bytes": 1042227200, "sha256": "a1a71ab72304c690af08c35d8823a9337630fc8be2a8e442d2fcc24181e66cc7", "samples": 17, "frames": 1274, "repo": "xxxspatialencoderwds2/data_2", "path": "wds_v20260913/CA-1M/ca_1m-train-000000.tar", "commit": "a8d90972c890d03c6a5b0874686452fc9b2263e2", "verified_at": 1789348901.6397967, "url": "https://huggingface.co/datasets/xxxspatialencoderwds2/data_2/resolve/a8d90972c890d03c6a5b0874686452fc9b2263e2/wds_v20260913/CA-1M/ca_1m-train-000000.tar"}, {"schema": "spatialencoder-wds-native-v1", "dataset": "HSSD", "split": "train", "shard_index": 1, "file": "hssd-train-000001.tar", "bytes": 1044736000, "sha256": "0aed8da5b8f4cb14c140e92f1ee7ed9e41db1569768eb23ae1603802a3cac4a2", "samples": 34, "frames": 204, "repo": "xxxspatialencoderwds2/data_2", "path": "wds_v20260913/HSSD/hssd-train-000001.tar", "commit": "6f5ad487ae9c6c36e5e56d37aff246af2f422b86", "verified_at": 1789348674.290689, "url": "https://huggingface.co/datasets/xxxspatialencoderwds2/data_2/resolve/6f5ad487ae9c6c36e5e56d37aff246af2f422b86/wds_v20260913/HSSD/hssd-train-000001.tar"}], "updated_at": 1789348901.7521486}

reader.py

@@ -1,8 +1,8 @@
 """S3/HTTPS WebDataset consumer using SpatialEncoder's native geometry code.
 
-Shards stream through WebDataset; only the current sample lives in RAM.
-Linux memfd handles allow existing PIL/OpenCV/HDF5 readers to seek inside a
-sample, without fetching a complete shard or accessing original dataset paths.
+Shards stream through WebDataset; only the current sample is materialized.
+Linux memfd handles allow existing PIL/OpenCV readers to seek inside a sample.
+HDF5 requires a real path and uses a bounded per-sample temporary file instead.
 Install this module alongside the matching SpatialEncoder checkout.
 """
 from __future__ import annotations
@@ -37,9 +37,9 @@ def header_of(sample):
 
 @contextmanager
 def memory_assets(sample):
-    """Suffix-preserving local handles; assets themselves never touch disk."""
+    """Suffix-preserving handles; HDF5 alone needs a real temporary file."""
     descriptors = []
-    with tempfile.TemporaryDirectory(prefix="spatial-wds-") as directory:
+    with tempfile.TemporaryDirectory(prefix="spatial-wds-", dir=os.environ.get("SPATIAL_WDS_TMPDIR")) as directory:
         paths = {}
         try:
             for field, data in sample.items():
@@ -47,12 +47,19 @@ def memory_assets(sample):
                     continue
                 if not re.fullmatch(r"[a-zA-Z0-9_]+\.[a-zA-Z0-9]+", field):
                     raise ValueError(f"Unsafe asset field: {field}")
+                path = Path(directory) / field
+                if field.endswith(".h5"):
+                    # HDF5's sec2 driver resolves realpath and rejects Linux
+                    # anonymous memfd objects. Cache only this sequence, not
+                    # the shard; the context always removes it after decoding.
+                    path.write_bytes(data)
+                    paths[field] = str(path)
+                    continue
                 descriptor = os.memfd_create("spatial-wds", flags=os.MFD_CLOEXEC)
                 descriptors.append(descriptor)
                 view = memoryview(data)
                 while view:
                     view = view[os.write(descriptor, view):]
-                path = Path(directory) / field
                 path.symlink_to(f"/proc/{os.getpid()}/fd/{descriptor}")
                 paths[field] = str(path)
             yield Path(directory), paths

urls/train.txt

@@ -1 +1,2 @@
+https://huggingface.co/datasets/xxxspatialencoderwds2/data_2/resolve/a8d90972c890d03c6a5b0874686452fc9b2263e2/wds_v20260913/CA-1M/ca_1m-train-000000.tar
 https://huggingface.co/datasets/xxxspatialencoderwds2/data_2/resolve/6f5ad487ae9c6c36e5e56d37aff246af2f422b86/wds_v20260913/HSSD/hssd-train-000001.tar