Upload 4 files

examples/01_create_featureless_anndata.py

@@ -1,20 +1,24 @@
 #!/usr/bin/env python3
 """
-Create a featureless AnnData (.h5ad) from the distributed DRMHB VibFrame archive.
+Create a featureless DRMHB AnnData from the distributed VibFrame archive.
 
-This example uses the out-of-core API from vibframe-anndata and is intended for
-the full DRMHB archive, which is too large to treat as a small in-memory example.
+Validated with:
+    vibframe-anndata==0.2.2
+
+DRMHB is distributed canonically as VibFrame. This script demonstrates the
+recommended out-of-core conversion into a reusable featureless H5AD.
+
+The 0.2.2 release can also ingest DRMHB snapshot-level evaluation truth from
+evaluation/snapshot_truth/ and store it in obsm["ground_truth"] without exposing
+those labels as model features.
 
 Install:
-    python -m pip install "vibframe-anndata==0.2.0"
+    python -m pip install "vibframe-anndata==0.2.2"
 
 Usage:
     python examples/01_create_featureless_anndata.py \
         DRMHB.vibframe.zip \
         DRMHB.float32.h5ad
-
-The resulting AnnData contains raw spectra/waveforms and metadata, but no
-engineered features in X.
 """
 
 from __future__ import annotations
@@ -28,20 +32,22 @@ from vibframe_anndata import (
     validate_streamed_h5ad,
 )
 
+EXPECTED_VERSION = "0.2.2"
+
 
 def parse_args() -> argparse.Namespace:
     parser = argparse.ArgumentParser(
-        description="Convert a DRMHB VibFrame archive into a featureless AnnData."
+        description="Convert DRMHB VibFrame into a featureless AnnData."
     )
     parser.add_argument(
         "vibframe",
         type=Path,
-        help="Path to the DRMHB .vibframe.zip/.zip archive.",
+        help="Path to the DRMHB VibFrame ZIP archive.",
     )
     parser.add_argument(
         "output",
         type=Path,
-        help="Destination .h5ad file.",
+        help="Destination .h5ad path.",
     )
     parser.add_argument(
         "--block-size-mib",
@@ -55,15 +61,23 @@ def parse_args() -> argparse.Namespace:
         default="float32",
         help="Numerical storage dtype (default: float32).",
     )
+    parser.add_argument(
+        "--without-ground-truth",
+        action="store_true",
+        help=(
+            "Do not ingest evaluation/snapshot_truth into obsm['ground_truth']. "
+            "By default DRMHB ground truth is imported explicitly for evaluation."
+        ),
+    )
     return parser.parse_args()
 
 
 def main() -> None:
     args = parse_args()
 
-    if vibframe_anndata_version != "0.2.0":
+    if vibframe_anndata_version != EXPECTED_VERSION:
         raise RuntimeError(
-            "This example is validated for vibframe-anndata==0.2.0; "
+            f"This example is validated for vibframe-anndata=={EXPECTED_VERSION}; "
             f"active version: {vibframe_anndata_version!r}"
         )
 
@@ -75,16 +89,22 @@ def main() -> None:
     config = {
         "version": 1,
         "raw_import": {
-            # DRMHB contains both signal families. Keep the explicit settings
-            # here so that the example documents the intended conversion.
             "waveforms": True,
             "spectra": True,
             "on_missing_signal": "nan",
         },
+        "ground_truth": {
+            "enabled": not args.without_ground_truth,
+            "on_missing": "error",
+        },
         "output": {
             "dtype": args.dtype,
+            # Explicitly use the 0.2.2 default. Compression is lossless but the
+            # package benchmark shows that compressed raw payloads materially
+            # slow later feature enrichment for this DRMHB-sized workload.
+            "raw_compression": "none",
         },
-        # Empty by design: the base AnnData is a raw reusable representation.
+        # Empty by design: the base AnnData contains raw reusable signals only.
         "features": [],
     }
 
@@ -100,7 +120,11 @@ def main() -> None:
     print(f"Created: {output}")
     print(f"Observations: {validation['n_obs']:,}")
     print(f"Variables/features: {validation['n_vars']:,}")
-    print("The featureless base should report n_vars = 0.")
+    print("Expected featureless base: n_vars = 0")
+    print(
+        "Ground truth:",
+        "disabled" if args.without_ground_truth else "stored in obsm['ground_truth']",
+    )
 
 
 if __name__ == "__main__":

examples/02_inspect_anndata.py

@@ -1,12 +1,12 @@
 #!/usr/bin/env python3
 """
-Inspect a DRMHB AnnData produced locally from the distributed VibFrame archive.
+Inspect a DRMHB AnnData generated with vibframe-anndata 0.2.2.
 
-The file is opened in backed read-only mode so the raw multi-gigabyte arrays are
-not loaded into RAM.
+The H5AD is opened in backed read-only mode, so the multi-gigabyte raw signal
+payload is not loaded into RAM.
 
 Install:
-    python -m pip install "vibframe-anndata==0.2.0" "anndata>=0.10,<0.13"
+    python -m pip install "vibframe-anndata==0.2.2" "anndata>=0.10,<0.13"
 
 Usage:
     python examples/02_inspect_anndata.py DRMHB.float32.h5ad
@@ -19,12 +19,17 @@ from pathlib import Path
 
 import anndata as ad
 
-from vibframe_anndata import validate_streamed_h5ad
+from vibframe_anndata import (
+    __version__ as vibframe_anndata_version,
+    validate_streamed_h5ad,
+)
+
+EXPECTED_VERSION = "0.2.2"
 
 
 def parse_args() -> argparse.Namespace:
     parser = argparse.ArgumentParser(
-        description="Inspect a DRMHB AnnData without loading the full raw payload."
+        description="Inspect a locally generated DRMHB AnnData."
     )
     parser.add_argument("h5ad", type=Path, help="Path to a generated DRMHB .h5ad.")
     return parser.parse_args()
@@ -33,6 +38,12 @@ def parse_args() -> argparse.Namespace:
 def main() -> None:
     args = parse_args()
 
+    if vibframe_anndata_version != EXPECTED_VERSION:
+        raise RuntimeError(
+            f"This example is validated for vibframe-anndata=={EXPECTED_VERSION}; "
+            f"active version: {vibframe_anndata_version!r}"
+        )
+
     if not args.h5ad.is_file():
         raise FileNotFoundError(args.h5ad)
 
@@ -47,24 +58,37 @@ def main() -> None:
         print(adata)
         print()
         print("X shape:", adata.shape)
-        print("obs columns:")
+
+        print("\nobs columns:")
         for column in adata.obs.columns:
             print(" -", column)
 
-        print()
-        print("obsm keys:")
+        print("\nobsm keys:")
         for key in adata.obsm.keys():
             print(" -", key)
 
-        print()
-        print("uns keys:")
+        print("\nuns keys:")
         for key in adata.uns.keys():
             print(" -", key)
 
-        print()
-        print("First observations:")
+        print("\nFirst observations:")
         print(adata.obs.head())
 
+        if "ground_truth" in adata.obsm:
+            gt = adata.obsm["ground_truth"]
+            print("\nGround truth:")
+            print(" - location: obsm['ground_truth']")
+            print(" - shape:", gt.shape)
+            print(
+                " - note: evaluation-only metadata; do not include it "
+                "automatically in unsupervised model inputs"
+            )
+        else:
+            print(
+                "\nGround truth was not imported. Recreate the H5AD without "
+                "--without-ground-truth if evaluation labels are required."
+            )
+
         if adata.n_vars == 0:
             print(
                 "\nThis is a featureless base AnnData. "
@@ -72,7 +96,6 @@ def main() -> None:
                 "raw signals already stored in the H5AD."
             )
     finally:
-        # Backed AnnData keeps an HDF5 handle open.
         if getattr(adata, "file", None) is not None:
             adata.file.close()
 

examples/03_add_features.py

@@ -1,21 +1,23 @@
 #!/usr/bin/env python3
 """
-Materialize a small example feature set from a locally generated DRMHB AnnData.
+Materialize a compact example feature set from a DRMHB AnnData.
 
-This does NOT reopen the VibFrame archive. Features are derived from the raw
-waveform/spectrum representation already stored in the featureless .h5ad.
+Validated with:
+    vibframe-anndata==0.2.2
+
+This step works exclusively from raw signals already persisted in the H5AD and
+does not reopen the original VibFrame archive.
 
 Install:
-    python -m pip install "vibframe-anndata==0.2.0"
+    python -m pip install "vibframe-anndata==0.2.2"
 
 Usage:
     python examples/03_add_features.py \
         DRMHB.float32.h5ad \
         DRMHB.example-features.h5ad
 
-The example is intentionally small. See the vibframe-anndata documentation for
-the complete catalog/configuration model:
-https://413hq.github.io/vibframe-anndata-docs/latest/
+Documentation:
+    https://413hq.github.io/vibframe-anndata-docs/latest/
 """
 
 from __future__ import annotations
@@ -29,6 +31,8 @@ from vibframe_anndata import (
     validate_streamed_h5ad,
 )
 
+EXPECTED_VERSION = "0.2.2"
+
 
 def parse_args() -> argparse.Namespace:
     parser = argparse.ArgumentParser(
@@ -37,7 +41,7 @@ def parse_args() -> argparse.Namespace:
     parser.add_argument(
         "source",
         type=Path,
-        help="Featureless DRMHB .h5ad created from the VibFrame archive.",
+        help="Featureless DRMHB .h5ad generated from the VibFrame archive.",
     )
     parser.add_argument(
         "output",
@@ -56,9 +60,9 @@ def parse_args() -> argparse.Namespace:
 def main() -> None:
     args = parse_args()
 
-    if vibframe_anndata_version != "0.2.0":
+    if vibframe_anndata_version != EXPECTED_VERSION:
         raise RuntimeError(
-            "This example is validated for vibframe-anndata==0.2.0; "
+            f"This example is validated for vibframe-anndata=={EXPECTED_VERSION}; "
             f"active version: {vibframe_anndata_version!r}"
         )
 
@@ -67,9 +71,6 @@ def main() -> None:
 
     args.output.parent.mkdir(parents=True, exist_ok=True)
 
-    # A deliberately compact demonstration spanning spectral and waveform
-    # features. Catalog-backed spectral descriptors use the metric metadata
-    # persisted in the source AnnData.
     config = {
         "version": 1,
         "output": {"dtype": "float32"},
@@ -99,6 +100,10 @@ def main() -> None:
     print(f"Created: {output}")
     print(f"Observations: {validation['n_obs']:,}")
     print(f"Variables/features: {validation['n_vars']:,}")
+    print(
+        "If ground truth was present in the source base H5AD, it remains "
+        "evaluation metadata rather than an X feature."
+    )
 
 
 if __name__ == "__main__":

examples/README.md

@@ -1,17 +1,23 @@
 # DRMHB examples
 
 The Hugging Face repository distributes **DRMHB in VibFrame form only**. AnnData
-files are generated locally with the companion
+representations are generated locally with the companion
 [`vibframe-anndata`](https://413hq.github.io/vibframe-anndata-docs/latest/)
 package.
 
-Validated package version for these examples:
+These examples target and validate against:
+
+```text
+vibframe-anndata==0.2.2
+```
+
+Install it with:
 
 ```bash
-python -m pip install "vibframe-anndata==0.2.0"
+python -m pip install "vibframe-anndata==0.2.2"
 ```
 
-## 1. Create the featureless AnnData
+## 1. Create a featureless AnnData
 
 ```bash
 python examples/01_create_featureless_anndata.py \
@@ -19,8 +25,36 @@ python examples/01_create_featureless_anndata.py \
     DRMHB.float32.h5ad
 ```
 
-This is the recommended starting point for ML workflows. It ingests the raw
-VibFrame signals and metadata but leaves `X` with zero variables/features.
+This performs the recommended out-of-core conversion and leaves `X` with zero
+engineered features.
+
+With version 0.2.2, DRMHB's partitioned snapshot truth under
+`evaluation/snapshot_truth/` is also imported explicitly into:
+
+```python
+adata.obsm["ground_truth"]
+```
+
+It remains separate from `X` and `obs`, so it is available for evaluation
+without becoming an implicit model input.
+
+If a pure learning representation without evaluation truth is desired:
+
+```bash
+python examples/01_create_featureless_anndata.py \
+    DRMHB.vibframe.zip \
+    DRMHB.float32.h5ad \
+    --without-ground-truth
+```
+
+The example also explicitly keeps:
+
+```yaml
+raw_compression: none
+```
+
+which is the 0.2.2 default and the recommended setting for DRMHB when repeated
+feature materialization/read performance matters.
 
 ## 2. Inspect the generated AnnData
 
@@ -28,8 +62,8 @@ VibFrame signals and metadata but leaves `X` with zero variables/features.
 python examples/02_inspect_anndata.py DRMHB.float32.h5ad
 ```
 
-The script opens the H5AD in backed read-only mode, prints `obs`, `obsm` and
-`uns` structure, and does not load the full raw payload into RAM.
+The script opens the H5AD in backed read-only mode and reports `obs`, `obsm`,
+`uns`, feature geometry and whether `obsm["ground_truth"]` is available.
 
 ## 3. Materialize a small feature set
 
@@ -39,10 +73,13 @@ python examples/03_add_features.py \
     DRMHB.example-features.h5ad
 ```
 
-This demonstrates the second stage of the package: deriving features from raw
-signals already stored in AnnData, without reopening the original VibFrame.
+This demonstrates the second stage of the package: features are derived from
+raw signals already stored in AnnData, without reopening the original VibFrame.
+
+The example intentionally materializes only a compact set of waveform and
+spectral descriptors.
 
-For the complete API, configuration schema, supported metrics and limitations,
-see:
+For the complete API, configuration schema, ground-truth semantics, compression
+options, supported metrics and limitations, see:
 
 https://413hq.github.io/vibframe-anndata-docs/latest/