From 5a1d3ea977a11c0ddedcd6004b0a56bc9cf2074a Mon Sep 17 00:00:00 2001
From: SikongJueluo <selfconfusion@gmail.com>
Date: Thu, 21 May 2026 16:05:47 +0800
Subject: [PATCH] refactor(verification): extract shared text_labels cell and
 hash codec

- Move inline _text_labels lists into a shared text_labels cell
- Replace inline np.packbits hash encoding with bits_tensor_to_hash_bytes
- Add static tests to verify notebook API usage patterns
---
 notebooks/verification.py                  | 56 +++++++++++-----------
 tests/test_notebook_verification_static.py | 20 ++++++++
 2 files changed, 47 insertions(+), 29 deletions(-)
 create mode 100644 tests/test_notebook_verification_static.py

diff --git a/notebooks/verification.py b/notebooks/verification.py
index 0fef577..98ec317 100644
--- a/notebooks/verification.py
+++ b/notebooks/verification.py
@@ -39,6 +39,7 @@ def project_imports():
         ObjectNode,
         RoomNode,
         SimpleSceneGraph,
+        bits_tensor_to_hash_bytes,
         query_image_against_scene_graph,
     )
     from simulator import (
@@ -59,6 +60,7 @@ def project_imports():
         RoomNode,
         SimpleSceneGraph,
         TopDownSceneElements,
+        bits_tensor_to_hash_bytes,
         cfg_manager,
         collect_room_views_by_room,
         create_habitat_simulator,
@@ -70,6 +72,24 @@ def project_imports():
     )
 
 
+@app.cell
+def text_labels():
+    """Shared text labels for detection during graph build and query."""
+    text_labels = [
+        "a chair",
+        "a table",
+        "a sofa",
+        "a cabinet",
+        "a shelf",
+        "a lamp",
+        "a picture",
+        "a window",
+        "a door",
+        "a plant",
+    ]
+    return (text_labels,)
+
+
 @app.cell
 def habitat_setup(HabitatSimulatorConfig, RoomNode, create_habitat_simulator, np):
     """Initialize Habitat simulator and sample room nodes."""
@@ -170,6 +190,7 @@ def collect_views(
 def build_scene_graph(
     ObjectNode,
     SimpleSceneGraph,
+    bits_tensor_to_hash_bytes,
     cfg_manager,
     mo,
     np,
@@ -178,6 +199,7 @@ def build_scene_graph(
     room_view_dataset,
     save_object_image,
     save_room_view,
+    text_labels,
     torch,
 ):
     scene_graph = SimpleSceneGraph(
@@ -194,18 +216,6 @@ def build_scene_graph(
     _images = [item[2] for item in room_view_dataset]
     _metadata = [(item[0], item[1]) for item in room_view_dataset]
 
-    _text_labels = [
-        "a chair",
-        "a table",
-        "a sofa",
-        "a cabinet",
-        "a shelf",
-        "a lamp",
-        "a picture",
-        "a window",
-        "a door",
-        "a plant",
-    ]
     inference_batch_size = 4
     image_batches = [
         _images[index : index + inference_batch_size]
@@ -229,7 +239,7 @@ def build_scene_graph(
     ):
         _batch_output = pipeline.process_batch(
             _batch_images,
-            _text_labels,
+            text_labels,
             batch_size=inference_batch_size,
             return_debug_details=True,
         )
@@ -296,9 +306,7 @@ def build_scene_graph(
 
         _obj_id = f"{_room_id}_v{_view_idx:03d}_m{_mask_idx:02d}"
 
-        _bits_array = _hash_bits.detach().cpu().numpy().reshape(-1)
-        _bits_binary = (_bits_array > 0).astype(np.uint8)
-        _hash_bytes = np.packbits(_bits_binary).tobytes()
+        _hash_bytes = bits_tensor_to_hash_bytes(_hash_bits)
 
         object_images[_obj_id] = _cropped
         save_object_image(output_dir, _room_id, _obj_id, _view_idx, _mask_idx, _cropped)
@@ -372,6 +380,7 @@ def query_matching(
     pipeline,
     query_image_against_scene_graph,
     scene_graph,
+    text_labels,
 ):
     from io import BytesIO
 
@@ -384,23 +393,11 @@ def query_matching(
 
     _query_image = Image.open(BytesIO(_file_contents)).convert("RGB")
 
-    _text_labels = [
-        "a chair",
-        "a table",
-        "a sofa",
-        "a cabinet",
-        "a shelf",
-        "a lamp",
-        "a picture",
-        "a window",
-        "a door",
-        "a plant",
-    ]
     _query_results = query_image_against_scene_graph(
         image=_query_image,
         pipeline=pipeline,
         scene_graph=scene_graph,
-        text_labels=_text_labels,
+        text_labels=text_labels,
         top_k=5,
         batch_size=1,
     )
@@ -425,6 +422,7 @@ def query_matching(
             "query_cropped": query_cropped,
             "query_hash_hex": _best_result.query_hash.hex(),
             "top_matches": top_matches,
+            "num_query_results": len(_query_results),
         }
 
     return query_cropped, query_result, top_matches
diff --git a/tests/test_notebook_verification_static.py b/tests/test_notebook_verification_static.py
new file mode 100644
index 0000000..ddfa7f4
--- /dev/null
+++ b/tests/test_notebook_verification_static.py
@@ -0,0 +1,20 @@
+from __future__ import annotations
+
+from pathlib import Path
+
+
+NOTEBOOK = Path(__file__).resolve().parents[1] / "notebooks" / "verification.py"
+
+
+def test_verification_notebook_uses_scenegraph_query_api():
+    source = NOTEBOOK.read_text()
+
+    assert "query_image_against_scene_graph" in source
+    assert "text_labels=text_labels" in source
+
+
+def test_verification_notebook_uses_hash_codec_for_object_hashes():
+    source = NOTEBOOK.read_text()
+
+    assert "bits_tensor_to_hash_bytes" in source
+    assert "np.packbits" not in source