Select group of cells

.gitignore

@@ -192,5 +192,5 @@ cython_debug/
 
 .history/
 
-# Claude
+#Claude
 .claude/

package.json

@@ -6,7 +6,7 @@
   "scripts": {
     "dev": "vite",
     "build": "tsc && vite build",
-    "build:python": "tsc && vite build && mkdir -p python/src/intracktive/frontend/assets && cp dist/index.html dist/spark1.png dist/Biohub_logo_60x60.png dist/icon_zebrafish.png python/src/intracktive/frontend/ && cp dist/assets/* python/src/intracktive/frontend/assets/",
+    "build:python": "tsc && vite build && rm -rf python/src/intracktive/frontend && mkdir -p python/src/intracktive/frontend/assets && cp dist/index.html dist/spark1.png dist/Biohub_logo_60x60.png dist/icon_zebrafish.png python/src/intracktive/frontend/ && cp dist/assets/* python/src/intracktive/frontend/assets/",
     "preview": "vite preview",
     "lint": "eslint src",
     "format": "prettier --write src",

python/src/intracktive/add_annotations.py

@@ -0,0 +1,113 @@
+"""
+Standalone script to precompute annotated track point IDs for an inTRACKtive zarr store.
+
+For each hex-binary attribute, finds all cells with a non-background annotation value,
+looks up their track IDs, and stores one representative point ID per unique annotated track
+in attributes.attrs["annot_point_ids"] (a list of lists, parallel to attribute_names).
+
+Usage:
+    python -m intracktive.add_annotations <zarr_path>
+"""
+
+import sys
+from pathlib import Path
+
+import numpy as np
+import zarr
+
+# The hex value used to represent "no annotation" (grey: #404040)
+NO_ANNOTATION_VALUE = 4210752
+# Sentinel value used to pad the jagged array
+INF_SPACE = -9000.0
+
+
+def add_annotated_point_ids(zarr_path: str | Path) -> None:
+    """
+    Process an existing inTRACKtive zarr store and write annot_point_ids to attributes.attrs.
+
+    For each hex-binary attribute column, scans all timepoints to find annotated cells,
+    resolves their track IDs via points_to_tracks, and stores one representative point ID
+    per unique annotated track. Non-hex-binary attributes get an empty list.
+
+    Parameters
+    ----------
+    zarr_path : str or Path
+        Path to the inTRACKtive zarr store.
+    """
+    zarr_path = Path(zarr_path)
+    store = zarr.open(zarr_path.as_posix(), mode="r+")
+
+    if "attributes" not in store:
+        print("No attributes array found in zarr — nothing to do.")
+        return
+
+    attributes = store["attributes"]
+    zattrs = dict(attributes.attrs)
+
+    attribute_names = zattrs.get("attribute_names", [])
+    attribute_types = zattrs.get("attribute_types", [])
+
+    if not attribute_names:
+        print("No attribute_names found in attributes.zattrs — nothing to do.")
+        return
+
+    n_attributes = len(attribute_names)
+    max_pts = attributes.shape[1] // n_attributes
+
+    # Load points_to_tracks CSR arrays once
+    indptr = store["points_to_tracks/indptr"][:]
+    indices = store["points_to_tracks/indices"][:]
+
+    annot_point_ids: list[list[int]] = []
+    annot_colors: list[int | None] = []
+
+    for col_idx, (name, attr_type) in enumerate(zip(attribute_names, attribute_types)):
+        if attr_type != "hex-binary":
+            annot_point_ids.append([])
+            annot_colors.append(None)
+            continue
+
+        print(f"Processing hex-binary attribute '{name}' (column {col_idx})...")
+
+        # Load full attribute column for all timepoints: shape (n_times, max_pts)
+        start_col = col_idx * max_pts
+        end_col = start_col + max_pts
+        attr_data = attributes[:, start_col:end_col]
+
+        # Find all (t, i) that are annotated: not background and not INF_SPACE padding
+        mask = (attr_data != NO_ANNOTATION_VALUE) & (attr_data > INF_SPACE)
+        t_indices, i_indices = np.where(mask)
+        point_ids = (t_indices * max_pts + i_indices).astype(np.int64)
+
+        # For each point, look up its track(s) and keep one representative point ID per track
+        seen_tracks: set[int] = set()
+        representative_point_ids: list[int] = []
+
+        for point_id in point_ids:
+            pid = int(point_id)
+            start = int(indptr[pid])
+            end = int(indptr[pid + 1])
+            for track_id in indices[start:end]:
+                tid = int(track_id)
+                if tid not in seen_tracks:
+                    seen_tracks.add(tid)
+                    representative_point_ids.append(pid)
+
+        # Extract tissue color (all annotated cells share the same hex color for hex-binary)
+        annotated_values = attr_data[mask]
+        tissue_color = int(annotated_values[0]) if len(annotated_values) > 0 else None
+        annot_colors.append(tissue_color)
+
+        print(f"  → {len(seen_tracks)} annotated tracks found")
+        annot_point_ids.append(representative_point_ids)
+
+    attributes.attrs["annot_point_ids"] = annot_point_ids
+    attributes.attrs["annot_colors"] = annot_colors
+    print(f"Saved annot_point_ids to {zarr_path / 'attributes' / '.zattrs'}")
+
+
+if __name__ == "__main__":
+    if len(sys.argv) != 2:
+        print("Usage: python -m intracktive.add_annotations <zarr_path>")
+        sys.exit(1)
+    add_annotated_point_ids(sys.argv[1])

python/src/intracktive/frontend/index.html

@@ -10,7 +10,7 @@
         rel="stylesheet"
     />
     <title>inTRACKtive</title>
-    <script type="module" crossorigin src="/assets/index-CLy-n6J5.js"></script>
+    <script type="module" crossorigin src="/assets/index-BLTjMDY-.js"></script>
     <link rel="stylesheet" crossorigin href="/assets/index-3tFKGWso.css">
   </head>
   <body>