Add script to subset an h5ad file, use it to reduce the size of the s…

…ample.h5ad to make loading faster initially from github pages

Makefile

@@ -1,7 +1,11 @@
+serve:
+	python -m http.server 3000
+
 check-model:
 	python -m onnxruntime.tools.check_onnx_model_mobile_usability --log_level debug data/sims.onnx
 
 deploy:
-	rsync -v ./*.{js,html} [email protected]:~/public_html/sims/
-	rsync -avz models [email protected]:~/public_html/sims/
-	# rsync -avz data [email protected]:~/public_html/sims/
+	git checkout gh-pages
+	git merge main
+	git push origin gh-pages
+	git checkout main

index.html

@@ -39,8 +39,8 @@ <h2 class="mb-0" style="font-weight: 300">SIMS Web</h2>
                     </div>
                     <div class="form-group">
                         <label for="cellRange">% Of Cells To Process:</label>
-                        <input type="range" class="custom-range" id="cellRange" min="0" max="100" step="1" value="1">
-                        <small id="cellRangeValue" class="form-text text-muted">1%</small>
+                        <input type="range" class="custom-range" id="cellRange" min="0" max="100" step="1" value="10">
+                        <small id="cellRangeValue" class="form-text text-muted">10%</small>
                     </div>
                     <button id="predict_btn" class="btn btn-primary" autofocus>Predict</button>
                     <button id="stop_btn" class="btn btn-danger">Stop</button>

subset-anndata.py

@@ -0,0 +1,53 @@
+#!/usr/bin/env python3
+
+import anndata
+import numpy as np
+import argparse
+import sys
+
+def sample_anndata(input_file: str, output_file: str, n_cells: int):
+    """
+    Reads an .h5ad AnnData file, randomly selects a subset of cells, and saves a new .h5ad file.
+
+    Parameters:
+        input_file (str): Path to the input .h5ad file.
+        output_file (str): Path where the output .h5ad file will be saved.
+        n_cells (int): Number of cells to randomly select.
+    """
+    # Read the input AnnData file
+    adata = anndata.read_h5ad(input_file)
+
+    # Total number of cells in the dataset
+    total_cells = adata.n_obs
+
+    # Check if requested number of cells is valid
+    if n_cells > total_cells:
+        raise ValueError(f"Requested number of cells ({n_cells}) exceeds total cells in the dataset ({total_cells}).")
+
+    # Randomly select indices of cells to keep
+    np.random.seed(42)  # Optional: Set seed for reproducibility
+    selected_cells = np.random.choice(total_cells, size=n_cells, replace=False)
+
+    # Subset the AnnData object to the selected cells
+    adata_subset = adata[selected_cells, :].copy()
+
+    # Save the subsetted AnnData object to a new file
+    adata_subset.write_h5ad(output_file)
+
+def main():
+    parser = argparse.ArgumentParser(description='Sample a subset of cells from an AnnData .h5ad file.')
+    parser.add_argument('input_file', type=str, help='Path to the input .h5ad file.')
+    parser.add_argument('output_file', type=str, help='Path to the output .h5ad file.')
+    parser.add_argument('--n_cells', '-n', type=int, default=1000, help='Number of cells to sample (default: 1000).')
+
+    args = parser.parse_args()
+
+    try:
+        sample_anndata(args.input_file, args.output_file, args.n_cells)
+        print(f"Successfully saved {args.n_cells} cells to {args.output_file}")
+    except Exception as e:
+        print(f"Error: {e}", file=sys.stderr)
+        sys.exit(1)
+
+if __name__ == '__main__':
+    main()