fix the bootstrapping component to align with what one expects from a…

… bootstrapping method (#61)

src/task/process_dataset/bootstrap/config.vsh.yaml

@@ -11,56 +11,51 @@ functionality:
   argument_groups:
     - name: Inputs
       arguments:
-      - name: --train_h5ad
+      - name: --input
         type: file
         required: true
         direction: input
-        example: resources/neurips-2023-kaggle/de_train.h5ad
-      - name: --test_h5ad
-        type: file
-        required: true
-        direction: input
-        example: resources/neurips-2023-kaggle/de_test.h5ad
+        example: resources/neurips-2023-raw/sc_counts_reannotated_with_counts.h5ad
     - name: Outputs
       arguments:
-        - name: --output_train_h5ad
+        - name: --output
           type: file
           required: true
           direction: output
-          multiple: true
-          example: [de_train_bootstrap.h5ad]
-        - name: --output_test_h5ad
-          type: file
-          required: true
-          direction: output
-          multiple: true
-          example: [de_test_bootstrap.h5ad]
-    - name: Arguments
+          example: sc_counts_bootstrap.h5ad
+    - name: Sampling parameters
+      description: Parameters for sampling the bootstraps.
       arguments:
-        - name: --num_replicates
-          type: integer
-          required: true
-          default: 10
-          description: Number of bootstraps to generate.
+        - name: --bootstrap_obs
+          type: boolean
+          default: true
+          description: Whether to sample observations.
         - name: --obs_fraction
           type: double
-          required: true
-          default: 0.95
-          description: Fraction of the training dataset obs to include in each bootstrap.
+          default: 1
+          description: Fraction of the obs of the sc_counts to include in each bootstrap.
+        - name: --obs_replace
+          type: boolean
+          default: true
+          description: Whether to sample with replacement.
+        - name: --bootstrap_var
+          type: boolean
+          default: false
+          description: Whether to sample variables.
         - name: --var_fraction
           type: double
-          required: true
-          default: 0.95
-          description: Fraction of the training & test dataset var to include in each bootstrap.
+          default: 1
+          description: Fraction of the var of the sc_counts to include in each bootstrap.
+        - name: --var_replace
+          type: boolean
+          default: true
+          description: Whether to sample with replacement.
   resources:
-    - type: r_script
-      path: script.R
+    - type: python_script
+      path: script.py
 platforms:
   - type: docker
-    image: ghcr.io/openproblems-bio/base_r:1.0.4
-    setup:
-      - type: r
-        cran: [ dplyr, tidyr, purrr, tibble, arrow ]
+    image: ghcr.io/openproblems-bio/base_python:1.0.4
   - type: nextflow
     directives:
-      label: [ midtime, midmem, lowcpu ]
+      label: [ midtime, highmem, midcpu ]

src/task/process_dataset/bootstrap/script.py

@@ -0,0 +1,39 @@
+import anndata as ad
+import numpy as np
+
+# VIASH START
+par = {
+    "input": "resources/neurips-2023-raw/sc_counts_reannotated_with_counts.h5ad",
+    "output": "output/sc_counts_bootstrapped_*.h5ad",
+    "bootstrap_obs": True,
+    "obs_fraction": 1,
+    "obs_replace": True,
+    "bootstrap_var": False,
+    "var_fraction": 1,
+    "var_replace": True
+}
+# VIASH END
+
+# Load data
+data = ad.read_h5ad(par["input"])
+
+if par["bootstrap_obs"]:
+    # Sample indices
+    obs_ix = np.random.choice(
+        range(data.n_obs),
+        int(data.n_obs * par["obs_fraction"]),
+        replace=par["obs_replace"]
+    )
+    data = data[obs_ix, :]
+
+if par["bootstrap_var"]:
+    # Sample indices
+    var_ix = np.random.choice(
+        range(data.n_vars),
+        int(data.n_vars * par["var_fraction"]),
+        replace=par["var_replace"]
+    )
+    data = data[:, var_ix]
+
+# Write output
+data.write_h5ad(par["output"], compression="gzip")

src/task/workflows/run_stability_analysis/config.vsh.yaml

@@ -20,21 +20,30 @@ functionality:
     - name: Bootstrapping arguments
       description: Define the sampling strategy for the stability analysis.
       arguments:
-        - name: --bootstrap_num_replicates
-          type: integer
-          required: true
-          default: 10
-          description: Number of bootstraps to generate.
+        - name: --bootstrap_obs
+          type: boolean
+          default: true
+          description: Whether to sample observations.
         - name: --bootstrap_obs_fraction
           type: double
-          required: true
-          default: 0.95
+          default: 1
           description: Fraction of the obs of the sc_counts to include in each bootstrap.
+        - name: --bootstrap_obs_replace
+          type: boolean
+          default: true
+          description: Whether to sample with replacement.
+        - name: --bootstrap_var
+          type: boolean
+          default: false
+          description: Whether to sample variables.
         - name: --bootstrap_var_fraction
           type: double
-          required: true
           default: 1
           description: Fraction of the var of the sc_counts to include in each bootstrap.
+        - name: --bootstrap_var_replace
+          type: boolean
+          default: true
+          description: Whether to sample with replacement.
     - name: Outputs
       arguments:
         - name: "--scores"
@@ -60,7 +69,7 @@ functionality:
     - type: file
       path: "../../api/task_info.yaml"
   dependencies:
-    - name: process_dataset/bootstrap_sc_counts
+    - name: process_dataset/bootstrap
     - name: workflows/process_dataset
     - name: workflows/run_benchmark
   repositories:

src/task/workflows/run_stability_analysis/main.nf

@@ -28,11 +28,15 @@ workflow run_wf {
       }
     }
 
-    | bootstrap_sc_counts.run(
+    | bootstrap.run(
       fromState: [
         input: "sc_counts",
+        bootstrap_obs: "bootstrap_obs",
         obs_fraction: "bootstrap_obs_fraction",
-        var_fraction: "bootstrap_var_fraction"
+        obs_replace: "bootstrap_obs_replace",
+        bootstrap_var: "bootstrap_var",
+        var_fraction: "bootstrap_var_fraction",
+        var_replace: "bootstrap_var_replace"
       ],
       toState: [
         sc_counts: "output"