From 5b0a66dd9e77b96e6e819b274ac3f9425c7df912 Mon Sep 17 00:00:00 2001
From: Dries Schaumont <5946712+DriesSchaumont@users.noreply.github.com>
Date: Fri, 16 Aug 2024 16:52:18 +0200
Subject: [PATCH] Bump liana to 1.3.0 (#862)

---
 CHANGELOG.md                          |  2 +-
 src/interpret/lianapy/config.vsh.yaml |  5 ++---
 src/interpret/lianapy/script.py       | 16 ++++++++++++----
 src/interpret/lianapy/test.py         | 10 +++++-----
 4 files changed, 20 insertions(+), 13 deletions(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 083aaa0b2ac..5e16d4b1ddb 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -10,7 +10,7 @@
 
 * Removed `--disable_library_compatibility_check` in favour of `--check_library_compatibility` to the `mapping/cellranger_multi` component and the `ingestion/cellranger_multi` workflow (PR #818).
 
-* `lianapy`: bumped version to `1.2.1` (PR #827).
+* `lianapy`: bumped version to `1.3.0` (PR #827 and PR #862). Additionally, `groupby` is now a required argument.
 
 * `XGBoost`: bump version to `2.0.3` (PR #646).
 
diff --git a/src/interpret/lianapy/config.vsh.yaml b/src/interpret/lianapy/config.vsh.yaml
index dd1ebf99907..942fa3b709d 100644
--- a/src/interpret/lianapy/config.vsh.yaml
+++ b/src/interpret/lianapy/config.vsh.yaml
@@ -35,8 +35,7 @@ functionality:
       description: "Layer in anndata.AnnData.layers to use. If None, use mudata.mod[modality].X."
     - name: "--groupby"
       type: string
-      default: "bulk_labels"
-      required: false
+      required: true
       description: "The key of the observations grouping to consider."
     - name: "--resource_name"
       type: string
@@ -93,7 +92,7 @@ platforms:
       - type: python
         __merge__: [/src/base/requirements/anndata_mudata.yaml, .]
         packages:
-          - liana~=1.2.1
+          - liana~=1.3.0
     test_setup:
       - type: python
         __merge__: [ /src/base/requirements/viashpy.yaml, .]
diff --git a/src/interpret/lianapy/script.py b/src/interpret/lianapy/script.py
index 66c1310d9ca..cee30e73e5f 100644
--- a/src/interpret/lianapy/script.py
+++ b/src/interpret/lianapy/script.py
@@ -3,16 +3,17 @@
 # TODO: Remove when grouping labels exist
 # For sign/PCA/
 import numpy as np
+import pandas as pd
 
 ### VIASH START
 par = {
-    "input": "integration.pca.output.h5mu",
+    "input": "resources_test/pbmc_1k_protein_v3/pbmc_1k_protein_v3_mms.h5mu",
     "output": "foo.h5mu",
     "output_compression": "gzip",
     "modality": "rna",
     "layer": None,
     "gene_symbol": "gene_symbol",
-    "groupby": "bulk_labels",
+    "groupby": "harmony_integration_leiden_1.0",
     "resource_name": "consensus",
     "expr_prop": 0.1,
     "min_cells": 5,
@@ -31,8 +32,12 @@ def main():
 
     # Add dummy grouping labels when they do not exist
     if par['groupby'] not in mod.obs:
-        foo = mod.obsm.to_df().iloc[:, 0]
-        mod.obs[par['groupby']] = np.sign(foo).astype('category')
+        raise ValueError(f"Column {par['groupy']} does not exist in "
+                         f".obs for modality {par['modality']}.")
+    mod_col = mod.obs[par['groupby']]
+    original_groupby_col = mod_col.copy()
+    if not isinstance(mod_col, pd.CategoricalDtype):
+        mod.obs[par['groupby']] = mod_col.astype(str).astype('category')
 
     # Solve gene labels
     orig_gene_label = mod.var.index
@@ -57,6 +62,9 @@ def main():
     # Return original gene labels
     mod.var_names = orig_gene_label
 
+    # Undo modifications to groupby column
+    mod.obs[par["groupby"]] = original_groupby_col
+
     # TODO: make sure compression is needed
     mdata.write_h5mu(par['output'].strip(), compression=par['output_compression'])
 
diff --git a/src/interpret/lianapy/test.py b/src/interpret/lianapy/test.py
index f6f7b150d53..cad96d05381 100644
--- a/src/interpret/lianapy/test.py
+++ b/src/interpret/lianapy/test.py
@@ -5,7 +5,7 @@
 
 ## VIASH START
 meta = {
-    'executable': './target/native/interpret/lianapy/',
+    'executable': './target/executable/interpret/lianapy/',
     'resources_dir': './resources_test/pbmc_1k_protein_v3/'
 }
 ## VIASH END
@@ -21,8 +21,8 @@ def test_lianapy(run_component, tmp_path):
         "--input", input_file,
         "--output_compression", "gzip",
         "--modality", "rna",
-        # "--layer", "log_normalized",
-        "--groupby", "bulk_labels",
+        "--layer", "log_normalized",
+        "--groupby", "harmony_integration_leiden_1.0",
         "--resource_name", "consensus",
         "--gene_symbol", "gene_symbol",
         "--expr_prop", "0.1",
@@ -39,8 +39,8 @@ def test_lianapy(run_component, tmp_path):
     np.testing.assert_array_equal(output_data.mod['rna'].X.data, input_data.mod['rna'].X.data)
     np.testing.assert_array_equal(input_data.mod['rna'].var.index, output_data.mod['rna'].var.index)
     assert "liana_res" in output_data.mod["rna"].uns
-    assert all(elem in output_data.mod['rna'].obs['bulk_labels'].values for elem in output_data.mod['rna'].uns['liana_res']['source'].unique())
-    assert all(elem in output_data.mod['rna'].obs['bulk_labels'].values for elem in output_data.mod['rna'].uns['liana_res']['target'].unique())
+    assert all(elem in output_data.mod['rna'].obs['harmony_integration_leiden_1.0'].values for elem in output_data.mod['rna'].uns['liana_res']['source'].unique())
+    assert all(elem in output_data.mod['rna'].obs['harmony_integration_leiden_1.0'].values for elem in output_data.mod['rna'].uns['liana_res']['target'].unique())
 
 if __name__ == "__main__":
     sys.exit(pytest.main([__file__]))
\ No newline at end of file