Merge pull request #78 from PixelgenTechnologies/chore/fix-a-bunch-of…

…-warnings

Fix a bunch of warnings generated in the tests

pyproject.toml

@@ -50,7 +50,7 @@ pyarrow = ">=13,<15"
 semver = "^3.0.0"
 ruamel-yaml = "^0.17.21"
 pydantic = "^1.10.7"
-polars = "^0.19"
+polars = "^0.20"
 importlib-resources = "^5.12.0"
 fsspec = "^2023.9.0"
 fastparquet = "^2023.8.0"

src/pixelator/analysis/colocalization/__init__.py

@@ -66,7 +66,9 @@ def colocalization_from_component_edgelist(
     graph = Graph.from_edgelist(
         edgelist=edgelist,
         add_marker_counts=True,
-        simplify=False,
+        # If we do A-node projection, we will simplify anyway.
+        # This just removes the warning.
+        simplify=False if use_full_bipartite else True,
         use_full_bipartite=use_full_bipartite,
     )
     return colocalization_from_component_graph(
@@ -244,7 +246,9 @@ def data():
             graph = Graph.from_edgelist(
                 edgelist=component_df,
                 add_marker_counts=True,
-                simplify=False,
+                # If we do A-node projection, we will simplify anyway.
+                # This just removes the warning.
+                simplify=False if use_full_bipartite else True,
                 use_full_bipartite=use_full_bipartite,
             )
             if len(graph.vs) < 2:

src/pixelator/analysis/colocalization/prepare.py

@@ -41,7 +41,9 @@ def prepare_from_edgelist(
         .groupby(by=group_by)["marker"]
         .value_counts()
         .reset_index()
-    ).pivot_table(index=group_by, columns="marker", values="count", fill_value=0)
+    ).pivot_table(
+        index=group_by, columns="marker", values="count", fill_value=0, observed=False
+    )
     markers_per_pixel.columns.name = "markers"
     markers_per_pixel.index.name = "node"
     markers_per_pixel.index = markers_per_pixel.index.astype("string[pyarrow]")

src/pixelator/analysis/colocalization/statistics.py

@@ -24,7 +24,7 @@ def _wide_correlation_matrix_to_long_correlation_results(
     df_lower_tri = df.mask(~lower_triangle)
     df_lower_tri.index.set_names(["marker_cols"], inplace=True)
     df_lower_tri.columns.set_names(["marker_rows"], inplace=True)
-    correlation_values = df_lower_tri.stack(dropna=True).reset_index()
+    correlation_values = df_lower_tri.stack(future_stack=True).dropna().reset_index()
     correlation_values.columns = ["marker_1", "marker_2", stat_name]
     correlation_values.set_index(["marker_1", "marker_2"], inplace=True)
     correlation_values.index.rename(["marker_1", "marker_2"])

src/pixelator/config/panel.py

@@ -183,23 +183,25 @@ def _transform_legacy_panels(cls, df: pd.DataFrame) -> pd.DataFrame:
         :param df: DataFrame with data of the panel to validate
         :returns pd.DataFrame: The in-place modified input dataframe
         """
-        new_df = df.copy()
-
         # update control and nuclear column to boolean
-        TR_TABLE = {"(?i)no": "False", "(?i)yes": "True"}
-
-        new_df.replace(
-            {
-                "control": TR_TABLE,
-                "nuclear": TR_TABLE,
-            },
-            regex=True,
-            inplace=True,
+        TR_TABLE = {"(?i)yes": "True", "(?i)no": "False"}
+
+        df["control"] = (
+            df["control"]
+            .astype("string[pyarrow]")
+            .fillna("")
+            .replace(TR_TABLE, regex=True)
+            .astype(bool)
+        )
+        df["nuclear"] = (
+            df["nuclear"]
+            .astype("string[pyarrow]")
+            .fillna("")
+            .replace(TR_TABLE, regex=True)
+            .astype(bool)
         )
-        new_df.fillna({"control": "False", "nuclear": "False"}, inplace=True)
-        new_df["control"] = new_df["control"].map({"True": True, "False": False})
-        new_df["nuclear"] = new_df["nuclear"].map({"True": True, "False": False})
-        return new_df
+
+        return df
 
     @classmethod
     def _parse_header(cls, file: Path) -> AntibodyPanelMetadata:

src/pixelator/report/qcreport/collect.py

@@ -116,9 +116,9 @@ def collect_reads_per_umi_frequency(dataset: PixelDataset) -> str:
 
     freq = freq.with_columns(
         pl.col("reads_per_umi"),
-        frequency=pl.col("counts") / pl.col("counts").sum(),
+        frequency=pl.col("count") / pl.col("count").sum(),
     )
-    pd_freq = freq.rename({"counts": "count"}).sort(by="reads_per_umi").to_pandas()
+    pd_freq = freq.sort(by="reads_per_umi").to_pandas()
 
     return pd_freq.to_csv(index=False)
 

tests/data/param_files/test_data_pe_T1.amplicon.meta.json

@@ -1,10 +1,8 @@
 {
-    "cli": {
-        "command": "pixelator single-cell amplicon",
-        "options": {
-            "--input1-pattern": "_R1",
-            "--input2-pattern": "_R2",
-            "--output": "."
-        }
+  "cli": {
+    "command": "pixelator single-cell amplicon",
+    "options": {
+      "--output": "."
     }
+  }
 }

tests/data/report/info_and_metrics.json

@@ -364,18 +364,6 @@
       {
         "command": "pixelator single-cell amplicon",
         "options": [
-          {
-            "name": "--input1-pattern",
-            "value": "_R1",
-            "default_value": "_R1",
-            "description": "The string pattern to use to identify forward (R1) files"
-          },
-          {
-            "name": "--input2-pattern",
-            "value": "_R2",
-            "default_value": "_R2",
-            "description": "The string pattern to use to identify reverse (R2) files"
-          },
           {
             "name": "--output",
             "value": ".",

tests/graph/test_graph.py

@@ -25,8 +25,8 @@ def create_simple_edge_list_from_graph(
 
     df = graph.get_edge_dataframe()
     df_vert = graph.get_vertex_dataframe()
-    df["source"].replace(df_vert["name"], inplace=True)
-    df["target"].replace(df_vert["name"], inplace=True)
+    df["source"] = df["source"].replace(df_vert["name"])
+    df["target"] = df["target"].replace(df_vert["name"])
 
     # rename source/target columns
     df = df.rename(columns={"source": "upib", "target": "upia"})

tests/plot/test_plot.py

@@ -3,8 +3,8 @@
 Copyright (c) 2023 Pixelgen Technologies AB.
 """
 
-
 import numpy as np
+import pytest
 from numpy.testing import assert_almost_equal
 from pixelator.graph import Graph
 from pixelator.plot import (
@@ -460,10 +460,17 @@ def test_plot_3d_heatmap(edgelist):
         simplify=True,
         use_full_bipartite=False,
     )
-    # For now, just making sure that this doesn't crash when running.
-    _ = plot_3d_heatmap(
-        graph,
-        layout_algorithm="fruchterman_reingold_3d",
-        marker="CD3",
-        distance_cutoff=0.4,
-    )
+    with pytest.warns(
+        UserWarning,
+        match=(
+            "The function `plot_3d_heatmap` is experimental, "
+            "it might be removed or the API might change without notice."
+        ),
+    ):
+        # For now, just making sure that this doesn't crash when running.
+        _ = plot_3d_heatmap(
+            graph,
+            layout_algorithm="fruchterman_reingold_3d",
+            marker="CD3",
+            distance_cutoff=0.4,
+        )