Fixed warnings, added read_csv, fix installation issue with windows

src/FlowSOM/__init__.py

@@ -5,4 +5,4 @@
 
 __all__ = ["FlowSOM", "flowsom_clustering", "io", "pl", "pp", "tl"]
 
-__version__ = version("FlowSOM")
+__version__ = version("flowsom")

src/FlowSOM/io/__init__.py

@@ -1,3 +1,3 @@
-from .read_fcs import read_FCS
+from .read_fcs import read_FCS, read_csv
 
-__all__ = ["read_FCS"]
+__all__ = ["read_FCS", "read_csv"]

src/FlowSOM/io/read_fcs.py

@@ -3,6 +3,8 @@
 import re
 
 import pytometry as pm
+import anndata as ad
+import pandas as pd
 
 
 def read_FCS(filepath):
@@ -32,3 +34,13 @@ def read_FCS(filepath):
         f.var.rename(index=index_markers, inplace=True)
         f.uns["meta"]["channels"]["$PnS"] = [index_markers[key] for key in f.uns["meta"]["channels"].index]
     return f
+
+
+def read_csv(filepath, spillover=None, **kwargs):
+    ff = ad.read_csv(filepath, **kwargs)
+    ff.var = pd.DataFrame(
+        {"n": range(ff.shape[1]), "channel": ff.var_names, "marker": ff.var_names}, index=ff.var.index
+    )
+    if spillover is not None:
+        ff.uns["meta"]["SPILL"] = pd.read_csv(spillover)
+    return ff

src/FlowSOM/main.py

@@ -12,7 +12,7 @@
 from scipy.stats import median_abs_deviation
 from sklearn.cluster import AgglomerativeClustering
 
-from .io import read_FCS
+from .io import read_FCS, read_csv
 from .tl import SOM, ConsensusCluster, get_channels, get_markers, map_data_to_codes
 
 
@@ -48,7 +48,10 @@ def read_input(self, inp):
         if isinstance(inp, ad.AnnData):
             data = inp.X
         elif isinstance(inp, str):
-            inp = read_FCS(inp)
+            if inp.endswith(".csv"):
+                inp = read_csv(inp)
+            elif inp.endswith(".fcs"):
+                inp = read_FCS(inp)
             data = inp.X
         channels = np.asarray(inp.var["channel"])
         markers = np.asarray(inp.var["marker"])

src/FlowSOM/pl/_plot_helper_functions.py

@@ -131,7 +131,7 @@ def plot_FlowSOM(
     # Add background
     if background_values is not None:
         if equal_background_size:
-            background_size = np.repeat(max_node_size * background_size, len(background_values))
+            background_size = np.repeat(np.max(node_sizes) * background_size, len(background_values))
         else:
             background_size = (
                 parse_node_sizes(
@@ -217,8 +217,10 @@ def plot_star_legend(fig, ax, markers, coords=(0, 0), cmap=FlowSOM_colors(), max
     n_left_right = pd.crosstab(n_left_right, columns="x")
     if n_left_right.shape[0] != 1:
         by = 1 if len(markers) <= 8 else 0.65
-        left = np.linspace(start=0, stop=(n_left_right.x[0] - 1) * by, num=n_left_right.x[0])
-        right = np.multiply(-1, np.linspace(start=0, stop=(n_left_right.x[1] - 1) * by, num=n_left_right.x[1]))
+        left = np.linspace(start=0, stop=(n_left_right.x.iloc[0] - 1) * by, num=n_left_right.x.iloc[0])
+        right = np.multiply(
+            -1, np.linspace(start=0, stop=(n_left_right.x.iloc[1] - 1) * by, num=n_left_right.x.iloc[1])
+        )
         segments_left = segments[segments[:, 1] < 0, :]
         segments_left = segments_left[segments_left[:, 2].argsort()]
         segments_right = segments[segments[:, 1] >= 0]
@@ -266,6 +268,8 @@ def plot_star_legend(fig, ax, markers, coords=(0, 0), cmap=FlowSOM_colors(), max
 
 
 def scale_star_heights(median_values, node_sizes):
+    if isinstance(node_sizes, pd.Series):
+        node_sizes = node_sizes.to_numpy()
     max_all_nodes = median_values[~np.isnan(median_values)].max()
     min_all_nodes = median_values[~np.isnan(median_values)].min()
     scaled_row = [
@@ -297,7 +301,9 @@ def auto_max_node_size(layout, overlap):
     return min_distance / 2 * overlap
 
 
-def add_text(ax, layout, text, text_size=20, text_color="black", ha=None, va=None):
+def add_text(ax, layout, text, text_size=25, text_color="black", ha=None, va=None):
+    if isinstance(text, pd.Series):
+        text = text.to_numpy()
     if va is None:
         va = ["center"]
     if ha is None:
@@ -327,16 +333,22 @@ def parse_edges(fsom):
 
 
 def add_nodes(layout, heights):
+    if isinstance(heights, pd.Series):
+        heights = heights.to_numpy()
     patches = [Circle((row[0], row[1]), heights[i]) for i, row in enumerate(layout)]
     return patches
 
 
 def add_stars(layout, heights):
+    if isinstance(heights, pd.Series):
+        heights = heights.to_numpy()
     patches = np.hstack([add_wedges((row[0], row[1]), heights[i, :]) for i, row in enumerate(layout)])
     return patches
 
 
 def add_wedges(coord, heights, angles=None):
+    if isinstance(heights, pd.Series):
+        heights = heights.to_numpy()
     if angles is None:
         part = 360 / len(heights)
         angles = np.arange(0, 360.01, part)

src/FlowSOM/pl/plot_functions.py

@@ -369,7 +369,7 @@ def plot_pies(
             if 0 not in angles:
                 angles = np.insert(angles, 0, 0)
             row = layout[cl, :]
-            patches = add_wedges(tuple(row), heights=np.repeat(scaled_node_size[cl], len(angles)), angles=angles)
+            patches = add_wedges(tuple(row), heights=np.repeat(scaled_node_size.iloc[cl], len(angles)), angles=angles)
             p = mc.PatchCollection(patches)
             p.set_facecolor([color_dict.get(key) for key in table.index.values])
             p.set_edgecolor("black")