Merge pull request #18 from artuurC/main

Solutions for issue #11 #12 #17

Author: berombau

pyproject.toml

@@ -25,19 +25,19 @@ urls.Documentation = "https://flowsom.readthedocs.io/en/latest/"
 urls.Source = "https://github.com/saeyslab/FlowSOM_Python"
 urls.Home-page = "https://github.com/saeyslab/FlowSOM_Python"
 dependencies = [
-    "mudata<=0.2.4",
+    "mudata",
     "numpy",
     "matplotlib",
     "pandas",
     "scipy",
-    "readfcs<2",
+    "readfcs",
     "scikit-learn",
     "igraph",
     "session-info",
     "numba",
     "scanpy",
     "seaborn",
-    "anndata<=0.10.8",
+    "anndata",
     "loguru",
 ]
 

src/flowsom/io/read_fcs.py

@@ -17,22 +17,22 @@ def read_FCS(filepath):
         f = readfcs.read(filepath, reindex=True)
         f.var.n = f.var.n.astype(int)
         f.var = f.var.sort_values(by="n")
-        f.uns["meta"]["channels"].index = f.uns["meta"]["channels"].index.astype(int)
-        f.uns["meta"]["channels"] = f.uns["meta"]["channels"].sort_index()
+        #f.uns["meta"]["channels"].index = f.uns["meta"]["channels"].index.astype(int)
+        #f.uns["meta"]["channels"] = f.uns["meta"]["channels"].sort_index()
     except ValueError:
         f = readfcs.read(filepath, reindex=False)
         markers = {
             str(re.sub("S$", "", re.sub("^P", "", string))): f.uns["meta"][string]
             for string in f.uns["meta"].keys()
-            if re.match("^P[0-9]+S$", string)
+            if re.match("^p[0-9]+s$", string)
         }
         fluo_channels = list(markers.keys())
         non_fluo_channels = {
-            i: f.uns["meta"]["channels"]["$PnN"][i] for i in f.uns["meta"]["channels"].index if i not in fluo_channels
+            i: f.var["channel"][i] for i in f.var.index if i not in fluo_channels
         }
         index_markers = dict(markers, **non_fluo_channels)
         f.var.rename(index=index_markers, inplace=True)
-        f.uns["meta"]["channels"]["$PnS"] = [index_markers[key] for key in f.uns["meta"]["channels"].index]
+        #f.uns["meta"]["channels"]["$PnS"] = [index_markers[key] for key in f.uns["meta"]["channels"].index]
     return f
 
 

src/flowsom/main.py

@@ -398,7 +398,7 @@ def new_data(self, inp, mad_allowed=4):
         madAllowed * MAD. Default is 4.
         :type mad_allowed: int
         """
-        fsom_new = copy.deepcopy(self)
+        fsom_new = self.copy()
         fsom_new.read_input(inp)
         fsom_new.mad_allowed = mad_allowed
         X = fsom_new.get_cell_data()[:, self.cols_to_use].X
@@ -412,7 +412,7 @@ def subset(self, ids):
         :param ids: An array of ids to subset
         :type ids: np.array
         """
-        fsom_subset = copy.deepcopy(self)
+        fsom_subset = self.copy()
         fsom_subset.mudata.mod["cell_data"] = fsom_subset.mudata["cell_data"][ids, :].copy()
         fsom_subset.model.subset(ids)
         fsom_subset._update_derived_values()
@@ -425,6 +425,31 @@ def get_cell_data(self):
     def get_cluster_data(self):
         """Get the cluster data."""
         return self.mudata["cluster_data"]
+    
+    def copy(self):
+        """
+        Returns a copy of the FlowSOM instance, leveraging MuData's built-in copy method.
+            
+        Returns:
+            FlowSOM: A new instance of FlowSOM with all data copied.
+        """
+        # Create a new instance without calling __init__
+        fsom_copy = self.__class__.__new__(self.__class__)
+        
+        # Copy attributes
+        fsom_copy.cols_to_use = self.cols_to_use
+        fsom_copy.mad_allowed = self.mad_allowed
+        fsom_copy.xdim = self.xdim
+        fsom_copy.ydim = self.ydim
+        fsom_copy.rlen = self.rlen
+        fsom_copy.mst = self.mst
+        fsom_copy.alpha = self.alpha
+        fsom_copy.seed = self.seed
+        fsom_copy.n_clusters = self.n_clusters
+        fsom_copy.model = self.model
+        fsom_copy.mudata = self.mudata.copy()
+        return fsom_copy
+
 
 
 def flowsom_clustering(inp: ad.AnnData, cols_to_use=None, n_clusters=10, xdim=10, ydim=10, **kwargs):

src/flowsom/tl/getter_functions.py

@@ -30,8 +30,8 @@ def get_channels(obj, markers: np.ndarray, exact=True):
             ]
         )
     else:
-        object_markers = np.asarray(obj.uns["meta"]["channels"]["$PnS"])
-        object_channels = np.asarray(obj.uns["meta"]["channels"]["$PnN"])
+        object_markers = np.asarray(obj.var["marker"])
+        object_channels = np.asarray(obj.var["channel"])
 
     channelnames = {}
     for marker in markers:
@@ -78,8 +78,8 @@ def get_markers(obj, channels, exact=True):
             ]
         )
     else:
-        object_markers = np.asarray(obj.uns["meta"]["channels"]["$PnS"])
-        object_channels = np.asarray(obj.uns["meta"]["channels"]["$PnN"])
+        object_markers = np.asarray(obj.var["marker"])
+        object_channels = np.asarray(obj.var["channel"])
 
     markernames = {}
     for channel in channels:
@@ -166,7 +166,7 @@ def get_cluster_percentages_positive(fsom, cutoffs, cols_used=False, pretty_coln
         channels = {key: value for key, value in channels.items() if key in markers}
 
     perc_pos = np.empty((len(clusters), len(channels)))
-    perc_pos.fill(np.NaN)
+    perc_pos.fill(np.nan)
     for i, cluster in enumerate(clusters):
         data_per_cluster = fsom.get_cell_data().to_df().loc[cl_per_cell == cluster, list(channels.keys())]
         if data_per_cluster.shape[0] != 0:
@@ -207,7 +207,7 @@ def get_metacluster_percentages_positive(fsom, cutoffs, cols_used=False, pretty_
         channels = {key: value for key, value in channels.items() if key in markers}
 
     perc_pos = np.empty((len(metaclusters), len(channels)))
-    perc_pos.fill(np.NaN)
+    perc_pos.fill(np.nan)
     for i, cluster in enumerate(metaclusters):
         data_per_metacluster = fsom.get_cell_data().to_df().loc[mcl_per_cell == cluster, list(channels.keys())]
         if data_per_metacluster.shape[0] != 0: