Switch from _is_na to pd.isna

Author: grst

src/scirpy/ir_dist/__init__.py

@@ -1,3 +1,5 @@
+import pandas as pd
+
 """Compute distances between immune receptor sequences"""
 
 from collections.abc import Sequence
@@ -8,7 +10,7 @@
 from scipy.sparse import csr_matrix
 
 from scirpy.get import airr as get_airr
-from scirpy.util import DataHandler, _doc_params, _is_na, deprecated
+from scirpy.util import DataHandler, _doc_params, deprecated
 
 from . import metrics
 
@@ -237,7 +239,7 @@ def _get_unique_seqs(tmp_adata, chain_type):
         tmp_seqs = np.concatenate(
             [get_airr(tmp_adata, key, f"{chain_type}_{chain_id}").values for chain_id in ["1", "2"]]  # type: ignore
         )
-        return np.unique([x.upper() for x in tmp_seqs[~_is_na(tmp_seqs)]])
+        return np.unique([x.upper() for x in tmp_seqs[~pd.isna(tmp_seqs)]])
 
     for i, tmp_params in enumerate([params, params_ref]):
         if tmp_params is not None:

src/scirpy/pl/_vdj_usage.py

@@ -4,10 +4,11 @@
 
 import matplotlib.pyplot as plt
 import numpy as np
+import pandas as pd
 
 from scirpy.get import airr as get_airr
 from scirpy.io import AirrCell
-from scirpy.util import DataHandler, _is_na, _normalize_counts
+from scirpy.util import DataHandler, _normalize_counts
 
 from .styling import _init_ax
 
@@ -163,7 +164,7 @@ def vdj_usage(
 
         # Draw gene segments
         for i, (segment_size, gene) in list(enumerate(zip(segment_sizes, genes, strict=False)))[:max_segments][::-1]:
-            if _is_na(gene):
+            if pd.isna(gene):
                 gene = "none"
             gene_tops[col_name][gene] = bottom + segment_size
             if draw_bars:
@@ -229,15 +230,15 @@ def vdj_usage(
                 try:
                     tmp_gene = ribbon[target_pair[0]]
                     tmp_col = target_pair[0]
-                    tmp_gene = "none" if _is_na(tmp_gene) else tmp_gene
+                    tmp_gene = "none" if pd.isna(tmp_gene) else tmp_gene
                     ribbon_color = gene_colors[tmp_col][tmp_gene]
                 except KeyError:
                     # Don't draw ribbon if the source gene is not drawn.
                     continue
 
             for col_name in target_pair:
                 gene = ribbon[col_name]
-                if _is_na(gene):
+                if pd.isna(gene):
                     gene = "none"
                 if gene not in tmp_gene_tops[col_name]:
                     gene = "other"

src/scirpy/tl/_chain_qc.py

@@ -7,7 +7,7 @@
 from scanpy import logging
 
 from scirpy import get
-from scirpy.util import DataHandler, _is_na
+from scirpy.util import DataHandler
 
 
 @DataHandler.inject_param_docs()
@@ -176,10 +176,10 @@ def _chain_pairing(
 
     logging.debug("Done initalizing")
 
-    mask_has_vj1 = ~_is_na(get.airr(params, "junction_aa", "VJ_1").values)
-    mask_has_vdj1 = ~_is_na(get.airr(params, "junction_aa", "VDJ_1").values)
-    mask_has_vj2 = ~_is_na(get.airr(params, "junction_aa", "VJ_2").values)
-    mask_has_vdj2 = ~_is_na(get.airr(params, "junction_aa", "VDJ_2").values)
+    mask_has_vj1 = ~pd.isna(get.airr(params, "junction_aa", "VJ_1").values)
+    mask_has_vdj1 = ~pd.isna(get.airr(params, "junction_aa", "VDJ_1").values)
+    mask_has_vj2 = ~pd.isna(get.airr(params, "junction_aa", "VJ_2").values)
+    mask_has_vdj2 = ~pd.isna(get.airr(params, "junction_aa", "VDJ_2").values)
 
     logging.debug("Done with masks")
 

src/scirpy/tl/_clonal_expansion.py

@@ -5,7 +5,7 @@
 import numpy as np
 import pandas as pd
 
-from scirpy.util import DataHandler, _is_na, _normalize_counts
+from scirpy.util import DataHandler, _normalize_counts
 
 
 def _clip_and_count(
@@ -48,7 +48,7 @@ def _get_interval(value: int) -> str:
         .assign(tmp_count=lambda X: pd.Categorical(_get_interval(X["tmp_count"].values), categories=categories))
     )
     clipped_count = obs.merge(clonotype_counts, how="left", on=groupby_cols)["tmp_count"]
-    clipped_count[_is_na(obs[target_col])] = "nan"
+    clipped_count[pd.isna(obs[target_col])] = "nan"
     clipped_count.index = obs.index
 
     if inplace:
@@ -170,7 +170,7 @@ def summarize_clonal_expansion(
     obs[tmp_col] = expansion
 
     # filter NA values
-    obs = obs.loc[~_is_na(obs[target_col]), :]
+    obs = obs.loc[~pd.isna(obs[target_col]), :]
 
     if summarize_by == "clone_id":
         obs.drop_duplicates(inplace=True)

src/scirpy/tl/_clonotype_modularity.py

@@ -2,13 +2,14 @@
 from typing import Literal
 
 import numpy as np
+import pandas as pd
 import scipy.sparse
 import scipy.stats
 from mudata import MuData
 from scanpy import logging
 from statsmodels.stats.multitest import fdrcorrection
 
-from scirpy.util import DataHandler, _is_na, tqdm
+from scirpy.util import DataHandler, tqdm
 from scirpy.util._negative_binomial import fit_nbinom
 from scirpy.util.graph import _get_igraph_from_adjacency
 
@@ -107,7 +108,7 @@ def clonotype_modularity(
         n_permutations = 1000 if permutation_test == "approx" else 10000
 
     clonotype_per_cell = params.get_obs(target_col)
-    cells_with_valid_clonotype = clonotype_per_cell[~_is_na(clonotype_per_cell.values)].index
+    cells_with_valid_clonotype = clonotype_per_cell[~pd.isna(clonotype_per_cell.values)].index
     data_subset = params.data[cells_with_valid_clonotype.values, :]
     try:
         connectivities = data_subset.obsp[connectivity_key]

src/scirpy/tl/_diversity.py

@@ -4,7 +4,7 @@
 import numpy as np
 import pandas as pd
 
-from scirpy.util import DataHandler, _is_na
+from scirpy.util import DataHandler
 
 
 def _shannon_entropy(counts: np.ndarray):
@@ -105,7 +105,7 @@ def alpha_diversity(
     """
     params = DataHandler(adata, airr_mod)
     ir_obs = params.get_obs([target_col, groupby])
-    ir_obs = ir_obs.loc[~_is_na(ir_obs[target_col]), :]
+    ir_obs = ir_obs.loc[~pd.isna(ir_obs[target_col]), :]
     clono_counts = ir_obs.groupby([groupby, target_col], observed=True).size().reset_index(name="count")
 
     diversity = {}

src/scirpy/tl/_group_abundance.py

@@ -5,7 +5,7 @@
 import pandas as pd
 
 from scirpy.get import _has_ir
-from scirpy.util import DataHandler, _is_na, _normalize_counts
+from scirpy.util import DataHandler, _normalize_counts
 
 
 def _group_abundance(
@@ -17,7 +17,7 @@ def _group_abundance(
     sort: Literal["count", "alphabetical"] | Sequence[str] = "count",
 ) -> pd.DataFrame:
     # remove NA rows
-    na_mask = _is_na(ir_obs[groupby]) | _is_na(ir_obs[target_col])
+    na_mask = pd.isna(ir_obs[groupby]) | pd.isna(ir_obs[target_col])
     ir_obs = ir_obs.loc[~na_mask, :]
 
     # normalize to fractions

src/scirpy/tl/_ir_query.py

@@ -10,7 +10,7 @@
 
 from scirpy.ir_dist import MetricType, _get_metric_key
 from scirpy.ir_dist._clonotype_neighbors import ClonotypeNeighbors
-from scirpy.util import DataHandler, _is_na, read_cell_indices, tqdm
+from scirpy.util import DataHandler, read_cell_indices, tqdm
 
 from ._clonotypes import _common_doc, _common_doc_parallelism, _doc_clonotype_definition, _validate_parameters
 
@@ -407,7 +407,7 @@ def reduce_fun(x):
 
     # convert nan-equivalents to real nan values.
     for col in df_res:
-        df_res.loc[_is_na(df_res[col]), col] = None
+        df_res.loc[pd.isna(df_res[col]), col] = None
 
     if inplace:
         for col in df_res:

src/scirpy/tl/_repertoire_overlap.py

@@ -5,7 +5,7 @@
 from scipy.cluster import hierarchy as sc_hierarchy
 from scipy.spatial import distance as sc_distance
 
-from scirpy.util import DataHandler, _is_na, _normalize_counts
+from scirpy.util import DataHandler, _normalize_counts
 
 
 @DataHandler.inject_param_docs()
@@ -65,7 +65,7 @@ def repertoire_overlap(
         obs[normalize] = params.get_obs(normalize)
 
     # Remove NA rows
-    na_mask = _is_na(obs[groupby]) | _is_na(obs[target_col])
+    na_mask = pd.isna(obs[groupby]) | pd.isna(obs[target_col])
     df = obs.loc[~na_mask, :].copy()
 
     # Normalize to fractions