From c9fe7b7ad47f4798b8888716531689e13acc974f Mon Sep 17 00:00:00 2001
From: Daniel McDonald <danielmcdonald@ucsd.edu>
Date: Tue, 16 Jan 2024 12:09:47 -0800
Subject: [PATCH 1/2] BUG: read_count_threshold assumes samples in metadata are
 present in table

---
 q2_katharoseq/tests/test_method.py | 34 ++++++++++++++++++++++++++++++
 1 file changed, 34 insertions(+)

diff --git a/q2_katharoseq/tests/test_method.py b/q2_katharoseq/tests/test_method.py
index 17b6891..a548c3e 100644
--- a/q2_katharoseq/tests/test_method.py
+++ b/q2_katharoseq/tests/test_method.py
@@ -58,6 +58,40 @@ def setUp(self):
         folder = '../../example'
         self.fp = join(dirname(abspath(getfile(currentframe()))), folder)
 
+    def test_read_count_threshold_works_with_metadata_superset(self):
+        # make sure we work when a katharoseq control isn't present in the
+        # table
+        pos_control_col = self.positive_control_column.to_series().copy()
+        pos_control_col_name = pos_control_col.name
+        pos_control_col_idxname = pos_control_col.index.name
+        pos_control_col = pd.concat([pos_control_col,
+                                    pd.Series(['a', ], index=['foo', ])])
+        pos_control_col.name = pos_control_col_name
+        pos_control_col.index.name = pos_control_col_idxname
+        pos_control_col = CategoricalMetadataColumn(pos_control_col)
+
+        cell_count_col = self.cell_count_column.to_series().copy()
+        cell_count_col_name = cell_count_col.name
+        cell_count_col_idxname = cell_count_col.index.name
+        cell_count_col = pd.concat([cell_count_col,
+                                    pd.Series([1000000, ], index=['foo', ])])
+        cell_count_col.name = cell_count_col_name
+        cell_count_col.index.name = cell_count_col_idxname
+        cell_count_col = NumericMetadataColumn(cell_count_col)
+
+        with tempfile.TemporaryDirectory() as output_dir:
+            read_count_threshold(
+                output_dir,
+                self.threshold,
+                self.positive_control_value,
+                pos_control_col,
+                cell_count_col,
+                self.table,
+                self.control)
+
+            index_fp = os.path.join(output_dir, 'index.html')
+            self.assertTrue(os.path.exists(index_fp))
+
     def test_outputs_index(self):
         with tempfile.TemporaryDirectory() as output_dir:
             read_count_threshold(

From bba25a2c5699065916c23cb3f84d5493249910eb Mon Sep 17 00:00:00 2001
From: Daniel McDonald <danielmcdonald@ucsd.edu>
Date: Tue, 16 Jan 2024 12:21:04 -0800
Subject: [PATCH 2/2] BUG: allow metadata to be a superset of the table

---
 q2_katharoseq/_methods.py          | 23 ++++++++++++++---------
 q2_katharoseq/tests/test_method.py | 17 +++--------------
 2 files changed, 17 insertions(+), 23 deletions(-)

diff --git a/q2_katharoseq/_methods.py b/q2_katharoseq/_methods.py
index c641c11..4f609e3 100644
--- a/q2_katharoseq/_methods.py
+++ b/q2_katharoseq/_methods.py
@@ -109,22 +109,30 @@ def read_count_threshold(
     positive_control_column = positive_control_column.to_series()
     cell_count_column = cell_count_column.to_series()
 
-    # # FILTER COLUMNS
+    if set(positive_control_column.index) != set(cell_count_column.index):
+        raise ValueError("Columns have different samples")
+
+    md_samples = set(positive_control_column.index)
+    tab_samples = set(table.index)
+    if not md_samples.issuperset(tab_samples):
+        raise ValueError("Table contains samples not in metadata")
+
+    # overlap to what's present in the feature table
+    positive_control_column = positive_control_column.loc[list(tab_samples)]
+    cell_count_column = cell_count_column.loc[list(tab_samples)]
+
+    # FILTER COLUMNS
     positive_controls = positive_control_column[
         positive_control_column == positive_control_value]
 
     if not positive_controls.shape[0]:
         raise ValueError('No positive controls found in ' +
-                         'positive control column.')
+                         'positive control column which are in the table.')
     positive_controls = pd.Series(positive_controls)
 
     cell_counts = cell_count_column.loc[positive_controls.index]
 
-    # # CHECK SHAPES
     inds = positive_controls.index.intersection(table.index)
-    print(inds)
-    if len(inds) == 0:
-        raise KeyError('No positive controls found in table.')
     table_positive = table.loc[inds]
 
     if threshold > 100 or threshold < 0:
@@ -135,9 +143,6 @@ def read_count_threshold(
     # READS IN HIGHEST INPUT SAMPLE
     max_cell_counts = cell_counts.idxmax()
 
-    if max_cell_counts not in df.index.values:
-        raise KeyError('No positive controls found in table.')
-
     max_input = df.loc[max_cell_counts]
     max_inputT = max_input.T
     max_inputT = max_inputT.sort_values(ascending=False).head(10)
diff --git a/q2_katharoseq/tests/test_method.py b/q2_katharoseq/tests/test_method.py
index a548c3e..dfb581d 100644
--- a/q2_katharoseq/tests/test_method.py
+++ b/q2_katharoseq/tests/test_method.py
@@ -163,23 +163,12 @@ def test_no_positive_controls_in_col(self):
                 self.control)
 
     def test_no_positive_controls_in_table(self):
-
-        ind = pd.Index(
-                ['s5', 's6', 's7', 's8'],
-                name='sampleid')
-        table = pd.DataFrame(
-            [[0, 1, 2, 3],
-             [0, 1, 2, 3],
-             [5, 4, 3, 2],
-             [7, 2, 3, 4]],
-            index=ind,  # change index
-            columns=['f1', 'f2', 'f3', 'f4'])
+        table = self.table.loc[['s2', 's4']]
 
         with tempfile.TemporaryDirectory() as output_dir, \
             self.assertRaisesRegex(
-                KeyError,
-                'No positive controls found '
-                'in table.'):
+                ValueError,
+                'No positive controls found'):
 
             read_count_threshold(
                 output_dir,