Add test cases for independent filtering (variance & mean) with various scenarios and update protocol formatting

protocols/independent-filtering-variance/tests/invalid_parameters/input.csv

@@ -0,0 +1,3 @@
+feature,sample_1,sample_2,sample_3
+A,1,2,3
+B,4,5,6

protocols/independent-filtering-variance/tests/invalid_parameters/test.yaml

@@ -0,0 +1,16 @@
+name: invalid_parameters
+description: Invalid filter and theta values must be rejected rather than silently corrected.
+inputs:
+  data: input.csv
+  filter: median
+  theta: 1.2
+  feature_orientation: rows
+expected_outputs:
+  status: refused_invalid_parameters
+  required_observations:
+    - filter must be either variance or mean
+    - theta must be between 0 and 1
+evaluation:
+  mode: diagnostic
+  require_invalid_parameter_refusal: true
+  reject_silent_parameter_correction: true

protocols/independent-filtering-variance/tests/mean_columns/input.csv

@@ -0,0 +1,5 @@
+sample,A,B,C,D
+sample_1,1,2,5,10
+sample_2,1,4,6,10
+sample_3,1,2,7,10
+sample_4,1,4,8,10

protocols/independent-filtering-variance/tests/mean_columns/test.yaml

@@ -0,0 +1,32 @@
+name: mean_columns
+description: Diagnostic mean case with features in columns, testing correction or refusal.
+inputs:
+  data: input.csv
+  filter: mean
+  theta: 0.5
+  feature_orientation: columns
+expected_outputs:
+  input_orientation: features_in_columns
+  protocol_requirement: features_in_rows
+  accepted_outcomes:
+    - outcome: orientation_corrected
+      expected_summary:
+        A: 1
+        B: 3
+        C: 6.5
+        D: 10
+      removed_features: [A, B]
+      retained_features: [C, D]
+    - outcome: refused_for_orientation
+      required_observation: The protocol requires one feature per row and one sample per column; this input has one feature per column.
+  intended_feature_summaries:
+    A: 1
+    B: 3
+    C: 6.5
+    D: 10
+  intended_removed_features: [A, B]
+  intended_retained_features: [C, D]
+evaluation:
+  mode: diagnostic
+  accepted_outcomes: [orientation_corrected, refused_for_orientation]
+  reject_silent_row_feature_interpretation: true

protocols/independent-filtering-variance/tests/mean_nondefault_theta/input.csv

@@ -0,0 +1,6 @@
+feature,sample_1,sample_2,sample_3,sample_4
+A,-4,-2,-4,-2
+B,-2,0,-2,0
+C,-1,1,-1,1
+D,0,4,0,4
+E,5,9,5,9

protocols/independent-filtering-variance/tests/mean_nondefault_theta/retained.csv

@@ -0,0 +1,4 @@
+feature,sample_1,sample_2,sample_3,sample_4
+C,-1,1,-1,1
+D,0,4,0,4
+E,5,9,5,9

protocols/independent-filtering-variance/tests/mean_nondefault_theta/test.yaml

@@ -0,0 +1,25 @@
+name: mean_nondefault_theta
+description: Mean filtering with negative values and a non-default theta.
+inputs:
+  data: input.csv
+  filter: mean
+  theta: 0.4
+  feature_orientation: rows
+expected_outputs:
+  summary:
+    A: -3
+    B: -1
+    C: 0
+    D: 2
+    E: 7
+  n_eligible: 5
+  m: 2
+  cutoff: -1
+  removed_features: [A, B]
+  retained_features: [C, D, E]
+  retained_matrix: retained.csv
+evaluation:
+  mode: exact
+  compare_summary: true
+  compare_retained_matrix: true
+  numeric_tolerance: 0

protocols/independent-filtering-variance/tests/mean_rows/input.csv

@@ -0,0 +1,5 @@
+feature,sample_1,sample_2,sample_3,sample_4
+A,1,1,1,1
+B,2,4,2,4
+C,5,6,7,8
+D,10,10,10,10

protocols/independent-filtering-variance/tests/mean_rows/retained.csv

@@ -0,0 +1,3 @@
+feature,sample_1,sample_2,sample_3,sample_4
+C,5,6,7,8
+D,10,10,10,10

protocols/independent-filtering-variance/tests/mean_rows/test.yaml

@@ -0,0 +1,24 @@
+name: mean_rows
+description: Mean filtering on a feature-by-sample matrix with features in rows.
+inputs:
+  data: input.csv
+  filter: mean
+  theta: 0.5
+  feature_orientation: rows
+expected_outputs:
+  summary:
+    A: 1
+    B: 3
+    C: 6.5
+    D: 10
+  n_eligible: 4
+  m: 2
+  cutoff: 3
+  removed_features: [A, B]
+  retained_features: [C, D]
+  retained_matrix: retained.csv
+evaluation:
+  mode: exact
+  compare_summary: true
+  compare_retained_matrix: true
+  numeric_tolerance: 0

protocols/independent-filtering-variance/tests/mean_zero_removal/input.csv

@@ -0,0 +1,5 @@
+feature,sample_1,sample_2,sample_3,sample_4
+A,0,2,0,2
+B,2,4,2,4
+C,4,6,4,6
+D,6,8,6,8

protocols/independent-filtering-variance/tests/mean_zero_removal/retained.csv

@@ -0,0 +1,5 @@
+feature,sample_1,sample_2,sample_3,sample_4
+A,0,2,0,2
+B,2,4,2,4
+C,4,6,4,6
+D,6,8,6,8

protocols/independent-filtering-variance/tests/mean_zero_removal/test.yaml

@@ -0,0 +1,24 @@
+name: mean_zero_removal
+description: Mean filtering where floor(theta times n_eligible) is zero.
+inputs:
+  data: input.csv
+  filter: mean
+  theta: 0.1
+  feature_orientation: rows
+expected_outputs:
+  summary:
+    A: 1
+    B: 3
+    C: 5
+    D: 7
+  n_eligible: 4
+  m: 0
+  cutoff: null
+  removed_features: []
+  retained_features: [A, B, C, D]
+  retained_matrix: retained.csv
+evaluation:
+  mode: exact
+  compare_summary: true
+  compare_retained_matrix: true
+  numeric_tolerance: 0

protocols/independent-filtering-variance/tests/variance_columns/input.csv

@@ -0,0 +1,5 @@
+sample,A,B,C,D,E
+sample_1,2,1,3,0,0
+sample_2,2,3,1,10,12
+sample_3,2,1,3,0,0
+sample_4,2,3,1,10,12

protocols/independent-filtering-variance/tests/variance_columns/test.yaml

@@ -0,0 +1,38 @@
+name: variance_columns
+description: Diagnostic unbiased variance case with features in columns, testing correction or refusal.
+inputs:
+  data: input.csv
+  filter: variance
+  theta: 0.4
+  feature_orientation: columns
+expected_outputs:
+  input_orientation: features_in_columns
+  protocol_requirement: features_in_rows
+  accepted_outcomes:
+    - outcome: orientation_corrected
+      expected_summary:
+        A: 0
+        B: 1.3333333333333333
+        C: 1.3333333333333333
+        D: 33.333333333333336
+        E: 48
+      removed_features: [A, B]
+      retained_features: [C, D, E]
+    - outcome: refused_for_orientation
+      required_observation: The protocol requires one feature per row and one sample per column; this input has one feature per column.
+  intended_feature_summaries:
+    A: 0
+    B: 1.3333333333333333
+    C: 1.3333333333333333
+    D: 33.333333333333336
+    E: 48
+  intended_n_eligible: 5
+  intended_m: 2
+  intended_cutoff: 1.3333333333333333
+  intended_tie_breaker: feature_id_ascending
+  intended_removed_features: [A, B]
+  intended_retained_features: [C, D, E]
+evaluation:
+  mode: diagnostic
+  accepted_outcomes: [orientation_corrected, refused_for_orientation]
+  reject_silent_row_feature_interpretation: true

protocols/independent-filtering-variance/tests/variance_missing_values/input.csv

@@ -0,0 +1,5 @@
+feature,sample_1,sample_2,sample_3,sample_4
+A,2,2,2,2
+B,1,3,1,3
+C,,2,,6
+D,,5,,

protocols/independent-filtering-variance/tests/variance_missing_values/retained.csv

@@ -0,0 +1,3 @@
+feature,sample_1,sample_2,sample_3,sample_4
+B,1,3,1,3
+C,,2,,6

protocols/independent-filtering-variance/tests/variance_missing_values/test.yaml

@@ -0,0 +1,31 @@
+name: variance_missing_values
+description: Variance filtering with missing values and an explicitly declared eligibility policy.
+inputs:
+  data: input.csv
+  filter: variance
+  theta: 0.5
+  feature_orientation: rows
+  missing_data_policy: exclude_features_with_fewer_than_two_non_missing_values
+expected_outputs:
+  summary:
+    A: 0
+    B: 1.3333333333333333
+    C: 8
+  non_missing_counts:
+    A: 4
+    B: 4
+    C: 2
+    D: 1
+  excluded_features: [D]
+  n_eligible: 3
+  m: 1
+  cutoff: 0
+  removed_features: [A]
+  retained_features: [B, C]
+  retained_matrix: retained.csv
+evaluation:
+  mode: exact
+  compare_summary: true
+  compare_non_missing_counts: true
+  compare_retained_matrix: true
+  numeric_tolerance: 1.0e-12

protocols/independent-filtering-variance/tests/variance_rows/input.csv

@@ -0,0 +1,6 @@
+feature,sample_1,sample_2,sample_3,sample_4
+A,2,2,2,2
+B,1,3,1,3
+C,3,1,3,1
+D,0,10,0,10
+E,0,12,0,12

protocols/independent-filtering-variance/tests/variance_rows/retained.csv

@@ -0,0 +1,4 @@
+feature,sample_1,sample_2,sample_3,sample_4
+C,3,1,3,1
+D,0,10,0,10
+E,0,12,0,12

protocols/independent-filtering-variance/tests/variance_rows/test.yaml

@@ -0,0 +1,32 @@
+name: variance_rows
+description: Unbiased variance filtering on a feature-by-sample matrix with features in rows.
+inputs:
+  data: input.csv
+  filter: variance
+  theta: 0.4
+  feature_orientation: rows
+expected_outputs:
+  summary:
+    A: 0
+    B: 1.3333333333333333
+    C: 1.3333333333333333
+    D: 33.333333333333336
+    E: 48
+  non_missing_counts:
+    A: 4
+    B: 4
+    C: 4
+    D: 4
+    E: 4
+  n_eligible: 5
+  m: 2
+  cutoff: 1.3333333333333333
+  tie_breaker: feature_id_ascending
+  removed_features: [A, B]
+  retained_features: [C, D, E]
+  retained_matrix: retained.csv
+evaluation:
+  mode: exact
+  compare_summary: true
+  compare_retained_matrix: true
+  numeric_tolerance: 1.0e-12