add to docstring, add more tests

Author: rxu17

genie_registry/vcf.py

@@ -133,9 +133,7 @@ def _validate(self, vcfdf):
                 )
             if vcfdf["FORMAT"].isnull().values.any():
                 total_error += "vcf: Must not have missing values in FORMAT column.\n"
-        total_error += self.validate_tumor_and_normal_sample_columns_exist(
-            input_df=vcfdf
-        )
+        total_error += self.validate_tumor_and_normal_sample_columns(input_df=vcfdf)
         # Require that they report variants mapped to
         # either GRCh37 or hg19 without
         # the chr-prefix.
@@ -175,37 +173,33 @@ def _validate(self, vcfdf):
         # and output with warnings or errors if the format is not adhered too
         return total_error, warning
 
-    def validate_tumor_and_normal_sample_columns_exist(
-        self, input_df: pd.DataFrame
-    ) -> str:
-        """Validates that the expected tumor sample column and optional normal
-            sample columns are present in the VCF depending on how many
-            columns you have present in the VCF and they have no missing values
-
-            Rules:
-                - VCFs can only have a max of 11 columns including the 9 required columns
-                - For 11 columns VCFs, it is assumed this is a matched tumor normal vcf file
-                    which means there should be a tumor sample and normal sample
-                    column present
-                - For 10 column VCFs, it is assumed this is a single sample vcf file
-                    which means there should be a tumor sample column present
-                - Anything lower than 10 columns is INVALID because you must have at
-                least a tumor sample column on top of the 9 required VCF columns
-
-                - If tumor sample and normal sample columns are present, they must not have
-                any missing values.
-
-            Examples:
-
-            VCF with Matched Tumor Normal columns:
-            | GENIE-GOLD-1-1-tumor | GENIE-GOLD-1-1-normal |
-            | -------------------- | --------------------- |
-            |                      |                       |
-
-            VCF with Single Tumor VCF column:
-            | TUMOR |
-            | ----- |
-            |       |
+    def validate_tumor_and_normal_sample_columns(self, input_df: pd.DataFrame) -> str:
+        """
+        Validates that the expected tumor sample column and optional normal
+        sample columns are present in the VCF depending on how many
+        columns you have present in the VCF and they have no missing values
+
+        Rules:
+            - VCFs can only have a max of 11 columns including the 9 required columns
+            - For 11 columns VCFs, it is assumed this is a matched tumor normal vcf file
+                which means there should be a tumor sample and normal sample
+                column present
+            - For 10 column VCFs, it is assumed this is a single sample vcf file
+                which means there should be a tumor sample column present
+            - Anything lower than 10 columns is INVALID because you must have at
+            least a tumor sample column on top of the 9 required VCF columns
+            - If tumor sample and/or normal sample columns are present, they must not have
+            any missing values.
+
+        Example: VCF with Matched Tumor Normal columns
+            | OTHER_VCF_COLUMNS | GENIE-GOLD-1-1-tumor | GENIE-GOLD-1-1-normal |
+            | ----------------- | -------------------- | --------------------- |
+            |        ...        |         ...          |          ...          |
+
+        Example: VCF with Single Tumor VCF column
+            | OTHER_VCF_COLUMNS | TUMOR |
+            | ----------------- | ----- |
+            |        ...        |  ...  |
 
         Args:
             input_df (pd.DataFrame): input vcf data to be validated

tests/test_vcf.py

@@ -198,6 +198,26 @@ def test_validation_invalid_two_samples_normal(vcf_class):
     assert warning == ""
 
 
+def test_validation_invalid_format_has_nas(vcf_class):
+    vcfDf = pd.DataFrame(
+        {
+            "#CHROM": ["2", "9", "12"],
+            "POS": [69688533, 99401860, 53701241],
+            "ID": ["AAK1", "AAED1", "AAAS"],
+            "REF": ["AANT", "AACG", "AAAN"],
+            "ALT": ["AAK1", "AAED1", "AAAS"],
+            "QUAL": ["AAK1", "AAED1", "AAAS"],
+            "FILTER": ["AAK1", "AAED1", "AAAS"],
+            "INFO": ["AAK1", "AAED1", "AAAS"],
+            "FORMAT": [None, "AAED1", "AAAS"],
+            "TUMOR": ["AAK1", "AAED1", "AAAS"],
+        }
+    )
+    error, warning = vcf_class._validate(vcfDf)
+    assert error == "vcf: Must not have missing values in FORMAT column.\n"
+    assert warning == ""
+
+
 def test_validation_invalid_white_space(vcf_class):
     vcfDf = pd.DataFrame(
         {
@@ -427,11 +447,11 @@ def test_that__get_dataframe_uses_correct_columns_to_replace(
                 "FILTER",
                 "INFO",
                 "FORMAT",
-                "GENIE-SAGE-1-1-normal",
                 "GENIE-SAGE-1-1-tumor",
+                "GENIE-SAGE-1-1-normal",
             ],
             [[1, 2, 3, 4, 5, 6, 7, 8, 9, "a", None]],
-            "vcf: Must not have missing values in GENIE-SAGE-1-1-tumor column.\n",
+            "vcf: Must not have missing values in GENIE-SAGE-1-1-normal column.\n",
         ),
         # Case 7: 11 columns, wrong normal/tumor naming
         (
@@ -480,13 +500,14 @@ def test_that__get_dataframe_uses_correct_columns_to_replace(
         "invalid_tumor_sample_name",
         "tumor_sample_col_has_nas",
         "valid_matched_tumor_normal",
+        "normal_sample_col_has_nas",
         "invalid_normal_sample_name",
         "more_than_11_cols",
     ],
 )
-def test_validate_tumor_and_normal_sample_columns_exist(
+def test_validate_tumor_and_normal_sample_columns(
     columns, data, expected_in_error, vcf_class
 ):
     df = pd.DataFrame(data, columns=columns)
-    result = vcf_class.validate_tumor_and_normal_sample_columns_exist(df)
+    result = vcf_class.validate_tumor_and_normal_sample_columns(df)
     assert expected_in_error == result