From 4df61fd6e5f9c3d31ddd42204c03fcce7076132a Mon Sep 17 00:00:00 2001
From: rxu17 <26471741+rxu17@users.noreply.github.com>
Date: Thu, 2 Nov 2023 18:03:24 -0700
Subject: [PATCH] add to be class attributes

---
 genie_registry/maf.py | 21 +++++++++++++--------
 genie_registry/vcf.py | 15 +++++++--------
 2 files changed, 20 insertions(+), 16 deletions(-)

diff --git a/genie_registry/maf.py b/genie_registry/maf.py
index 4575d189..bb1f2cd1 100644
--- a/genie_registry/maf.py
+++ b/genie_registry/maf.py
@@ -70,6 +70,8 @@ class maf(FileTypeFormat):
     _fileType = "maf"
 
     _process_kwargs = []
+    _allele_cols = ["REFERENCE_ALLELE", "TUMOR_SEQ_ALLELE1", "TUMOR_SEQ_ALLELE2"]
+    _allowed_alleles = ["A", "T", "C", "G", "N", " ", "-"]
 
     def _validateFilename(self, filePath):
         """
@@ -294,21 +296,24 @@ def _validate(self, mutationDF):
             )
             total_error.write(errors)
             warning.write(warnings)
-            
-        # TODO: add these lists as class attribute or global
-        allele_cols = ["REFERENCE_ALLELE", "TUMOR_SEQ_ALLELE1", "TUMOR_SEQ_ALLELE2"]
-        allowed_alleles = ['A','T','C','G','N', ' ', '-']
-        for allele_col in allele_cols:
+
+        for allele_col in self._allele_cols:
             if process_functions.checkColExist(mutationDF, allele_col):
                 invalid_indices = validate.get_invalid_allele_rows(
-                    mutationDF, allele_col, allowed_alleles = allowed_alleles, ignore_case = True
+                    mutationDF,
+                    allele_col,
+                    allowed_alleles=self._allowed_alleles,
+                    ignore_case=True,
                 )
                 errors, warnings = validate.get_allele_validation_message(
-                    invalid_indices, invalid_col = allele_col, allowed_alleles = allowed_alleles, fileformat="maf"
+                    invalid_indices,
+                    invalid_col=allele_col,
+                    allowed_alleles=self._allowed_alleles,
+                    fileformat=self._fileType,
                 )
                 total_error.write(errors)
                 warning.write(warnings)
-                
+
         return total_error.getvalue(), warning.getvalue()
 
     def _cross_validate(self, mutationDF: pd.DataFrame) -> tuple:
diff --git a/genie_registry/vcf.py b/genie_registry/vcf.py
index 75525bd2..78406d92 100644
--- a/genie_registry/vcf.py
+++ b/genie_registry/vcf.py
@@ -18,6 +18,8 @@ class vcf(FileTypeFormat):
     _fileType = "vcf"
 
     _process_kwargs = []
+    _allele_col = "REF"
+    _allowed_alleles = ["A", "T", "C", "G", "N"]
 
     def _validateFilename(self, filePath):
         basename = os.path.basename(filePath[0])
@@ -137,18 +139,15 @@ def _validate(self, vcfdf):
         total_error += error
         warning += warn
 
-        # TODO: add this as class attribute or global
-        allele_col = "REF"
-        allowed_alleles = ["A", "T", "C", "G", "N"]
-        if process_functions.checkColExist(vcfdf, allele_col):
+        if process_functions.checkColExist(vcfdf, self._allele_col):
             invalid_indices = validate.get_invalid_allele_rows(
-                vcfdf, allele_col, allowed_alleles=allowed_alleles, ignore_case=True
+                vcfdf, self._allele_col, allowed_alleles=self._allowed_alleles, ignore_case=True
             )
             errors, warnings = validate.get_allele_validation_message(
                 invalid_indices,
-                invalid_col=allele_col,
-                allowed_alleles=allowed_alleles,
-                fileformat="vcf",
+                invalid_col=self._allele_col,
+                allowed_alleles=self._allowed_alleles,
+                fileformat=self._fileType,
             )
             total_error += errors
             warning += warnings