fix row skipping during file opening

databio · Sep 17, 2024 · ccf0c5c · ccf0c5c
1 parent e8be1fe
commit ccf0c5c
Show file tree

Hide file tree

Showing 2 changed files with 29 additions and 2 deletions.
diff --git a/bedboss/refgenome_validator/utils.py b/bedboss/refgenome_validator/utils.py
@@ -41,11 +41,32 @@ def _read_file_pd(*args, **kwargs) -> pd.DataFrame:
                     f"Skipped {row_count} rows while standardization. File: '{args}'"
                 )
             df = df.dropna(axis=1)
-            return df
+            for index, row in df.iterrows():
+                if (
+                    isinstance(row[0], str)
+                    and isinstance(row[1], int)
+                    and isinstance(row[2], int)
+                ):
+                    return df
+                else:
+                    if isinstance(row[1], str):
+                        try:
+                            _ = int(row[1])
+                            df[1] = pd.to_numeric(df[1])
+                        except Exception:
+                            row_count += 1
+                            break
+                    if isinstance(row[2], str):
+                        try:
+                            _ = int(row[2])
+                            df[2] = pd.to_numeric(df[2])
+                        except Exception:
+                            row_count += 1
+                            break
+                    return df
         except (pd.errors.ParserError, pd.errors.EmptyDataError) as _:
             if row_count <= max_rows:
                 row_count += 1
-        # if can't open file after 5 attempts try to open it with gzip
     return _read_gzipped_file(*args)
 
 

diff --git a/scripts/ref_genome_validating/validate_genome.py b/scripts/ref_genome_validating/validate_genome.py
@@ -109,6 +109,11 @@ def main():
         chrom_sizes_file="/home/drc/GITHUB/bedboss/bedboss/bedboss/refgenome_validator/chrom_sizes/ucsc_mm10.chrom.sizes",
     )
 
+    ucsc_mm39 = GenomeModel(
+        genome_alias="ucsc_mm10",
+        chrom_sizes_file="/home/drc/GITHUB/bedboss/bedboss/bedboss/refgenome_validator/chrom_sizes/ucsc_mm39.chrom.sizes",
+    )
+
     ucsc_pantro6 = GenomeModel(
         genome_alias="ucsc_pantro6",
         chrom_sizes_file="/home/drc/GITHUB/bedboss/bedboss/bedboss/refgenome_validator/chrom_sizes/ucsc_panTro6.chrom.sizes",
@@ -117,6 +122,7 @@ def main():
     all_genome_models.append(ucsc_hg38)
     all_genome_models.append(ncbi_hg38)
     all_genome_models.append(ucsc_mm10)
+    all_genome_models.append(ucsc_mm39)
     all_genome_models.append(ucsc_hg19)
     all_genome_models.append(ensembl_hg38)
     all_genome_models.append(ucsc_pantro6)