specifiable column hiearachy list

xlsx_dir_map/__main__.py

@@ -4,14 +4,17 @@ parser = argparse.ArgumentParser(description='Process some intzzz')
 parser.add_argument('--in', '-i', dest='file_in', required=True)
 parser.add_argument('--out', '-o')
 parser.add_argument('--range', '-r', dest='cell_range')
+parser.add_argument('column_groups', nargs='+')
 
 args = parser.parse_args()
 
 from spreadsheet import get_names_list 
 
-names = get_names_list(args.file_in, args.cell_range)
+names = get_names_list(args.file_in, args.cell_range, args.column_groups)
 print(names)
 
 from directory import create_directories
 
 create_directories(args.out, names)
+
+#print(args.column_groups)

xlsx_dir_map/spreadsheet.py

@@ -41,7 +41,21 @@ def parse_range(cellrange):
 
     return (min_row, min_col, max_row, max_col)
 
-def get_names_list(filename, cellrange):
+def get_value(row, colrange):
+    cols = colrange.split(":")
+    indices = [custom_index(row, lambda cell: cell.column_letter == col) for col in cols]
+    
+    name = " ".join([str(row[index].value) for index in indices])
+    return name
+
+
+def custom_index(array, compare_function):
+    for i, v in enumerate(array):
+        if compare_function(v):
+            return i
+
+        
+def get_names_list(filename, cellrange, column_groups):
     wb = load_workbook(filename = filename)
     names = []
 
@@ -50,8 +64,15 @@ def get_names_list(filename, cellrange):
     if cellrange:
         min_row, min_col, max_row, max_col = parse_range(cellrange)
 
-    for row in wb.active.iter_rows(min_row=min_row, min_col=min_col, max_row=max_row, max_col=max_col, values_only=True):
+    for row in wb.active.iter_rows(min_row=min_row, min_col=min_col, max_row=max_row, max_col=max_col):
-        name = " ".join([str(cell) for cell in row if cell is not None])
+        cells = [cell for cell in row if cell is not None]
-        names.append(name)
+        paths = []
         
-    return [name for name in names if name != ""]
+        for colgroup in column_groups:
+            path = get_value(cells, colgroup) 
+            paths.append(path)
+
+        if "None" not in paths:
+            names.append("/".join(paths))
+        
+    return names