From 131a8dd1bbbd90cbe52b9a5185ca7fd75966e5f9 Mon Sep 17 00:00:00 2001
From: To Huynh <37560480+tohuynh@users.noreply.github.com>
Date: Fri, 20 Aug 2021 10:51:17 -0700
Subject: [PATCH] Allow standard-institution to have sub_category field (#38)

---
 .../google_sheets_institution_extracter.py     | 17 +++++++++--------
 siglatools/institution_extracters/utils.py     | 18 ++++++++++++++++++
 .../tests/institution_extracters/test_utils.py | 17 +++++++++++++++++
 3 files changed, 44 insertions(+), 8 deletions(-)

diff --git a/siglatools/institution_extracters/google_sheets_institution_extracter.py b/siglatools/institution_extracters/google_sheets_institution_extracter.py
index d551545..11a05ba 100644
--- a/siglatools/institution_extracters/google_sheets_institution_extracter.py
+++ b/siglatools/institution_extracters/google_sheets_institution_extracter.py
@@ -16,8 +16,9 @@
     VariableType,
 )
 from . import exceptions
-from .constants import GoogleSheetsFormat as gs_format, MetaDataField
-from .utils import FormattedSheetData, SheetData
+from .constants import GoogleSheetsFormat as gs_format
+from .constants import MetaDataField
+from .utils import FormattedSheetData, SheetData, create_institution_sub_category
 
 ###############################################################################
 
@@ -94,12 +95,9 @@ def _get_multilple_sigla_answer_variable(
             InstitutionField.category: sheet_data.meta_data.get(
                 InstitutionField.category
             ),
-            InstitutionField.sub_category: [
-                " ".join(sub_cat.strip().split())
-                for sub_cat in sheet_data.meta_data.get(InstitutionField.sub_category)
-                .strip()
-                .split(";")
-            ],
+            InstitutionField.sub_category: create_institution_sub_category(
+                sheet_data.meta_data.get(InstitutionField.sub_category)
+            ),
             "childs": [
                 {
                     VariableField.heading: variable_row[0],
@@ -149,6 +147,9 @@ def _get_standard_institution(
             InstitutionField.category: sheet_data.meta_data.get(
                 InstitutionField.category
             ),
+            InstitutionField.sub_category: create_institution_sub_category(
+                sheet_data.meta_data.get(InstitutionField.sub_category)
+            ),
             "childs": [
                 {
                     VariableField.heading: variable_row[0],
diff --git a/siglatools/institution_extracters/utils.py b/siglatools/institution_extracters/utils.py
index b4bb7a6..1328159 100644
--- a/siglatools/institution_extracters/utils.py
+++ b/siglatools/institution_extracters/utils.py
@@ -118,3 +118,21 @@ def convert_col_to_name(col: int) -> str:
         col_num = int((col_num - 1) / 26)
 
     return col_str
+
+
+def create_institution_sub_category(sub_categories: str) -> List[str]:
+    """
+    Create a list of institution sub categories.
+
+    Parameters
+    ----------
+    sub_categories: str
+        The list of sub categories, separated by ;.
+    Returns
+    -------
+    sub_categories_list: List[str]
+        The list of sub categories.
+    """
+    if not sub_categories:
+        return []
+    return [" ".join(sub_cat.strip().split()) for sub_cat in sub_categories.split(";")]
diff --git a/siglatools/tests/institution_extracters/test_utils.py b/siglatools/tests/institution_extracters/test_utils.py
index cd3fe1e..6fc48a4 100644
--- a/siglatools/tests/institution_extracters/test_utils.py
+++ b/siglatools/tests/institution_extracters/test_utils.py
@@ -7,6 +7,7 @@
 from siglatools.institution_extracters.utils import (
     convert_col_to_name,
     convert_rowcol_to_A1_name,
+    create_institution_sub_category,
 )
 
 
@@ -54,3 +55,19 @@ def test_convert_col_to_name(col, expected):
 )
 def test_convert_rowcol_to_A1_name(row, col, expected):
     assert convert_rowcol_to_A1_name(row, col) == expected
+
+
+@pytest.mark.parametrize(
+    "sub_categories, expected",
+    [
+        (None, []),
+        ("", []),
+        ("a", ["a"]),
+        ("a;b", ["a", "b"]),
+        (" a ; b ", ["a", "b"]),
+        ("a; b c", ["a", "b c"]),
+        ("a; b c;d", ["a", "b c", "d"]),
+    ],
+)
+def test_create_institution_sub_category(sub_categories, expected):
+    assert create_institution_sub_category(sub_categories) == expected