fix(linker): Allow citations like Parshat Vayikra 2:3

sefaria/model/linker/referenceable_book_node.py

@@ -9,10 +9,31 @@
 
 def subref(ref: text.Ref, section: int):
     if ref.index_node.addressTypes[len(ref.sections)-1] == "Talmud":
-        d = ref._core_dict()
-        d['sections'][-1] += (section-1)
-        d['toSections'] = d['sections'][:]
-        return text.Ref(_obj=d)
+        return _talmud_subref(ref, section)
+    elif ref.index.categories == ['Tanakh', 'Torah']:
+        return _parsha_subref(ref, section)
+    else:
+        return ref.subref(section)
+
+
+def _talmud_subref(ref: text.Ref, section: int):
+    d = ref._core_dict()
+    d['sections'][-1] += (section-1)
+    d['toSections'] = d['sections'][:]
+    return text.Ref(_obj=d)
+
+
+def _parsha_subref(ref: text.Ref, section: int):
+    parsha_trefs = {n.wholeRef for n in ref.index.get_alt_struct_leaves()}
+    if ref.normal() in parsha_trefs:
+        book_subref = text.Ref(ref.index.title).subref(section)
+        if ref.contains(book_subref):
+            return book_subref
+        else:
+            # section doesn't fall within parsha
+            # Note, only validates that perek is in parsha range, doesn't check segment level.
+            # Edge case is Parshat Noach 6:3
+            raise InputError
     else:
         return ref.subref(section)
 

sefaria/model/linker/tests/linker_test.py

@@ -29,7 +29,6 @@ def test_resolved_raw_ref_clone():
 
 
 @pytest.mark.parametrize(('resolver_data', 'expected_trefs'), [
-    [crrd(["@Parshat Vayikra", "#2", "#3"], lang='en'), ('Leviticus 2:3',)],
     # Numbered JAs
     [crrd(["@Jerusalem", "@Talmud", "@Yoma", "#5a"], lang='en'), ("Jerusalem Talmud Yoma 1:1:20-25",)],
     [crrd(["@Babylonian", "@Talmud", "@Sukkah", "#49b"], lang='en'), ("Sukkah 49b",)],
@@ -51,6 +50,11 @@ def test_resolved_raw_ref_clone():
     [crrd(['@שבת', '#א', '#ב']), ["Mishnah Shabbat 1:2"]],  # shouldn't match Shabbat 2a by reversing order of parts
     [crrd(['@שבת', '#ב', '#א']), ["Shabbat 2a", "Mishnah Shabbat 2:1"]],  # ambiguous case
 
+    # Parsha -> sections
+    [crrd(["@Parshat Vayikra", "#2", "#3"], lang='en'), ('Leviticus 2:3',)],
+    [crrd(["@Parshat Tzav", "#2", "#3"], lang='en'), tuple()],  # validate that sections fall within parsha
+    pytest.param(crrd(["@Parshat Noach", "#6", "#3"], lang='en'), tuple(), marks=pytest.mark.xfail(reason="currently dont check if pasuk/perek pair fall in parsha, only perek")),
+
     # Aliases for perakim
     [crrd(["@משנה", "@ברכות", "#פרק קמא"]), ("Mishnah Berakhot 1",)],
     [crrd(["@משנה", "@ברכות", "#פרק בתרא"]), ("Mishnah Berakhot 9",)],