From 2d2f83e9e02cfbe76be45cd09003d4585349d1de Mon Sep 17 00:00:00 2001
From: jonavellecuerdo <cuerdojonavelle@gmail.com>
Date: Thu, 25 Jul 2024 17:36:06 -0400
Subject: [PATCH] notes

---
 tests/sources/xml/test_marc.py     |  89 +++++++++++++++
 transmogrifier/sources/xml/marc.py | 168 +++++++++++++++--------------
 2 files changed, 178 insertions(+), 79 deletions(-)

diff --git a/tests/sources/xml/test_marc.py b/tests/sources/xml/test_marc.py
index cef9d43..cc5686c 100644
--- a/tests/sources/xml/test_marc.py
+++ b/tests/sources/xml/test_marc.py
@@ -1461,6 +1461,95 @@ def test_marc_get_locations_transforms_correctly_if_fields_blank():
     assert Marc.get_locations(source_record) is None
 
 
+def test_get_notes_success():
+    source_record = create_marc_source_record_stub(
+        datafield_insert=(
+            """
+            <datafield tag="245" ind1="0" ind2="0">
+                <subfield code="c">arranged by the Arts Council of Great Britain.</subfield>
+            </datafield>
+            <datafield tag="500" ind1=" " ind2=" ">
+                <subfield code="a">Opera in 5 acts.</subfield>
+            </datafield>
+            <datafield tag="502" ind1=" " ind2=" ">
+                <subfield code="a">Thesis (D.SC.)--University of London.</subfield>
+            </datafield>
+            <datafield tag="504" ind1=" " ind2=" ">
+                <subfield code="a">Includes bibliographical references and index.</subfield>
+            </datafield>
+            <datafield tag="508" ind1=" " ind2=" ">
+                <subfield code="a">Producer, Toygun Kirali.</subfield>
+            </datafield>
+            <datafield tag="511" ind1="0" ind2=" ">
+                <subfield code="a">Lamoureux Concerts Orchestra ; Igor Markevitch, conductor.</subfield>
+            </datafield>
+            <datafield tag="515" ind1=" " ind2=" ">
+                <subfield code="a">Suspended publication 1944-52.</subfield>
+            </datafield>
+            <datafield tag="522" ind1=" " ind2=" ">
+                <subfield code="a">Canada.</subfield>
+            </datafield>
+            <datafield tag="533" ind1=" " ind2=" ">
+                <subfield code="a">Electronic reproduction.</subfield>
+                <subfield code="b">New York :</subfield>
+                <subfield code="c">Springer,</subfield>
+                <subfield code="d">2008.</subfield>
+            </datafield>
+            <datafield tag="534" ind1=" " ind2=" ">
+                <subfield code="p">Originally published</subfield>
+                <subfield code="c">New York : Garland, 1987.</subfield>
+            </datafield>
+            <datafield tag="588" ind1="0" ind2=" ">
+                <subfield code="a">Hard copy version record.</subfield>
+            </datafield>
+            <datafield tag="590" ind1=" " ind2=" ">
+                <subfield code="a">Rare Book copy: Advance copy notice inserted.</subfield>
+            </datafield>
+            """
+        )
+    )
+    assert Marc.get_notes(source_record) == [
+        timdex.Note(
+            value=["arranged by the Arts Council of Great Britain"],
+            kind="Title Statement of Responsibility",
+        ),
+        timdex.Note(value=["Opera in 5 acts"], kind="General Note"),
+        timdex.Note(
+            value=["Thesis (D.SC.)--University of London"], kind="Dissertation Note"
+        ),
+        timdex.Note(
+            value=["Includes bibliographical references and index"],
+            kind="Bibliography Note",
+        ),
+        timdex.Note(
+            value=["Producer, Toygun Kirali"], kind="Creation/Production Credits Note"
+        ),
+        timdex.Note(
+            value=["Lamoureux Concerts Orchestra ; Igor Markevitch, conductor"],
+            kind="Participant or Performer Note",
+        ),
+        timdex.Note(
+            value=["Suspended publication 1944-52"], kind="Numbering Peculiarities Note"
+        ),
+        timdex.Note(value=["Canada"], kind="Geographic Coverage Note"),
+        timdex.Note(
+            value=["Electronic reproduction. New York : Springer, 2008"],
+            kind="Reproduction Note",
+        ),
+        timdex.Note(
+            value=["Originally published New York : Garland, 1987"],
+            kind="Original Version Note",
+        ),
+        timdex.Note(
+            value=["Hard copy version record"],
+            kind="Source of Description Note",
+        ),
+        timdex.Note(
+            value=["Rare Book copy: Advance copy notice inserted"], kind="Local Note"
+        ),
+    ]
+
+
 def test_marc_record_missing_leader_skips_record(caplog):
     marc_xml_records = Marc.parse_source_file(
         "tests/fixtures/marc/marc_record_missing_leader.xml"
diff --git a/transmogrifier/sources/xml/marc.py b/transmogrifier/sources/xml/marc.py
index 70696f1..d6cd72f 100644
--- a/transmogrifier/sources/xml/marc.py
+++ b/transmogrifier/sources/xml/marc.py
@@ -90,85 +90,7 @@ def get_optional_fields(self, source_record: Tag) -> dict | None:
         fields["locations"] = self.get_locations(source_record)
 
         # notes
-        note_marc_fields = [
-            {
-                "tag": "245",
-                "subfields": "c",
-                "kind": "Title Statement of Responsibility",
-            },
-            {
-                "tag": "500",
-                "subfields": "a",
-                "kind": "General Note",
-            },
-            {
-                "tag": "502",
-                "subfields": "abcdg",
-                "kind": "Dissertation Note",
-            },
-            {
-                "tag": "504",
-                "subfields": "a",
-                "kind": "Bibliography Note",
-            },
-            {
-                "tag": "508",
-                "subfields": "a",
-                "kind": "Creation/Production Credits Note",
-            },
-            {
-                "tag": "511",
-                "subfields": "a",
-                "kind": "Participant or Performer Note",
-            },
-            {
-                "tag": "515",
-                "subfields": "a",
-                "kind": "Numbering Peculiarities Note",
-            },
-            {
-                "tag": "522",
-                "subfields": "a",
-                "kind": "Geographic Coverage Note",
-            },
-            {
-                "tag": "533",
-                "subfields": "abcdefmn",
-                "kind": "Reproduction Note",
-            },
-            {
-                "tag": "534",
-                "subfields": "abcefklmnoptxz",
-                "kind": "Original Version Note",
-            },
-            {
-                "tag": "588",
-                "subfields": "a",
-                "kind": "Source of Description Note",
-            },
-            {
-                "tag": "590",
-                "subfields": "a",
-                "kind": "Local Note",
-            },
-        ]
-        for note_marc_field in note_marc_fields:
-            for datafield in source_record.find_all(
-                "datafield", tag=note_marc_field["tag"]
-            ):
-                if note_value := (
-                    self.create_subfield_value_string_from_datafield(
-                        datafield,
-                        note_marc_field["subfields"],
-                        " ",
-                    )
-                ):
-                    fields.setdefault("notes", []).append(
-                        timdex.Note(
-                            value=[note_value.rstrip(" .")],
-                            kind=note_marc_field["kind"],
-                        )
-                    )
+        fields["notes"] = self.get_notes(source_record)
 
         # numbering
 
@@ -944,6 +866,94 @@ def get_locations(cls, source_record: Tag) -> list[timdex.Location] | None:
             )
         return locations or None
 
+    @classmethod
+    def get_notes(cls, source_record: Tag) -> list[timdex.Note] | None:
+        notes = []
+        note_marc_fields = [
+            {
+                "tag": "245",
+                "subfields": "c",
+                "kind": "Title Statement of Responsibility",
+            },
+            {
+                "tag": "500",
+                "subfields": "a",
+                "kind": "General Note",
+            },
+            {
+                "tag": "502",
+                "subfields": "abcdg",
+                "kind": "Dissertation Note",
+            },
+            {
+                "tag": "504",
+                "subfields": "a",
+                "kind": "Bibliography Note",
+            },
+            {
+                "tag": "508",
+                "subfields": "a",
+                "kind": "Creation/Production Credits Note",
+            },
+            {
+                "tag": "511",
+                "subfields": "a",
+                "kind": "Participant or Performer Note",
+            },
+            {
+                "tag": "515",
+                "subfields": "a",
+                "kind": "Numbering Peculiarities Note",
+            },
+            {
+                "tag": "522",
+                "subfields": "a",
+                "kind": "Geographic Coverage Note",
+            },
+            {
+                "tag": "533",
+                "subfields": "abcdefmn",
+                "kind": "Reproduction Note",
+            },
+            {
+                "tag": "534",
+                "subfields": "abcefklmnoptxz",
+                "kind": "Original Version Note",
+            },
+            {
+                "tag": "588",
+                "subfields": "a",
+                "kind": "Source of Description Note",
+            },
+            {
+                "tag": "590",
+                "subfields": "a",
+                "kind": "Local Note",
+            },
+        ]
+        for note_marc_field in note_marc_fields:
+            notes.extend(
+                [
+                    timdex.Note(
+                        value=[note_value.rstrip(" .")],
+                        kind=note_marc_field["kind"],
+                    )
+                    for datafield in source_record.find_all(
+                        "datafield", tag=note_marc_field["tag"]
+                    )
+                    if (
+                        note_value := (
+                            cls.create_subfield_value_string_from_datafield(
+                                datafield,
+                                note_marc_field["subfields"],
+                                " ",
+                            )
+                        )
+                    )
+                ]
+            )
+        return notes or None
+
     @staticmethod
     def get_main_titles(xml: Tag) -> list[str]:
         """