Ensure fixed fields are correct and avoid bad public notes

One more iteration based on the UWin team's set of feedback on the
hopefully penultimate processing round for the CRKN records.

Signed-off-by: Dan Scott <dscott@laurentian.ca>

diff --git a/tools/ebooks/prep_ebook_records.py b/tools/ebooks/prep_ebook_records.py
index 242eea2..80949de 100644 (file)
--- a/tools/ebooks/prep_ebook_records.py
+++ b/tools/ebooks/prep_ebook_records.py
@@ -334,6 +334,12 @@ def process_fields(record, options):
         # Strip out 300 fields that only contain placeholders
         elif field.tag == '300' and field['a'] == 'p. cm.':
             pass
+        elif field.tag == '008' and field.value()[23] != 's':
+            fixed_field = pymarc.Field(
+                tag='008',
+                data=field.value()[:22] + 's' + field.value()[24:]
+            )
+            new_record.add_field(fixed_field)
         else:
             new_record.add_field(field)
 
@@ -867,7 +873,10 @@ def get_subfields(field, data):
 
     # Check for a $z as the first 856; in Springer records, at least, this
     # indicates a multi-volume set that requires keeping the $z around
-    if field.subfields[0] == 'z':
+    if field.subfields[0] == 'z' and (
+        # However, we don't want to keep garbage-y public notes
+        not field.get_subfields('z')[0].startswith('Connect to MyiLibrary')
+    ):
         subs.extend([field.subfields[0], field.subfields[1]])
 
     subs.extend([