From: Jan Szejko <janek37@gmail.com>
Date: Thu, 27 Apr 2017 17:06:28 +0000 (+0200)
Subject: eisbn csv generator
X-Git-Url: https://git.mdrn.pl/wolnelektury.git/commitdiff_plain/d4d849867aedf29df6cd3d764fa5bd9c0f6e8c76?ds=sidebyside

eisbn csv generator
---

diff --git a/src/catalogue/management/commands/eisbn_csv.py b/src/catalogue/management/commands/eisbn_csv.py
new file mode 100644
index 000000000..ec1333203
--- /dev/null
+++ b/src/catalogue/management/commands/eisbn_csv.py
@@ -0,0 +1,82 @@
+# -*- coding: utf-8 -*-
+# This file is part of Wolnelektury, licensed under GNU Affero GPLv3 or later.
+# Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.
+#
+import csv
+import sys
+from django.core.management.base import BaseCommand
+from django.utils.timezone import localtime
+
+from catalogue.models import Book
+from librarian import RDFNS, DCNS
+
+
+FORMATS = ('HTML', 'PDF', 'TXT', 'EPUB', 'MOBI')
+
+FORMATS_WITH_CHILDREN = ('PDF', 'EPUB', 'MOBI')
+
+
+PRODUCT_FORMS_1 = {
+    'HTML': 'EC',
+    'PDF': 'EB',
+    'TXT': 'EB',
+    'EPUB': 'ED',
+    'MOBI': 'ED',
+}
+
+PRODUCT_FORMS_2 = {
+    'HTML': 'E105',
+    'PDF': 'E107',
+    'TXT': 'E112',
+    'EPUB': 'E101',
+    'MOBI': 'E127',
+}
+
+
+def is_institution(name):
+    return name.startswith(u'Zgromadzenie Ogólne')
+
+
+class Command(BaseCommand):
+    @staticmethod
+    def dc_values(desc, tag):
+        return [e.text for e in desc.findall('.//' + DCNS(tag))]
+
+    def handle(self, *args, **options):
+        writer = csv.writer(sys.stdout)
+        for book in Book.objects.all():
+            desc = book.wldocument().edoc.find('.//' + RDFNS('Description'))
+            formats = FORMATS_WITH_CHILDREN if book.children.exists() else FORMATS
+            for file_format in formats:
+                imprint = u'Fundacja Nowoczesna Polska'
+                title = book.title
+                subtitle = ''
+                year = ''
+                volume = ''
+                publication_date = localtime(book.created_at).date().isoformat()
+                info_date = publication_date
+                author = '; '.join(self.dc_values(desc, 'creator'))
+                author_person = author if not is_institution(author) else ''
+                author_institution = author if is_institution(author) else ''
+                publication_type = 'DGO'
+                edition = '1'
+                product_form1 = PRODUCT_FORMS_1[file_format]
+                product_form2 = PRODUCT_FORMS_2[file_format]
+                language = self.dc_values(desc, 'language')[0]
+                row = [
+                    imprint,
+                    title,
+                    subtitle,
+                    year,
+                    volume,
+                    publication_date,
+                    info_date,
+                    author_person,
+                    author_institution,
+                    publication_type,
+                    edition,
+                    product_form1,
+                    product_form2,
+                    language,
+                ]
+                writer.writerow([s.encode('utf-8') for s in row])