use metadata.opf or metadata if available

oml/media/opf.py

@@ -0,0 +1,45 @@
+# -*- coding: utf-8 -*-
+# vi:si:et:sw=4:sts=4:ts=4
+from __future__ import division
+
+import xml.etree.ElementTree as ET
+
+import stdnum.isbn
+
+from utils import normalize_isbn
+from ox import strip_tags
+import ox.iso
+
+def info(opf):
+    data = {}
+    with open(opf) as fd:
+        opf = ET.fromstring(fd.read())
+    ns = '{http://www.idpf.org/2007/opf}'
+    metadata = opf.findall(ns + 'metadata')[0]
+    for e in metadata.getchildren():
+        if e.text:
+            key = e.tag.split('}')[-1]
+            key = {
+                'creator': 'author',
+            }.get(key, key)
+            value = e.text
+            if key == 'identifier':
+                isbn = normalize_isbn(value)
+                if stdnum.isbn.is_valid(isbn):
+                    if not 'isbn' in data:
+                        data['isbn'] = [isbn]
+                    else:
+                        data['isbn'].append(isbn)
+                if e.attrib.get(ns + 'scheme') == 'AMAZON':
+                    if not 'asin' in data:
+                        data['asin'] = [value]
+                    else:
+                        data['asin'].append(value)
+            else:
+                data[key] = strip_tags(e.text)
+    #YYY-MM-DD
+    if 'date' in data and len(data['date']) > 10:
+        data['date'] =data['date'][:10]
+    if 'language' in data:
+        data['language'] = ox.iso.codeToLang(data['language'])
+    return data