[ckan#3390] datastore dump format=xml

ckanext/datastore/controller.py

@@ -19,12 +19,13 @@
     csv_writer,
     tsv_writer,
     json_writer,
+    xml_writer,
 )
 
 int_validator = get_validator('int_validator')
 boolean_validator = get_validator('boolean_validator')
 
-DUMP_FORMATS = 'csv', 'tsv', 'json'
+DUMP_FORMATS = 'csv', 'tsv', 'json', 'xml'
 PAGINATE_BY = 10000
 
 
@@ -48,6 +49,8 @@ def start_writer(columns):
                 return tsv_writer(response, columns, resource_id, bom)
             if fmt == 'json':
                 return json_writer(response, columns, resource_id, bom)
+            if fmt == 'xml':
+                return xml_writer(response, columns, resource_id, bom)
             abort(400, _(
                 u'format: must be one of %s') % u', '.join(DUMP_FORMATS))
 

ckanext/datastore/writer.py

@@ -3,6 +3,7 @@
 from contextlib import contextmanager
 from email.utils import encode_rfc2231
 import json
+from xml.etree.cElementTree import Element, SubElement, ElementTree
 
 import unicodecsv
 
@@ -113,3 +114,54 @@ def writerow(self, row):
             ensure_ascii=False,
             separators=(u',', u':'),
             sort_keys=True).encode(u'utf-8'))
+
+
+@contextmanager
+def xml_writer(response, columns, name=None, bom=False):
+    u'''Context manager for writing UTF-8 XML data to response
+
+    :param response: file-like or response-like object for writing
+        data and headers (response-like objects only)
+    :param columns: list of column names
+    :param name: file name (for headers, response-like objects only)
+    :param bom: True to include a UTF-8 BOM at the start of the file
+
+    >>> with xml_writer(response, fields) as d:
+    >>>    d.writerow(row1)
+    >>>    d.writerow(row2)
+    '''
+
+    if hasattr(response, u'headers'):
+        response.headers['Content-Type'] = (
+            b'text/xml; charset=utf-8')
+        if name:
+            response.headers['Content-disposition'] = (
+                b'attachment; filename="{name}.xml"'.format(
+                    name=encode_rfc2231(name)))
+    if bom:
+        response.write(UTF8_BOM)
+    response.write(b'<data>\n')
+    yield XMLWriter(response, columns)
+    response.write(b'</data>\n')
+
+
+class XMLWriter(object):
+    def __init__(self, response, columns):
+        self.response = response
+        self.id_col = columns[0] == u'_id'
+        if self.id_col:
+            columns = columns[1:]
+        self.columns = columns
+
+    def writerow(self, row):
+        root = Element(u'row')
+        if self.id_col:
+            root.attrib[u'_id'] = unicode(row[0])
+            row = row[1:]
+        for k, v in zip(self.columns, row):
+            if v is None:
+                SubElement(root, k).attrib[u'xsi:nil'] = u'true'
+                continue
+            SubElement(root, k).text = unicode(v)
+        ElementTree(root).write(self.response, encoding=u'utf-8')
+        self.response.write(b'\n')

doc/maintaining/datastore.rst

@@ -282,9 +282,10 @@ A DataStore resource can be downloaded in the `CSV`_ file format from ``{CKAN-UR
 
 For an Excel-compatible CSV file use ``{CKAN-URL}/datastore/dump/{RESOURCE-ID}?bom=true``.
 
-Other formats are also supported. For tab-separated values use
-``{CKAN-URL}/datastore/dump/{RESOURCE-ID}?format=tsv`` and for JSON use
-``{CKAN-URL}/datastore/dump/{RESOURCE-ID}?format=json``.
+Other formats supported include tab-separated values (``?format=tsv``),
+JSON (``?format=json``) and XML (``?format=xml``). E.g. to download an Excel-compatible
+tab-separated file use
+``{CKAN-URL}/datastore/dump/{RESOURCE-ID}?format=tsv&bom=true``.
 
 .. _CSV: https://en.wikipedia.org/wiki/Comma-separated_values