Question

I’m learning python for the first time.
I have an aim which is to take data from an API and output it as xml.

The output is stored in an array (“projectData”), here is an example of the output:

[{'code': 'demo',
 'created_at': datetime.datetime(2008, 6, 11, 7, 35, 19, tzinfo=<api.LocalTimezone object at 0x10072ab10>),
 'created_by': None,
 'id': 4,
 'image': 'https://website.com/files/0000/0000/blah.jpg',
 'name': 'Demo Project',
 'description': 'This is for demonstration purposes',
 'due': '2009-05-30',
 'start': '2009-05-06',
 'status': 'Active',
 'stype': 'Demo',
 'tag_list': [],
 'type': 'Project',
 'updated_at': datetime.datetime(2009, 5, 27, 1, 41, 55, tzinfo=<api.LocalTimezone object at 0x10072ab10>),
 'updated_by': {'id': 24, 'name': 'Test', 'type': 'HumanUser'},
 'users': [{'id': 19, 'name': 'User 1', 'type': 'HumanUser'},
           {'id': 18, 'name': 'User 2', 'type': 'HumanUser'},
           {'id': 17, 'name': 'User 3', 'type': 'HumanUser'},
           {'id': 16, 'name': 'User 4', 'type': 'HumanUser'},
           {'id': 15, 'name': 'User 5', 'type': 'HumanUser'},
           {'id': 14, 'name': 'User 6', 'type': 'HumanUser'},
           {'id': 13, 'name': 'User 7', 'type': 'HumanUser'},
           {'id': 12, 'name': 'User 8', 'type': 'HumanUser'},
           {'id': 20, 'name': 'Client 1', 'type': 'HumanUser'}]},

(etc.)

I’ve written some code which will output it as xml like so:

for _project in projectData:
  print "<Project>"
  for key in _project:
    value = _project[key]
    print "\t<" + str(key) + ">" + str(value) + "</" + str(key) + ">"
  print("</Project>\n")

Which actually gives me a result that works for me.

However, because I’m new to this, I suspect that this isn’t a very efficient approach, and may be susceptible to all sorts of bugs, I was hoping that someone more knowledgeable might have some pointers for me. The next thing I want to try with this is make it recursive, so that the “updated_by” element for example returns its own xml

Thanks.

Answer 1

Here’s an example using lxml.etree, incomplete.. and probably a bit naive. Really you should define a schema and make sure your output is consistent with it.

Edit, said it was incomplete, added None type and assumed a created_by is like an updated_by when populated

import datetime

projects = [{'code': 'demo',
 'created_at': datetime.datetime(2008, 6, 11, 7, 35, 19),
 'created_by': None,
 'id': 4,
 'image': 'https://website.com/files/0000/0000/blah.jpg',
 'name': 'Demo Project',
 'description': 'This is for demonstration purposes',
 'due': '2009-05-30',
 'start': '2009-05-06',
 'status': 'Active',
 'stype': 'Demo',
 'tag_list': [],
 'type': 'Project',
 'updated_at': datetime.datetime(2009, 5, 27, 1, 41, 55),
 'updated_by': {'id': 24, 'name': 'Test', 'type': 'HumanUser'},
 'users': [{'id': 19, 'name': 'User 1', 'type': 'HumanUser'},
           {'id': 18, 'name': 'User 2', 'type': 'HumanUser'},
           {'id': 17, 'name': 'User 3', 'type': 'HumanUser'},
           {'id': 16, 'name': 'User 4', 'type': 'HumanUser'},
           {'id': 15, 'name': 'User 5', 'type': 'HumanUser'},
           {'id': 14, 'name': 'User 6', 'type': 'HumanUser'},
           {'id': 13, 'name': 'User 7', 'type': 'HumanUser'},
           {'id': 12, 'name': 'User 8', 'type': 'HumanUser'},
           {'id': 20, 'name': 'Client 1', 'type': 'HumanUser'}]},
 ]

from lxml import etree

def E(tag, parent=None, content=None, children=None, **kw):
  e = etree.Element(tag)
  if not content is None:
    e.text = str(content)
  for k,v in kw.items():
    e.set(k, str(v))
  if not parent is None:
    parent.append(e)
  if not children is None:
    for c in children:
      e.append(c)
  return e

def processProject(data):
  attrs = ('name','type','id')
  p = E('Project')
  for item in attrs:
    p.set(item,str(data[item]))
  for k,v in [ x for x in data.items() if x[0] not in attrs ]:
    if v is None:
      E(k,parent=p)
    elif isinstance(v,basestring):
      E(k,content=v,parent=p)
    elif isinstance(v,(float,long,int)):
      E(k,content=str(v),parent=p)
    elif isinstance(v,datetime.datetime):
      E(k,content=v.strftime('%Y-%m-%d %H%M'),parent=p)
    elif k == 'users':
      users = E(k,parent=p)
      for u in v:
        E('user',parent=users,**dict([ (x,str(y)) for (x,y) in u.items()]))
    elif k in ('updated_by','created_by'):
      E(k,parent=p,**dict([ (x,str(y)) for (x,y) in v.items()]))
    elif k == 'tag_list':
      taglist = E(k,parent=p)
      for t in v:
        E('tag',parent=taglist,content=t)
  return p

>>> projxml = processProject(projects[0])
>>> etree.dump(projxml)
<Project name="Demo Project" type="Project" id="4">
  <status>Active</status>
  <code>demo</code>
  <created_at>2008-06-11 0735</created_at>
  <due>2009-05-30</due>
  <created_by/>
  <updated_at>2009-05-27 0141</updated_at>
  <start>2009-05-06</start>
  <image>https://website.com/files/0000/0000/blah.jpg</image>
  <updated_by type="HumanUser" id="24" name="Test"/>
  <users>
    <user type="HumanUser" id="19" name="User 1"/>
    <user type="HumanUser" id="18" name="User 2"/>
    <user type="HumanUser" id="17" name="User 3"/>
    <user type="HumanUser" id="16" name="User 4"/>
    <user type="HumanUser" id="15" name="User 5"/>
    <user type="HumanUser" id="14" name="User 6"/>
    <user type="HumanUser" id="13" name="User 7"/>
    <user type="HumanUser" id="12" name="User 8"/>
    <user type="HumanUser" id="20" name="Client 1"/>
  </users>
  <tag_list/>
  <stype>Demo</stype>
  <description>This is for demonstration purposes</description>
</Project>