xmlpumpkin 0.1

Usage

Expected usages are focused on chunk surfaces and dependency links:

>>> aisansan = xmlpumpkin.parse_to_tree(
...     u'愛燦々とこの身に降って心密かなうれしい涙を流したりして'
... )
>>> len(aisansan.chunks)
8
>>> print(aisansan.root.surface)
流したりして
>>> print(aisansan.root.func_surface)
て
>>> for dep in aisansan.root.linked:
...     print(dep.surface)
...
降って
涙を

You need CaboCha in your path, or shortly with prepared XML:

>>> tree = xmlpumpkin.Tree(xml_as_unicode)

Should you need an easy interface from Python to CaboCha:

>>> from xmlpumpkin import cabocha
>>> print(cabocha.txttree(
...     u'愛燦々とこの身に降って心密かなうれしい涙を流したりして'
... ))
    愛燦々と-----D
          この-D |
            身に-D
            降って-------D
            心密かな---D |
              うれしい-D |
                    涙を-D
              流したりして
EOS
>>> print(cabocha.as_xml(
...     u'愛燦々とこの身に降って心密かなうれしい涙を流したりして'
... ))
<sentence>
  ...
</sentence>

All I/Os are unicodes! If encodings other than UTF-8 is preferred, directly modify following constants:

>>> import xmlpumpkin.runner
>>> xmlpumpkin.runner.CABOCHA_ENCODING = 'SJIS'
>>>
>>> import xmlpumpkin.tree
>>> xmlpumpkin.tree.XML_ENCODING = 'SJIS'

Properties

Not enough but a few properties are provided via Tree and Chunk objects.

class xmlpumpkin.Tree(cabocha_xml)

• chunks - tuple of chunks

• root - root (not depending on any chunks) Chunk object

• chunk_by_id(chunk_id) - get Chunk object by its id generated by CaboCha

• _element - origin XML as lxml Element object

class xmlpumpkin.Chunk(element, parent)

• id - chunk id

• link_to_id - its depending chunk id

• linked_from_ids - tuple of chunk id depending to this chunk

• func_id - functional token id of this chunk

• dep - its depending Chunk object

• linked - list of all Chunk objects depending to this chunk

• surface - surface of this chunk

• func_surface - surface of this chunk’s functional token

• _tokens() - its containing tokens as lxml Element objects