penman.transform

Tree and graph transformations.

See also

The transformation functions in this module alter the content of the graph. Other functions may change the shape or form of the graph without altering its content, such as:

• penman.layout.rearrange()

• penman.layout.reconfigure()

• penman.tree.Tree.reset_variables()

penman.transform.canonicalize_roles(t, model)

Normalize roles in t so they are canonical according to model.

This is a tree transformation instead of a graph transformation because the orientation of the pure graph’s triples is not decided until the graph is configured into a tree.

Parameters

• t – a Tree object

• model – a model defining role normalizations

Returns

A new Tree object with canonicalized roles.

Example

>>> from penman.codec import PENMANCodec
>>> from penman.models.amr import model
>>> from penman.transform import canonicalize_roles
>>> codec = PENMANCodec()
>>> t = codec.parse('(c / chapter :domain-of 7)')
>>> t = canonicalize_roles(t, model)
>>> print(codec.format(t))
(c / chapter
   :mod 7)

penman.transform.reify_edges(g, model)

Reify all edges in g that have reifications in model.

Parameters

• g – a Graph object

• model – a model defining reifications

Returns

A new Graph object with reified edges.

Example

>>> from penman.codec import PENMANCodec
>>> from penman.models.amr import model
>>> from penman.transform import reify_edges
>>> codec = PENMANCodec(model=model)
>>> g = codec.decode('(c / chapter :mod 7)')
>>> g = reify_edges(g, model)
>>> print(codec.encode(g))
(c / chapter
   :ARG1-of (_ / have-mod-91
               :ARG2 7))

penman.transform.dereify_edges(g, model)

Dereify edges in g that have reifications in model.

Parameters

g – a Graph object

Returns

A new Graph object with dereified edges.

Example

>>> from penman.codec import PENMANCodec
>>> from penman.models.amr import model
>>> from penman.transform import dereify_edges
>>> codec = PENMANCodec(model=model)
>>> g = codec.decode(
...   '(c / chapter'
...   '   :ARG1-of (_ / have-mod-91'
...   '               :ARG2 7))')
>>> g = dereify_edges(g, model)
>>> print(codec.encode(g))
(c / chapter
   :mod 7)

penman.transform.reify_attributes(g)

Reify all attributes in g.

Parameters

g – a Graph object

Returns

A new Graph object with reified attributes.

Example

>>> from penman.codec import PENMANCodec
>>> from penman.models.amr import model
>>> from penman.transform import reify_attributes
>>> codec = PENMANCodec(model=model)
>>> g = codec.decode('(c / chapter :mod 7)')
>>> g = reify_attributes(g)
>>> print(codec.encode(g))
(c / chapter
   :mod (_ / 7))

penman.transform.indicate_branches(g, model)

Insert TOP triples in g indicating the tree structure.

Note

This depends on g containing the epigraphical layout markers from parsing; it will not work with programmatically constructed Graph objects or those whose epigraphical data were removed.

Parameters

• g – a Graph object

• model – a model defining the TOP role

Returns

A new Graph object with TOP roles indicating tree branches.

Example

>>> from penman.codec import PENMANCodec
>>> from penman.models.amr import model
>>> from penman.transform import indicate_branches
>>> codec = PENMANCodec(model=model)
>>> g = codec.decode('''
... (w / want-01
...    :ARG0 (b / boy)
...    :ARG1 (g / go-02
...             :ARG0 b))''')
>>> g = indicate_branches(g, model)
>>> print(codec.encode(g))
(w / want-01
   :TOP b
   :ARG0 (b / boy)
   :TOP g
   :ARG1 (g / go-02
            :ARG0 b))