This package models graphs encoded in PENMAN notation (e.g., AMR), such as the following for the boy wants to go:
(w / want-01
:ARG0 (b / boy)
:ARG1 (g / go
:ARG0 b))
The Penman package may be used as a Python library or as a script.
- Read and write PENMAN-serialized graphs or triple conjunctions
- Read metadata in comments (e.g.,
# ::id 1234
) - Read surface alignments (e.g.,
foo~e.1,2
) - Inspect and manipulate the graph or tree structures
- Customize graphs for writing:
- Adjust indentation and compactness
- Select a new top node
- Rearrange edges
- Restructure the tree shape
- Relabel node variables
- Transform the graph
- Canonicalize roles
- Reify and dereify edges
- Reify attributes
- Embed the tree structure with additional
TOP
triples
- AMR model: role inventory and transformations
- Check graphs for model compliance
- Tested (but not yet 100% coverage)
- Documented (see the documentation)
>>> import penman
>>> g = penman.decode('(b / bark-01 :ARG0 (d / dog))')
>>> g.triples
[('b', ':instance', 'bark-01'), ('b', ':ARG0', 'd'), ('d', ':instance', 'dog')]
>>> g.edges()
[Edge(source='b', role=':ARG0', target='d')]
>>> print(penman.encode(g, indent=3))
(b / bark-01
:ARG0 (d / dog))
>>> print(penman.encode(g, indent=None))
(b / bark-01 :ARG0 (d / dog))
$ echo "(w / want-01 :ARG0 (b / boy) :ARG1 (g / go :ARG0 b))" | penman
(w / want-01
:ARG0 (b / boy)
:ARG1 (g / go
:ARG0 b))
$ echo "(w / want-01 :ARG0 (b / boy) :ARG1 (g / go :ARG0 b))" | penman --make-variables="a{i}"
(a0 / want-01
:ARG0 (a1 / boy)
:ARG1 (a2 / go
:ARG0 a1))
For a demonstration of the API usage, see the included Jupyter notebook:
-
View it on GitHub: docs/api-demo.ipynb
-
Run it on mybinder.org:
(Note: clear the output before running: Cell > All Output > Clear):
A description of the PENMAN notation can be found in the
documentation.
This module expands the original notation slightly to allow for
untyped nodes (e.g., (x)
) and anonymous relations (e.g., (x : (y))
). It also accommodates slightly malformed graphs as well as
surface alignments.
If you make use of Penman in your work, please cite Goodman, 2020. The BibTeX is below:
@inproceedings{goodman-2020-penman,
title = "{P}enman: An Open-Source Library and Tool for {AMR} Graphs",
author = "Goodman, Michael Wayne",
booktitle = "Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics: System Demonstrations",
month = jul,
year = "2020",
address = "Online",
publisher = "Association for Computational Linguistics",
url = "https://www.aclweb.org/anthology/2020.acl-demos.35",
pages = "312--319",
abstract = "Abstract Meaning Representation (AMR) (Banarescu et al., 2013) is a framework for semantic dependencies that encodes its rooted and directed acyclic graphs in a format called PENMAN notation. The format is simple enough that users of AMR data often write small scripts or libraries for parsing it into an internal graph representation, but there is enough complexity that these users could benefit from a more sophisticated and well-tested solution. The open-source Python library Penman provides a robust parser, functions for graph inspection and manipulation, and functions for formatting graphs into PENMAN notation. Many functions are also available in a command-line tool, thus extending its utility to non-Python setups.",
}
For the graph transformation/normalization work, please use the following:
@inproceedings{Goodman:2019,
title = "{AMR} Normalization for Fairer Evaluation",
author = "Goodman, Michael Wayne",
booktitle = "Proceedings of the 33rd Pacific Asia Conference on Language, Information, and Computation",
year = "2019",
pages = "47--56",
address = "Hakodate"
}
This project is not affiliated with ISI, the PENMAN project, or the AMR project.