Commit
·
196184d
1
Parent(s):
943d45f
Create argmicro.py
Browse files- argmicro.py +196 -0
argmicro.py
ADDED
@@ -0,0 +1,196 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
"""The Argumentative Microtext Corpus for German and English Argumentation Mining."""
|
2 |
+
import glob
|
3 |
+
import logging
|
4 |
+
import os
|
5 |
+
from os.path import abspath, isdir
|
6 |
+
from pathlib import Path
|
7 |
+
from xml.etree import ElementTree
|
8 |
+
|
9 |
+
import datasets
|
10 |
+
|
11 |
+
_CITATION = """\
|
12 |
+
@inproceedings{peldszus2015annotated,
|
13 |
+
title={An annotated corpus of argumentative microtexts},
|
14 |
+
author={Peldszus, Andreas and Stede, Manfred},
|
15 |
+
booktitle={Argumentation and Reasoned Action: Proceedings of the 1st European Conference on Argumentation, Lisbon},
|
16 |
+
volume={2},
|
17 |
+
pages={801--815},
|
18 |
+
year={2015}
|
19 |
+
}
|
20 |
+
"""
|
21 |
+
|
22 |
+
_DESCRIPTION = "The Argumentative Microtext Corpus for German and English Argumentation Mining."
|
23 |
+
|
24 |
+
_HOMEPAGE = "http://angcl.ling.uni-potsdam.de/resources/argmicro.html"
|
25 |
+
|
26 |
+
_LICENSE = "Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License, see https://creativecommons.org/licenses/by-nc-sa/4.0/"
|
27 |
+
|
28 |
+
|
29 |
+
# The HuggingFace dataset library don't host the datasets but only point to the original files
|
30 |
+
# This can be an arbitrary nested dict/list of URLs (see below in `_split_generators` method)
|
31 |
+
_URL = "https://github.com/peldszus/arg-microtexts/archive/refs/heads/master.zip"
|
32 |
+
|
33 |
+
_VERSION = datasets.Version("1.0.0")
|
34 |
+
|
35 |
+
_STANCE_CLASS_LABELS = ["con", "pro", "unclear", "UNDEFINED"]
|
36 |
+
_ADU_CLASS_LABELS = ["opp", "pro"]
|
37 |
+
_EDGE_CLASS_LABELS = ["seg", "sup", "exa", "add", "reb", "und"]
|
38 |
+
|
39 |
+
|
40 |
+
logger = logging.getLogger(__name__)
|
41 |
+
|
42 |
+
|
43 |
+
class ArgMicro(datasets.GeneratorBasedBuilder):
|
44 |
+
"""ArgMicro is a argumentation mining dataset."""
|
45 |
+
|
46 |
+
BUILDER_CONFIGS = [
|
47 |
+
datasets.BuilderConfig(name="en"),
|
48 |
+
datasets.BuilderConfig(name="de"),
|
49 |
+
]
|
50 |
+
|
51 |
+
def _info(self):
|
52 |
+
features = datasets.Features(
|
53 |
+
{
|
54 |
+
"id": datasets.Value("string"),
|
55 |
+
"topic_id": datasets.Value("string"),
|
56 |
+
"stance": datasets.ClassLabel(names=_STANCE_CLASS_LABELS),
|
57 |
+
"text": datasets.Value("string"),
|
58 |
+
"edus": datasets.Sequence(
|
59 |
+
{
|
60 |
+
"id": datasets.Value("string"),
|
61 |
+
"start": datasets.Value("int32"),
|
62 |
+
"end": datasets.Value("int32"),
|
63 |
+
}
|
64 |
+
),
|
65 |
+
"adus": datasets.Sequence(
|
66 |
+
{
|
67 |
+
"id": datasets.Value("string"),
|
68 |
+
"type": datasets.ClassLabel(names=_ADU_CLASS_LABELS),
|
69 |
+
}
|
70 |
+
),
|
71 |
+
"edges": datasets.Sequence(
|
72 |
+
{
|
73 |
+
"id": datasets.Value("string"),
|
74 |
+
"src": datasets.Value("string"),
|
75 |
+
"trg": datasets.Value("string"),
|
76 |
+
"type": datasets.ClassLabel(names=_EDGE_CLASS_LABELS),
|
77 |
+
}
|
78 |
+
),
|
79 |
+
}
|
80 |
+
)
|
81 |
+
|
82 |
+
return datasets.DatasetInfo(
|
83 |
+
# This is the description that will appear on the datasets page.
|
84 |
+
description=_DESCRIPTION,
|
85 |
+
# This defines the different columns of the dataset and their types
|
86 |
+
features=features, # Here we define them above because they are different between the two configurations
|
87 |
+
# If there's a common (input, target) tuple from the features,
|
88 |
+
# specify them here. They'll be used if as_supervised=True in
|
89 |
+
# builder.as_dataset.
|
90 |
+
supervised_keys=None,
|
91 |
+
# Homepage of the dataset for documentation
|
92 |
+
homepage=_HOMEPAGE,
|
93 |
+
# License for the dataset if available
|
94 |
+
license=_LICENSE,
|
95 |
+
# Citation for the dataset
|
96 |
+
citation=_CITATION,
|
97 |
+
)
|
98 |
+
|
99 |
+
def _split_generators(self, dl_manager):
|
100 |
+
"""Returns SplitGenerators."""
|
101 |
+
# If several configurations are possible (listed in BUILDER_CONFIGS), the configuration selected by the user is in self.config.name
|
102 |
+
|
103 |
+
# dl_manager is a datasets.download.DownloadManager that can be used to download and extract URLs
|
104 |
+
# It can accept any type or nested list/dict and will give back the same structure with the url replaced with path to local files.
|
105 |
+
# By default the archives will be extracted and a path to a cached folder where they are extracted is returned instead of the archive
|
106 |
+
|
107 |
+
if dl_manager.manual_dir is not None:
|
108 |
+
base_path = abspath(dl_manager.manual_dir)
|
109 |
+
if not isdir(base_path):
|
110 |
+
base_path = os.path.join(dl_manager.extract(base_path), "arg-microtexts-master")
|
111 |
+
else:
|
112 |
+
# TODO: test this!
|
113 |
+
base_path = os.path.join(
|
114 |
+
dl_manager.download_and_extract(_URL), "arg-microtexts-master"
|
115 |
+
)
|
116 |
+
base_path = Path(base_path) / "corpus"
|
117 |
+
|
118 |
+
try:
|
119 |
+
from lxml import etree
|
120 |
+
|
121 |
+
dtd = etree.DTD(base_path / "arggraph.dtd")
|
122 |
+
except ModuleNotFoundError:
|
123 |
+
logger.warning("lxml not installed. Skipping DTD validation.")
|
124 |
+
dtd = None
|
125 |
+
etree = None
|
126 |
+
|
127 |
+
return [
|
128 |
+
datasets.SplitGenerator(
|
129 |
+
name=datasets.Split.TRAIN,
|
130 |
+
gen_kwargs={"path": base_path / self.config.name, "dtd": dtd, "etree": etree},
|
131 |
+
),
|
132 |
+
]
|
133 |
+
|
134 |
+
def _generate_examples(self, path, dtd=None, etree=None):
|
135 |
+
"""Yields examples."""
|
136 |
+
# This method will receive as arguments the `gen_kwargs` defined in the previous `_split_generators` method.
|
137 |
+
# It is in charge of opening the given file and yielding (key, example) tuples from the dataset
|
138 |
+
# The key is not important, it's more here for legacy reason (legacy from tfds)
|
139 |
+
|
140 |
+
_id = 0
|
141 |
+
text_file_names = sorted(glob.glob(f"{path}/*.txt"))
|
142 |
+
if len(text_file_names) == 0:
|
143 |
+
raise Exception(f"No text files found in {path}. Did you set the correct data_dir?")
|
144 |
+
invalid_files = []
|
145 |
+
for text_file_name in text_file_names:
|
146 |
+
txt_fn = Path(text_file_name)
|
147 |
+
ann_fn = txt_fn.with_suffix(".xml")
|
148 |
+
with open(txt_fn, encoding="utf-8") as f:
|
149 |
+
text = f.read()
|
150 |
+
|
151 |
+
# validate xml file, if dtd is available
|
152 |
+
if dtd is not None and etree is not None:
|
153 |
+
e = etree.parse(ann_fn)
|
154 |
+
v = dtd.validate(e)
|
155 |
+
if not v:
|
156 |
+
logger.error(f"{ann_fn} is INVALID:")
|
157 |
+
logger.error(dtd.error_log.filter_from_errors()[0])
|
158 |
+
invalid_files.append(ann_fn)
|
159 |
+
continue
|
160 |
+
|
161 |
+
annotations = ElementTree.parse(ann_fn).getroot()
|
162 |
+
edus = []
|
163 |
+
start_pos = 0
|
164 |
+
for edu in annotations.findall("edu"):
|
165 |
+
start = text.find(edu.text, start_pos)
|
166 |
+
if start == -1:
|
167 |
+
raise Exception(f"Cannot find {edu.text} in {text}")
|
168 |
+
end = start + len(edu.text)
|
169 |
+
edus.append({"id": edu.attrib["id"], "start": start, "end": end})
|
170 |
+
start_pos = end
|
171 |
+
adus = [
|
172 |
+
{"id": adu.attrib["id"], "type": adu.attrib["type"]}
|
173 |
+
for adu in annotations.findall("adu")
|
174 |
+
]
|
175 |
+
edges = [
|
176 |
+
{
|
177 |
+
"id": edge.attrib["id"],
|
178 |
+
"src": edge.attrib["src"],
|
179 |
+
"trg": edge.attrib["trg"],
|
180 |
+
"type": edge.attrib["type"],
|
181 |
+
}
|
182 |
+
for edge in annotations.findall("edge")
|
183 |
+
]
|
184 |
+
yield _id, {
|
185 |
+
"id": annotations.attrib["id"],
|
186 |
+
"topic_id": annotations.attrib.get("topic_id", "UNDEFINED"),
|
187 |
+
"stance": annotations.attrib.get("stance", "UNDEFINED"),
|
188 |
+
"text": text,
|
189 |
+
"edus": edus,
|
190 |
+
"adus": adus,
|
191 |
+
"edges": edges,
|
192 |
+
}
|
193 |
+
_id += 1
|
194 |
+
|
195 |
+
if len(invalid_files) > 0:
|
196 |
+
raise Exception(f"Found {len(invalid_files)} invalid files: {invalid_files}")
|