umarbutler commited on
Commit
494f500
1 Parent(s): c84ed9a

Update open_australian_legal_embeddings.py

Browse files
Files changed (1) hide show
  1. open_australian_legal_embeddings.py +109 -0
open_australian_legal_embeddings.py ADDED
@@ -0,0 +1,109 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Copyright 2023 Umar Butler.
2
+ #
3
+ # Licensed under the Apache License, Version 2.0 (the "License");
4
+ # you may not use this file except in compliance with the License.
5
+ # You may obtain a copy of the License at
6
+ #
7
+ # http://www.apache.org/licenses/LICENSE-2.0
8
+ #
9
+ # Unless required by applicable law or agreed to in writing, software
10
+ # distributed under the License is distributed on an "AS IS" BASIS,
11
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12
+ # See the License for the specific language governing permissions and
13
+ # limitations under the License.
14
+ """Open Australian Legal Embeddings: the first open-source embeddings of Australian legislative and judicial documents"""
15
+
16
+ import datasets
17
+ for module in ('orjson', 'ujson', 'json'):
18
+ try:
19
+ json = __import__(module)
20
+
21
+ break
22
+ except ImportError:
23
+ pass
24
+
25
+ _CITATION = """\
26
+ @misc{butler-2023-open-australian-legal-embeddings,
27
+ author = {Butler, Umar},
28
+ year = {2023},
29
+ title = {Open Australian Legal Embeddings},
30
+ publisher = {Hugging Face},
31
+ version = {1.0.0},
32
+ url = {https://huggingface.co/datasets/umarbutler/open-australian-legal-embeddings}
33
+ }
34
+ """
35
+
36
+ _DESCRIPTION = """\
37
+ The Open Australian Legal Embeddings are the first open-source embeddings of Australian legislative and judicial documents.
38
+
39
+ Trained on the largest open database of Australian law, the [Open Australian Legal Corpus](https://huggingface.co/datasets/umarbutler/open-australian-legal-corpus), the Embeddings consist of roughly 5.2 million 384-dimensional vectors embedded with [`BAAI/bge-small-en-v1.5`](https://huggingface.co/BAAI/bge-small-en-v1.5).
40
+
41
+ The Embeddings open the door to a wide range of possibilities in the field of Australian legal AI, including the development of document classifiers, search engines and chatbots.
42
+
43
+ To ensure their accessibility to as wide an audience as possible, the Embeddings are distributed under the same licence as the [Open Australian Legal Corpus](https://huggingface.co/datasets/umarbutler/open-australian-legal-corpus/blob/main/LICENCE.md)."""
44
+
45
+ _HOMEPAGE = "https://huggingface.co/datasets/umarbutler/open-australian-legal-embeddings"
46
+
47
+ _LICENSE = """\
48
+ The Embeddings are distributed under the same licence as the [Open Australian Legal Corpus](https://huggingface.co/datasets/umarbutler/open-australian-legal-corpus/blob/main/LICENCE.md)."""
49
+
50
+ _URLS = {
51
+ 'embeddings' : 'data/embeddings.jsonl',
52
+ 'metadatas' : 'data/metadatas.jsonl',
53
+ 'texts' : 'data/texts.jsonl',
54
+ }
55
+
56
+ class OpenAustralianLegalEmbeddings(datasets.GeneratorBasedBuilder):
57
+ """Open Australian Legal Embeddings: the first open-source embeddings of Australian legislative and judicial documents"""
58
+
59
+ VERSION = datasets.Version("1.0.0")
60
+
61
+ DEFAULT_CONFIG_NAME = "train"
62
+
63
+ def _info(self):
64
+ return datasets.DatasetInfo(
65
+ description=_DESCRIPTION,
66
+ features=datasets.Features(
67
+ {
68
+ 'version_id' : datasets.Value('string'),
69
+ 'type' : datasets.Value('string'),
70
+ 'jurisdiction' : datasets.Value('string'),
71
+ 'source' : datasets.Value('string'),
72
+ 'citation' : datasets.Value('string'),
73
+ 'url' : datasets.Value('string'),
74
+ 'is_last_chunk' : datasets.Value('bool'),
75
+ 'text' : datasets.Value('string'),
76
+ 'embedding' : [datasets.Value('float32')]
77
+ }
78
+ ),
79
+ homepage=_HOMEPAGE,
80
+ license=_LICENSE,
81
+ citation=_CITATION,
82
+ )
83
+
84
+ def _split_generators(self, dl_manager):
85
+ downloaded_files = dl_manager.download_and_extract(_URLS)
86
+
87
+ return [
88
+ datasets.SplitGenerator(
89
+ name=datasets.Split.TRAIN,
90
+ gen_kwargs={
91
+ 'embeddings_path' : downloaded_files['embeddings'],
92
+ 'metadatas_path' : downloaded_files['metadatas'],
93
+ 'texts_path' : downloaded_files['texts'],
94
+ }
95
+ )
96
+ ]
97
+
98
+ def _generate_examples(self, embeddings_path, metadatas_path, texts_path):
99
+ with open(embeddings_path, 'rb') as embeddings_file, \
100
+ open(metadatas_path, 'rb') as metadatas_file, \
101
+ open(texts_path, 'rb') as texts_file:
102
+ i = -1
103
+
104
+ for embedding, metadata, text in zip(embeddings_file, metadatas_file, texts_file):
105
+ i += 1
106
+ yield i, json.loads(metadata) | {
107
+ 'text' : json.loads(text),
108
+ 'embedding' : json.loads(embedding)
109
+ }