File size: 5,005 Bytes
537088c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
import os
import io

import zstandard
import jsonlines
import datasets

import json


def parse_json(x):
    return json.loads(x)


_DESCRIPTION = "DaruLM for LLM"
_URLS = [
    "libru_accounting_0.jsonl.zst",
    "libru_antique_0.jsonl.zst",
    "libru_antique_1.jsonl.zst",
    "libru_aphorisms_0.jsonl.zst",
    "libru_art_0.jsonl.zst",
    "libru_biography_0.jsonl.zst",
    "libru_biography_1.jsonl.zst",
    "libru_biography_2.jsonl.zst",
    "libru_biography_3.jsonl.zst",
    "libru_biography_4.jsonl.zst",
    "libru_biology_0.jsonl.zst",
    "libru_business_0.jsonl.zst",
    "libru_cinema_0.jsonl.zst",
    "libru_computers_0.jsonl.zst",
    "libru_design_0.jsonl.zst",
    "libru_dramaturgy_0.jsonl.zst",
    "libru_economics_0.jsonl.zst",
    "libru_essay_0.jsonl.zst",
    "libru_essay_1.jsonl.zst",
    "libru_essay_2.jsonl.zst",
    "libru_fantasy_0.jsonl.zst",
    "libru_geography_0.jsonl.zst",
    "libru_guidebooks_0.jsonl.zst",
    "libru_guidebooks_1.jsonl.zst",
    "libru_history_0.jsonl.zst",
    "libru_history_1.jsonl.zst",
    "libru_history_2.jsonl.zst",
    "libru_history_3.jsonl.zst",
    "libru_history_4.jsonl.zst",
    "libru_history_5.jsonl.zst",
    "libru_humor_0.jsonl.zst",
    "libru_language_0.jsonl.zst",
    "libru_law_0.jsonl.zst",
    "libru_literature_0.jsonl.zst",
    "libru_medicine_0.jsonl.zst",
    "libru_military_0.jsonl.zst",
    "libru_music_0.jsonl.zst",
    "libru_philosophy_0.jsonl.zst",
    "libru_politic_0.jsonl.zst",
    "libru_prose_0.jsonl.zst",
    "libru_prose_1.jsonl.zst",
    "libru_prose_2.jsonl.zst",
    "libru_psychology_0.jsonl.zst",
    "libru_psychology_1.jsonl.zst",
    "libru_reference_0.jsonl.zst",
    "libru_religion_0.jsonl.zst",
    "libru_religion_1.jsonl.zst",
    "libru_religion_2.jsonl.zst",
    "libru_religion_3.jsonl.zst",
    "libru_science_0.jsonl.zst",
    "libru_science_1.jsonl.zst",
    "libru_science_2.jsonl.zst",
    "libru_sociology_0.jsonl.zst",
    "libru_textbook_0.jsonl.zst",
    "libru_UNDEFINED_0.jsonl.zst",
    "rulm_buriy_0.jsonl.zst",
    "rulm_buriy_1.jsonl.zst",
    "rulm_buriy_2.jsonl.zst",
    "rulm_gazeta_0.jsonl.zst",
    "rulm_habr_0.jsonl.zst",
    "rulm_habr_1.jsonl.zst",
    "rulm_lenta_0.jsonl.zst",
    "rulm_ods-tass_0.jsonl.zst",
    "rulm_ods-tass_1.jsonl.zst",
    "rulm_pikabu_0.jsonl.zst",
    "rulm_pikabu_1.jsonl.zst",
    "rulm_pikabu_2.jsonl.zst",
    "rulm_taiga-fontanka_0.jsonl.zst",
    "rulm_wiki_0.jsonl.zst",
    "rulm_wiki_1.jsonl.zst",
    "rulm_wiki_2.jsonl.zst",
    "wiki40_enwiki_0.jsonl.zst",
]


DOMAINS = sorted(set(url.split('_')[1] for url in _URLS))

_VERSION = datasets.Version("1.1.0")


class DarulmConfig(datasets.BuilderConfig):
    """BuilderConfig for DaruLM."""

    def __init__(self, domains="all", version=_VERSION, **kwargs):
        """BuilderConfig for DaruLM.
        Args:
            domains (str or list, default 'all'): domain names. 
            **kwargs: keyword arguments forwarded to super.
        """
        if isinstance(domains, str):
            domains = [domains]
        name = "+".join(domains)
        if name == "all":
            domains = DOMAINS

        missing_domains = set(domains)-set(DOMAINS)
        assert len(
            missing_domains) == 0, f"{missing_domains} not found in domain list"

        if version != _VERSION:
            name = f"{name}-{version}"
        super().__init__(name=name, version=version, **kwargs)
        self.domains = domains


class DarulmDataset(datasets.GeneratorBasedBuilder):
    VERSION = _VERSION

    BUILDER_CONFIG_CLASS = DarulmConfig
    BUILDER_CONFIGS = [
        DarulmConfig(
            domains="all",
            description="Full DaruLM collection.",
        ),
    ] + [
        DarulmConfig(
            domains=dn,
            description=f"DaruLM part with {dn} texts",
        )
        for dn in sorted(DOMAINS)
    ]
    DEFAULT_CONFIG_NAME = "all"

    def _info(self):
        features = datasets.Features(
            {
                "text": datasets.Value("string"),
                "domain": datasets.Value("string"),
            }
        )
        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=features
        )

    def _split_generators(self, dl_manager):
        downloaded_files = dl_manager.download(_URLS)
        paths = [
            path for dn in self.config.domains for path in downloaded_files if f"_{dn}_" in path
        ]
        return [
            datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={
                                    "paths": paths}),
        ]

    def _generate_examples(self, paths):
        for path in paths:
            with open(path, "rb") as f:
                cctx = zstandard.ZstdDecompressor()
                reader_stream = io.BufferedReader(cctx.stream_reader(f))
                reader = jsonlines.Reader(reader_stream, loads=parse_json)
                for id_, item in enumerate(reader):
                    yield id_, item