Datasets:

Modalities:
Text
Formats:
json
Languages:
Thai
ArXiv:
License:
yacht commited on
Commit
b35eaa3
·
0 Parent(s):

Initial release

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .gitattributes +61 -0
  2. LICENSE +202 -0
  3. NOTICE +37 -0
  4. README.md +228 -0
  5. benchmark.py +119 -0
  6. build_testsets.py +289 -0
  7. checksums.json +17 -0
  8. data/pud.jsonl +0 -0
  9. data/pudA.jsonl +0 -0
  10. data/pudB.jsonl +0 -0
  11. data/tud.jsonl +0 -0
  12. data/vistec.jsonl +3 -0
  13. data/ws160.jsonl +160 -0
  14. fetch_sources.py +65 -0
  15. predictions/attacut-1.0.6-c/bestdd.words.jsonl.gz +3 -0
  16. predictions/attacut-1.0.6-c/khm.words.jsonl.gz +3 -0
  17. predictions/attacut-1.0.6-c/khmdd.words.jsonl.gz +3 -0
  18. predictions/attacut-1.0.6-c/orchidndd.words.jsonl.gz +3 -0
  19. predictions/attacut-1.0.6-c/pud.tokens.jsonl.gz +3 -0
  20. predictions/attacut-1.0.6-c/pud.words.jsonl.gz +3 -0
  21. predictions/attacut-1.0.6-c/tud.tokens.jsonl.gz +3 -0
  22. predictions/attacut-1.0.6-c/tud.words.jsonl.gz +3 -0
  23. predictions/attacut-1.0.6-c/vistec.tokens.jsonl.gz +3 -0
  24. predictions/attacut-1.0.6-c/vistec.words.jsonl.gz +3 -0
  25. predictions/attacut-1.0.6-c/ws160.tokens.jsonl.gz +3 -0
  26. predictions/attacut-1.0.6-c/ws160.words.jsonl.gz +3 -0
  27. predictions/attacut-1.0.6-sc/bestdd.words.jsonl.gz +3 -0
  28. predictions/attacut-1.0.6-sc/khm.words.jsonl.gz +3 -0
  29. predictions/attacut-1.0.6-sc/khmdd.words.jsonl.gz +3 -0
  30. predictions/attacut-1.0.6-sc/orchidndd.words.jsonl.gz +3 -0
  31. predictions/attacut-1.0.6-sc/pud.tokens.jsonl.gz +3 -0
  32. predictions/attacut-1.0.6-sc/pud.words.jsonl.gz +3 -0
  33. predictions/attacut-1.0.6-sc/tud.tokens.jsonl.gz +3 -0
  34. predictions/attacut-1.0.6-sc/tud.words.jsonl.gz +3 -0
  35. predictions/attacut-1.0.6-sc/vistec.tokens.jsonl.gz +3 -0
  36. predictions/attacut-1.0.6-sc/vistec.words.jsonl.gz +3 -0
  37. predictions/attacut-1.0.6-sc/ws160.tokens.jsonl.gz +3 -0
  38. predictions/attacut-1.0.6-sc/ws160.words.jsonl.gz +3 -0
  39. predictions/deepcut-0.7.0.0/bestdd.words.jsonl.gz +3 -0
  40. predictions/deepcut-0.7.0.0/khm.words.jsonl.gz +3 -0
  41. predictions/deepcut-0.7.0.0/khmdd.words.jsonl.gz +3 -0
  42. predictions/deepcut-0.7.0.0/orchidndd.words.jsonl.gz +3 -0
  43. predictions/deepcut-0.7.0.0/pud.tokens.jsonl.gz +3 -0
  44. predictions/deepcut-0.7.0.0/pud.words.jsonl.gz +3 -0
  45. predictions/deepcut-0.7.0.0/tud.tokens.jsonl.gz +3 -0
  46. predictions/deepcut-0.7.0.0/tud.words.jsonl.gz +3 -0
  47. predictions/deepcut-0.7.0.0/vistec.tokens.jsonl.gz +3 -0
  48. predictions/deepcut-0.7.0.0/vistec.words.jsonl.gz +3 -0
  49. predictions/deepcut-0.7.0.0/ws160.tokens.jsonl.gz +3 -0
  50. predictions/deepcut-0.7.0.0/ws160.words.jsonl.gz +3 -0
.gitattributes ADDED
@@ -0,0 +1,61 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ *.7z filter=lfs diff=lfs merge=lfs -text
2
+ *.arrow filter=lfs diff=lfs merge=lfs -text
3
+ *.avro filter=lfs diff=lfs merge=lfs -text
4
+ *.bin filter=lfs diff=lfs merge=lfs -text
5
+ *.bz2 filter=lfs diff=lfs merge=lfs -text
6
+ *.ckpt filter=lfs diff=lfs merge=lfs -text
7
+ *.ftz filter=lfs diff=lfs merge=lfs -text
8
+ *.gz filter=lfs diff=lfs merge=lfs -text
9
+ *.h5 filter=lfs diff=lfs merge=lfs -text
10
+ *.joblib filter=lfs diff=lfs merge=lfs -text
11
+ *.lfs.* filter=lfs diff=lfs merge=lfs -text
12
+ *.lz4 filter=lfs diff=lfs merge=lfs -text
13
+ *.mds filter=lfs diff=lfs merge=lfs -text
14
+ *.mlmodel filter=lfs diff=lfs merge=lfs -text
15
+ *.model filter=lfs diff=lfs merge=lfs -text
16
+ *.msgpack filter=lfs diff=lfs merge=lfs -text
17
+ *.npy filter=lfs diff=lfs merge=lfs -text
18
+ *.npz filter=lfs diff=lfs merge=lfs -text
19
+ *.onnx filter=lfs diff=lfs merge=lfs -text
20
+ *.ot filter=lfs diff=lfs merge=lfs -text
21
+ *.parquet filter=lfs diff=lfs merge=lfs -text
22
+ *.pb filter=lfs diff=lfs merge=lfs -text
23
+ *.pickle filter=lfs diff=lfs merge=lfs -text
24
+ *.pkl filter=lfs diff=lfs merge=lfs -text
25
+ *.pt filter=lfs diff=lfs merge=lfs -text
26
+ *.pth filter=lfs diff=lfs merge=lfs -text
27
+ *.rar filter=lfs diff=lfs merge=lfs -text
28
+ *.safetensors filter=lfs diff=lfs merge=lfs -text
29
+ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
30
+ *.tar.* filter=lfs diff=lfs merge=lfs -text
31
+ *.tar filter=lfs diff=lfs merge=lfs -text
32
+ *.tflite filter=lfs diff=lfs merge=lfs -text
33
+ *.tgz filter=lfs diff=lfs merge=lfs -text
34
+ *.wasm filter=lfs diff=lfs merge=lfs -text
35
+ *.xz filter=lfs diff=lfs merge=lfs -text
36
+ *.zip filter=lfs diff=lfs merge=lfs -text
37
+ *.zst filter=lfs diff=lfs merge=lfs -text
38
+ *tfevents* filter=lfs diff=lfs merge=lfs -text
39
+ # Audio files - uncompressed
40
+ *.pcm filter=lfs diff=lfs merge=lfs -text
41
+ *.sam filter=lfs diff=lfs merge=lfs -text
42
+ *.raw filter=lfs diff=lfs merge=lfs -text
43
+ # Audio files - compressed
44
+ *.aac filter=lfs diff=lfs merge=lfs -text
45
+ *.flac filter=lfs diff=lfs merge=lfs -text
46
+ *.mp3 filter=lfs diff=lfs merge=lfs -text
47
+ *.ogg filter=lfs diff=lfs merge=lfs -text
48
+ *.wav filter=lfs diff=lfs merge=lfs -text
49
+ # Image files - uncompressed
50
+ *.bmp filter=lfs diff=lfs merge=lfs -text
51
+ *.gif filter=lfs diff=lfs merge=lfs -text
52
+ *.png filter=lfs diff=lfs merge=lfs -text
53
+ *.tiff filter=lfs diff=lfs merge=lfs -text
54
+ # Image files - compressed
55
+ *.jpg filter=lfs diff=lfs merge=lfs -text
56
+ *.jpeg filter=lfs diff=lfs merge=lfs -text
57
+ *.webp filter=lfs diff=lfs merge=lfs -text
58
+ # Video files - compressed
59
+ *.mp4 filter=lfs diff=lfs merge=lfs -text
60
+ *.webm filter=lfs diff=lfs merge=lfs -text
61
+ data/vistec.jsonl filter=lfs diff=lfs merge=lfs -text
LICENSE ADDED
@@ -0,0 +1,202 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ Apache License
3
+ Version 2.0, January 2004
4
+ http://www.apache.org/licenses/
5
+
6
+ TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
7
+
8
+ 1. Definitions.
9
+
10
+ "License" shall mean the terms and conditions for use, reproduction,
11
+ and distribution as defined by Sections 1 through 9 of this document.
12
+
13
+ "Licensor" shall mean the copyright owner or entity authorized by
14
+ the copyright owner that is granting the License.
15
+
16
+ "Legal Entity" shall mean the union of the acting entity and all
17
+ other entities that control, are controlled by, or are under common
18
+ control with that entity. For the purposes of this definition,
19
+ "control" means (i) the power, direct or indirect, to cause the
20
+ direction or management of such entity, whether by contract or
21
+ otherwise, or (ii) ownership of fifty percent (50%) or more of the
22
+ outstanding shares, or (iii) beneficial ownership of such entity.
23
+
24
+ "You" (or "Your") shall mean an individual or Legal Entity
25
+ exercising permissions granted by this License.
26
+
27
+ "Source" form shall mean the preferred form for making modifications,
28
+ including but not limited to software source code, documentation
29
+ source, and configuration files.
30
+
31
+ "Object" form shall mean any form resulting from mechanical
32
+ transformation or translation of a Source form, including but
33
+ not limited to compiled object code, generated documentation,
34
+ and conversions to other media types.
35
+
36
+ "Work" shall mean the work of authorship, whether in Source or
37
+ Object form, made available under the License, as indicated by a
38
+ copyright notice that is included in or attached to the work
39
+ (an example is provided in the Appendix below).
40
+
41
+ "Derivative Works" shall mean any work, whether in Source or Object
42
+ form, that is based on (or derived from) the Work and for which the
43
+ editorial revisions, annotations, elaborations, or other modifications
44
+ represent, as a whole, an original work of authorship. For the purposes
45
+ of this License, Derivative Works shall not include works that remain
46
+ separable from, or merely link (or bind by name) to the interfaces of,
47
+ the Work and Derivative Works thereof.
48
+
49
+ "Contribution" shall mean any work of authorship, including
50
+ the original version of the Work and any modifications or additions
51
+ to that Work or Derivative Works thereof, that is intentionally
52
+ submitted to Licensor for inclusion in the Work by the copyright owner
53
+ or by an individual or Legal Entity authorized to submit on behalf of
54
+ the copyright owner. For the purposes of this definition, "submitted"
55
+ means any form of electronic, verbal, or written communication sent
56
+ to the Licensor or its representatives, including but not limited to
57
+ communication on electronic mailing lists, source code control systems,
58
+ and issue tracking systems that are managed by, or on behalf of, the
59
+ Licensor for the purpose of discussing and improving the Work, but
60
+ excluding communication that is conspicuously marked or otherwise
61
+ designated in writing by the copyright owner as "Not a Contribution."
62
+
63
+ "Contributor" shall mean Licensor and any individual or Legal Entity
64
+ on behalf of whom a Contribution has been received by Licensor and
65
+ subsequently incorporated within the Work.
66
+
67
+ 2. Grant of Copyright License. Subject to the terms and conditions of
68
+ this License, each Contributor hereby grants to You a perpetual,
69
+ worldwide, non-exclusive, no-charge, royalty-free, irrevocable
70
+ copyright license to reproduce, prepare Derivative Works of,
71
+ publicly display, publicly perform, sublicense, and distribute the
72
+ Work and such Derivative Works in Source or Object form.
73
+
74
+ 3. Grant of Patent License. Subject to the terms and conditions of
75
+ this License, each Contributor hereby grants to You a perpetual,
76
+ worldwide, non-exclusive, no-charge, royalty-free, irrevocable
77
+ (except as stated in this section) patent license to make, have made,
78
+ use, offer to sell, sell, import, and otherwise transfer the Work,
79
+ where such license applies only to those patent claims licensable
80
+ by such Contributor that are necessarily infringed by their
81
+ Contribution(s) alone or by combination of their Contribution(s)
82
+ with the Work to which such Contribution(s) was submitted. If You
83
+ institute patent litigation against any entity (including a
84
+ cross-claim or counterclaim in a lawsuit) alleging that the Work
85
+ or a Contribution incorporated within the Work constitutes direct
86
+ or contributory patent infringement, then any patent licenses
87
+ granted to You under this License for that Work shall terminate
88
+ as of the date such litigation is filed.
89
+
90
+ 4. Redistribution. You may reproduce and distribute copies of the
91
+ Work or Derivative Works thereof in any medium, with or without
92
+ modifications, and in Source or Object form, provided that You
93
+ meet the following conditions:
94
+
95
+ (a) You must give any other recipients of the Work or
96
+ Derivative Works a copy of this License; and
97
+
98
+ (b) You must cause any modified files to carry prominent notices
99
+ stating that You changed the files; and
100
+
101
+ (c) You must retain, in the Source form of any Derivative Works
102
+ that You distribute, all copyright, patent, trademark, and
103
+ attribution notices from the Source form of the Work,
104
+ excluding those notices that do not pertain to any part of
105
+ the Derivative Works; and
106
+
107
+ (d) If the Work includes a "NOTICE" text file as part of its
108
+ distribution, then any Derivative Works that You distribute must
109
+ include a readable copy of the attribution notices contained
110
+ within such NOTICE file, excluding those notices that do not
111
+ pertain to any part of the Derivative Works, in at least one
112
+ of the following places: within a NOTICE text file distributed
113
+ as part of the Derivative Works; within the Source form or
114
+ documentation, if provided along with the Derivative Works; or,
115
+ within a display generated by the Derivative Works, if and
116
+ wherever such third-party notices normally appear. The contents
117
+ of the NOTICE file are for informational purposes only and
118
+ do not modify the License. You may add Your own attribution
119
+ notices within Derivative Works that You distribute, alongside
120
+ or as an addendum to the NOTICE text from the Work, provided
121
+ that such additional attribution notices cannot be construed
122
+ as modifying the License.
123
+
124
+ You may add Your own copyright statement to Your modifications and
125
+ may provide additional or different license terms and conditions
126
+ for use, reproduction, or distribution of Your modifications, or
127
+ for any such Derivative Works as a whole, provided Your use,
128
+ reproduction, and distribution of the Work otherwise complies with
129
+ the conditions stated in this License.
130
+
131
+ 5. Submission of Contributions. Unless You explicitly state otherwise,
132
+ any Contribution intentionally submitted for inclusion in the Work
133
+ by You to the Licensor shall be under the terms and conditions of
134
+ this License, without any additional terms or conditions.
135
+ Notwithstanding the above, nothing herein shall supersede or modify
136
+ the terms of any separate license agreement you may have executed
137
+ with Licensor regarding such Contributions.
138
+
139
+ 6. Trademarks. This License does not grant permission to use the trade
140
+ names, trademarks, service marks, or product names of the Licensor,
141
+ except as required for reasonable and customary use in describing the
142
+ origin of the Work and reproducing the content of the NOTICE file.
143
+
144
+ 7. Disclaimer of Warranty. Unless required by applicable law or
145
+ agreed to in writing, Licensor provides the Work (and each
146
+ Contributor provides its Contributions) on an "AS IS" BASIS,
147
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
148
+ implied, including, without limitation, any warranties or conditions
149
+ of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
150
+ PARTICULAR PURPOSE. You are solely responsible for determining the
151
+ appropriateness of using or redistributing the Work and assume any
152
+ risks associated with Your exercise of permissions under this License.
153
+
154
+ 8. Limitation of Liability. In no event and under no legal theory,
155
+ whether in tort (including negligence), contract, or otherwise,
156
+ unless required by applicable law (such as deliberate and grossly
157
+ negligent acts) or agreed to in writing, shall any Contributor be
158
+ liable to You for damages, including any direct, indirect, special,
159
+ incidental, or consequential damages of any character arising as a
160
+ result of this License or out of the use or inability to use the
161
+ Work (including but not limited to damages for loss of goodwill,
162
+ work stoppage, computer failure or malfunction, or any and all
163
+ other commercial damages or losses), even if such Contributor
164
+ has been advised of the possibility of such damages.
165
+
166
+ 9. Accepting Warranty or Additional Liability. While redistributing
167
+ the Work or Derivative Works thereof, You may choose to offer,
168
+ and charge a fee for, acceptance of support, warranty, indemnity,
169
+ or other liability obligations and/or rights consistent with this
170
+ License. However, in accepting such obligations, You may act only
171
+ on Your own behalf and on Your sole responsibility, not on behalf
172
+ of any other Contributor, and only if You agree to indemnify,
173
+ defend, and hold each Contributor harmless for any liability
174
+ incurred by, or claims asserted against, such Contributor by reason
175
+ of your accepting any such warranty or additional liability.
176
+
177
+ END OF TERMS AND CONDITIONS
178
+
179
+ APPENDIX: How to apply the Apache License to your work.
180
+
181
+ To apply the Apache License to your work, attach the following
182
+ boilerplate notice, with the fields enclosed by brackets "[]"
183
+ replaced with your own identifying information. (Don't include
184
+ the brackets!) The text should be enclosed in the appropriate
185
+ comment syntax for the file format. We also recommend that a
186
+ file or class name and description of purpose be included on the
187
+ same "printed page" as the copyright notice for easier
188
+ identification within third-party archives.
189
+
190
+ Copyright [yyyy] [name of copyright owner]
191
+
192
+ Licensed under the Apache License, Version 2.0 (the "License");
193
+ you may not use this file except in compliance with the License.
194
+ You may obtain a copy of the License at
195
+
196
+ http://www.apache.org/licenses/LICENSE-2.0
197
+
198
+ Unless required by applicable law or agreed to in writing, software
199
+ distributed under the License is distributed on an "AS IS" BASIS,
200
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
201
+ See the License for the specific language governing permissions and
202
+ limitations under the License.
NOTICE ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ OpenThai2.0-WordSeg-Benchmark
2
+
3
+ Code (*.py): Apache License 2.0 (LICENSE).
4
+
5
+ The test files in data/ are built from the sources below. Each keeps the licence of its source, and the files
6
+ built from a CC BY-SA source are shared under the same licence. All credit for the sources goes to their authors.
7
+
8
+ - tud.jsonl: UD_Thai-TUD, test split (Sriwirote et al., 2025, The Thai Universal Dependency Treebank, TACL 13).
9
+ Universal Dependencies, revision d4cb5fd3. https://github.com/UniversalDependencies/UD_Thai-TUD
10
+ Licence: CC BY-SA 4.0, https://creativecommons.org/licenses/by-sa/4.0/
11
+ - pud.jsonl, pudA.jsonl, pudB.jsonl: UD_Thai-PUD (Zeman et al., 2017, CoNLL 2017 Shared Task).
12
+ Universal Dependencies, revision 4e0e10fd. https://github.com/UniversalDependencies/UD_Thai-PUD
13
+ Licence: CC BY-SA 3.0, https://creativecommons.org/licenses/by-sa/3.0/
14
+ - ws160.jsonl: Wisesight-160 word segmentation samples (Suriyawongkul et al., 2024, PyThaiNLP/Wisesight
15
+ Sentiment Corpus with Word Tokenization Label, version 1.1).
16
+ PyThaiNLP, revision 6995f993. https://github.com/PyThaiNLP/wisesight-sentiment
17
+ Licence: CC0 1.0, https://creativecommons.org/publicdomain/zero/1.0/
18
+ - vistec.jsonl: VISTEC-TP-TH-2021, test split (Limkonchotiwat et al., 2021, Handling Cross- and Out-of-Domain
19
+ Samples in Thai Word Segmentation, Findings of ACL-IJCNLP 2021), from the OSKut repository, revision be9e16ac.
20
+ https://github.com/mrpeerat/OSKut
21
+ Licence: CC BY-SA 3.0, https://creativecommons.org/licenses/by-sa/3.0/
22
+
23
+ Changes made to these sources (build_testsets.py): each file is converted to natural text, one sentence or post
24
+ per line, with the character offsets of its gold words; a space always separates two words, so a source token
25
+ with a space inside is split; VISTEC's markup is removed; pudA and pudB are the two halves of the PUD test set.
26
+ Defects of the sources are kept as published.
27
+
28
+ Not redistributed (non-commercial licences); build_testsets.py rebuilds them from their original sources:
29
+ - BEST 2009 (Kosawat et al., 2009), NECTEC, CC BY-NC-SA 3.0. https://huggingface.co/datasets/nectec/best2009
30
+ - ORCHID (Sornlertlamvanich et al., 1999), NECTEC, CC BY-NC-SA 3.0.
31
+ https://github.com/wannaphong/corpus_mirror/releases/tag/orchid-v1.0
32
+ - khammoon test set (Sirichotedumrong et al., 2026), Typhoon, CC BY-NC 4.0; its documents come from
33
+ WangchanThaiInstruct (Limkonchotiwat et al., 2025). https://huggingface.co/datasets/typhoon-ai/khammoon-data
34
+
35
+ predictions/: the outputs of the systems listed in README.md. The word files hold line ids and character offsets
36
+ only, no text. The token files exist only for tud, pud, ws160 and vistec; they contain the text of these sets,
37
+ under its licence.
README.md ADDED
@@ -0,0 +1,228 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - th
4
+ license:
5
+ - cc-by-sa-4.0
6
+ - cc-by-sa-3.0
7
+ - cc0-1.0
8
+ - apache-2.0
9
+ pretty_name: OpenThai2.0 WordSeg Benchmark
10
+ task_categories:
11
+ - token-classification
12
+ tags:
13
+ - thai
14
+ - word-segmentation
15
+ - tokenization
16
+ - benchmark
17
+ size_categories:
18
+ - 10K<n<100K
19
+ configs:
20
+ - config_name: tud
21
+ default: true
22
+ data_files:
23
+ - split: test
24
+ path: data/tud.jsonl
25
+ - config_name: pud
26
+ data_files:
27
+ - split: test
28
+ path: data/pud.jsonl
29
+ - config_name: ws160
30
+ data_files:
31
+ - split: test
32
+ path: data/ws160.jsonl
33
+ - config_name: vistec
34
+ data_files:
35
+ - split: test
36
+ path: data/vistec.jsonl
37
+ ---
38
+
39
+ # OpenThai2.0-WordSeg-Benchmark
40
+
41
+ A standard test for Thai word segmentation, version 1.0: seven test sets, one scoring rule, and the results of
42
+ twelve systems. It accompanies the LATTE-WordSeg-Lite paper ([Citation](#citation)).
43
+
44
+ ## Quick start
45
+
46
+ ```bash
47
+ pip install -U huggingface_hub
48
+ hf download iapp/OpenThai2.0-WordSeg-Benchmark --repo-type dataset --local-dir wordseg-bench
49
+ cd wordseg-bench && pip install -r requirements.txt
50
+ ```
51
+
52
+ ```python
53
+ from benchmark import evaluate
54
+
55
+ evaluate(my_segmenter, name="my-segmenter") # my_segmenter(text) -> list of words
56
+ ```
57
+
58
+ This runs your segmenter on the 40,073 lines of the seven test sets. It prints word F1 and boundary F1 on each set
59
+ and their averages, next to the published systems, and saves `results-my-segmenter.json`. From the command line:
60
+ `python benchmark.py --module my_package:segment`, or `python benchmark.py --predictions my_outputs/` with one
61
+ file per set, `SET.jsonl`, holding `{"id": ..., "tokens": [...]}` for each line of `data/SET.jsonl`.
62
+
63
+ **Why the first run downloads files.** Four test sets are in this repository. The other three (BEST, ORCHID and
64
+ khammoon's test) have licences that do not allow sharing them. The first run downloads them from their owners
65
+ (about 50 MB), checks them and builds them here, once.
66
+
67
+ ## Test sets
68
+
69
+ | Set | Source | Licence | Included |
70
+ |---|---|---|---|
71
+ | `tud` | [UD Thai-TUD](https://github.com/UniversalDependencies/UD_Thai-TUD), test split | CC BY-SA 4.0 | yes |
72
+ | `pud` | [UD Thai-PUD](https://github.com/UniversalDependencies/UD_Thai-PUD) | CC BY-SA 3.0 | yes |
73
+ | `ws160` | [Wisesight-160](https://github.com/PyThaiNLP/wisesight-sentiment) | CC0 1.0 | yes |
74
+ | `vistec` | [VISTEC-TP-TH-2021](https://github.com/mrpeerat/OSKut), test split | CC BY-SA 3.0 | yes |
75
+ | `bestdd` | [BEST 2009](https://huggingface.co/datasets/nectec/best2009), our split | CC BY-NC-SA 3.0 | no |
76
+ | `orchidndd` | [ORCHID](https://github.com/wannaphong/corpus_mirror/releases/tag/orchid-v1.0), our split | CC BY-NC-SA 3.0 | no |
77
+ | `khm` | [khammoon data](https://huggingface.co/datasets/typhoon-ai/khammoon-data), test split | CC BY-NC 4.0 | no |
78
+
79
+ How the sets were made from their sources is listed in [NOTICE](NOTICE) and explained in the paper.
80
+
81
+ ## Results
82
+
83
+ Average F1 (%) over the seven test sets:
84
+
85
+ <!-- results:start -->
86
+ | System | Word F1 | Boundary F1 |
87
+ |---|---|---|
88
+ | **LATTE-WordSeg-Lite** | **85.6** | **93.3** |
89
+ | SEFR CUT, `best` | 84.5 | 92.6 |
90
+ | DeepCut | 84.5 | 92.6 |
91
+ | AttaCut-SC | 84.0 | 92.4 |
92
+ | SEFR CUT, `ws1000` | 84.0 | 92.5 |
93
+ | AttaCut-C | 83.2 | 91.9 |
94
+ | khammoon | 81.8 | 91.3 |
95
+ | OSKut, `scads` | 79.4 | 89.9 |
96
+ | TLTK | 75.5 | 87.9 |
97
+ | OSKut, `ws` | 75.3 | 88.1 |
98
+ | PyThaiNLP nlpO3 | 70.9 | 85.6 |
99
+ | PyThaiNLP newmm | 70.9 | 85.7 |
100
+ <!-- results:end -->
101
+
102
+ LATTE-WordSeg-Lite with its default unit, `short` (`long`: 83.1 and 92.0, `minimal`: 82.1 and 91.5). khammoon
103
+ segments morphemes, not words: six of the seven sets count words, so compare it with `minimal`. Boundary F1 scores
104
+ each cut between two characters, so it is higher than word F1 and closer between systems. Both scores on each set are
105
+ in `results/expected.json`, and every system's output is in `predictions/`. The model is at
106
+ [iapp/OpenThai2.0-LATTE-WordSeg-Lite](https://huggingface.co/iapp/OpenThai2.0-LATTE-WordSeg-Lite) and in the
107
+ Python package [iapp-ai](https://pypi.org/project/iapp-ai/).
108
+
109
+ ## How scores are computed
110
+
111
+ - **Word F1:** a word is correct when a gold word has the same start and end. **Boundary F1:** a cut between two
112
+ characters is correct when gold cuts there too. A space always separates words and is never scored: a token
113
+ that holds a space counts as the words between its spaces.
114
+ - Every system gets the same lines and the same rule, so its numbers can be compared with the table above.
115
+
116
+ ## Data format
117
+
118
+ Each line of `data/SET.jsonl` is one test line, with the start and end of every gold word:
119
+
120
+ ```json
121
+ {"id": 0, "text": "ยุ้ยจะบอกว่ามันนานแล้ว แต่เรารักกัน", "gold": [[0, 4], [4, 6], [6, 9], [9, 12], [12, 15], [15, 18], [18, 22], [23, 26], [26, 29], [29, 32], [32, 35]]}
122
+ ```
123
+
124
+ The four included sets also load with `datasets`: `load_dataset("iapp/OpenThai2.0-WordSeg-Benchmark", "tud")`.
125
+
126
+ ## Reproduce the paper
127
+
128
+ ```bash
129
+ python fetch_sources.py # every original file at its pinned revision, SHA-256 checked
130
+ python build_testsets.py # every test file, checked against checksums.json
131
+ python score.py table # every score of the paper, checked against results/expected.json
132
+ ```
133
+
134
+ ## Licences
135
+
136
+ - **Test files:** the licence of their source (table above). Files built from a CC BY-SA corpus are shared under
137
+ the same licence. [NOTICE](NOTICE) gives the attribution of each source.
138
+ - **Code** (`*.py`): Apache-2.0 ([LICENSE](LICENSE)).
139
+ - **Predictions:** the word offsets contain no text; the token files of the four included sets contain their
140
+ text, under its licence.
141
+
142
+ ## Citation
143
+
144
+ If you use the benchmark, please cite the paper and the source of each test set you use.
145
+
146
+ ```bibtex
147
+ @misc{chayintr2026lattewordseglite,
148
+ title = {{LATTE-WordSeg-Lite}: Multi-Granularity {Thai} Word Segmentation on a {CPU}, with a Reproducible Benchmark},
149
+ author = {Chay-intr, Thodsaporn and Viriyayudhakorn, Kobkrit and Theeramunkong, Thanaruk},
150
+ year = {2026}
151
+ }
152
+ ```
153
+
154
+ <details>
155
+ <summary>BibTeX of the sources</summary>
156
+
157
+ ```bibtex
158
+ @article{sriwirote2025tud,
159
+ title = {The {Thai} {Universal} {Dependency} Treebank},
160
+ author = {Sriwirote, Panyut and Leong, Wei Qi and Polpanumas, Charin and Thanyawong, Santhawat and Tjhi, William Chandra and Aroonmanakun, Wirote and Rutherford, Attapol T.},
161
+ journal = {Transactions of the Association for Computational Linguistics},
162
+ volume = {13},
163
+ pages = {376--391},
164
+ year = {2025},
165
+ doi = {10.1162/tacl_a_00745}
166
+ }
167
+
168
+ @inproceedings{zeman2017conll,
169
+ title = {{CoNLL} 2017 Shared Task: Multilingual Parsing from Raw Text to {Universal Dependencies}},
170
+ author = {Zeman, Daniel and others},
171
+ booktitle = {Proceedings of the {CoNLL} 2017 Shared Task: Multilingual Parsing from Raw Text to {Universal Dependencies}},
172
+ pages = {1--19},
173
+ year = {2017}
174
+ }
175
+
176
+ @misc{suriyawongkul2024wisesight,
177
+ title = {{PyThaiNLP/Wisesight} Sentiment Corpus with Word Tokenization Label},
178
+ author = {Suriyawongkul, Arthit and Chuangsuwanich, Ekapol and Chormai, Pattarawat and Chantarapratin, Nitchakarn and Prasertsom, Ponrawee and Sawatphol, Jitkapat and Yamada, Nozomi and Rutherford, Attapol and Polpanumas, Charin and Udomcharoenchaikit, Can},
179
+ publisher = {Zenodo},
180
+ note = {Version 1.1},
181
+ doi = {10.5281/zenodo.3457446},
182
+ year = {2024}
183
+ }
184
+
185
+ @inproceedings{limkonchotiwat2021oskut,
186
+ title = {Handling Cross- and Out-of-Domain Samples in {Thai} Word Segmentation},
187
+ author = {Limkonchotiwat, Peerat and Phatthiyaphaibun, Wannaphong and Sarwar, Raheem and Chuangsuwanich, Ekapol and Nutanong, Sarana},
188
+ booktitle = {Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021},
189
+ pages = {1003--1016},
190
+ year = {2021}
191
+ }
192
+
193
+ @inproceedings{kosawat2009best,
194
+ title = {{BEST} 2009: {Thai} Word Segmentation Software Contest},
195
+ author = {Kosawat, Krit and Boriboon, Monthika and Chootrakool, Patcharika and Chotimongkol, Ananlada and Klaithin, Supon and Kongyoung, Sarawoot and Kriengket, Kanyanut and Phaholphinyo, Sitthaa and Purodakananda, Sumonmas and Thanakulwarapas, Tipraporn and Wutiwiwatchai, Chai},
196
+ booktitle = {Eighth International Symposium on Natural Language Processing},
197
+ pages = {83--88},
198
+ year = {2009}
199
+ }
200
+
201
+ @article{sornlertlamvanich1999orchid,
202
+ title = {Building a {Thai} Part-of-Speech Tagged Corpus ({ORCHID})},
203
+ author = {Sornlertlamvanich, Virach and Takahashi, Naoto and Isahara, Hitoshi},
204
+ journal = {Journal of the Acoustical Society of Japan (E)},
205
+ volume = {20},
206
+ number = {3},
207
+ pages = {189--198},
208
+ year = {1999},
209
+ doi = {10.1250/ast.20.189}
210
+ }
211
+
212
+ @article{khammoon2026,
213
+ title = {khammoon: From a Linguistic Guideline to a {Thai} Morpheme Segmenter},
214
+ author = {Sirichotedumrong, Warit and Na-Thalang, Adisai and Wittayasakpan, Chanakan and Phatcharoen, Kritsadha and Pipatanakul, Kunat},
215
+ journal = {arXiv preprint arXiv:2611.13012},
216
+ year = {2026}
217
+ }
218
+
219
+ @inproceedings{limkonchotiwat2025wangchan,
220
+ title = {{WangchanThaiInstruct}: An Instruction-Following Dataset for Culture-Aware, Multitask, and Multi-domain Evaluation in {Thai}},
221
+ author = {Limkonchotiwat, Peerat and Tuchinda, Pume and Lowphansirikul, Lalita and Nonesung, Surapon and Tasawong, Panuthep and Aji, Alham Fikri and Udomcharoenchaikit, Can and Nutanong, Sarana},
222
+ booktitle = {Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing},
223
+ pages = {3535--3558},
224
+ year = {2025}
225
+ }
226
+ ```
227
+
228
+ </details>
benchmark.py ADDED
@@ -0,0 +1,119 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ """OpenThai2.0-WordSeg-Benchmark 1.0: score a Thai word segmenter on the seven test sets in one call.
2
+
3
+ from benchmark import evaluate
4
+ evaluate(segment, name="my-segmenter") # segment(text) -> list of tokens
5
+
6
+ python benchmark.py --module my_package:segment --name my-segmenter
7
+ python benchmark.py --predictions my_outputs/ --name my-segmenter # my_outputs/SET.jsonl, {"id", "tokens"}
8
+
9
+ It prints word F1 and boundary F1 on every set and their averages, next to the published systems, and saves
10
+ results-NAME.json. The first run builds the three test sets that may not be shared (BEST, ORCHID, khammoon's test):
11
+ it downloads their original files from their owners (about 50 MB), checks their SHA-256 and builds them, once.
12
+ """
13
+ from __future__ import annotations
14
+
15
+ import argparse
16
+ import hashlib
17
+ import importlib
18
+ import json
19
+ import sys
20
+ import time
21
+ from pathlib import Path
22
+
23
+ HERE = Path(__file__).resolve().parent
24
+ sys.path.insert(0, str(HERE))
25
+ from thaiws import boundary_counts, f1, load_set, read_jsonl, word_counts # noqa: E402
26
+
27
+ BENCHMARK, VERSION = "OpenThai2.0-WordSeg-Benchmark", "1.0"
28
+ SETS = [("tud", "TUD"), ("pud", "PUD"), ("ws160", "WS160"), ("vistec", "VISTEC"), ("bestdd", "BEST"),
29
+ ("orchidndd", "ORCHID"), ("khm", "khm")]
30
+ DATA, RAW = HERE / "data", HERE / "raw"
31
+
32
+
33
+ def prepare() -> None:
34
+ """Build, once, the test sets whose licence does not allow sharing them."""
35
+ missing = [s for s, _ in SETS if not (DATA / f"{s}.jsonl").exists()]
36
+ if not missing:
37
+ return
38
+ import build_testsets
39
+ import fetch_sources
40
+ groups = {"bestdd": build_testsets.build_best, "orchidndd": build_testsets.build_orchid,
41
+ "khm": lambda raw: build_testsets.build_khm(raw, dedup=False)}
42
+ unknown = [s for s in missing if s not in groups]
43
+ if unknown:
44
+ raise SystemExit(f"data/{unknown[0]}.jsonl is missing: download the benchmark again")
45
+ print("First run: BEST, ORCHID and khammoon's test may not be shared, so they are downloaded from their "
46
+ "owners and built here (about 50 MB, once).", flush=True)
47
+ bad = fetch_sources.fetch(RAW, sets=set(missing), quiet=True)
48
+ if bad:
49
+ raise SystemExit(f"these source files differ from the versions the benchmark was built from: {bad}")
50
+ want = json.loads((HERE / "checksums.json").read_text())
51
+ for s in missing:
52
+ for name, rows in groups[s](RAW).items():
53
+ blob = build_testsets.dump(rows)
54
+ if hashlib.sha256(blob).hexdigest() != want[f"{name}.jsonl"]:
55
+ raise SystemExit(f"{name}: the built file differs from the published one")
56
+ (DATA / f"{name}.jsonl").write_bytes(blob)
57
+ print("Test sets ready.\n", flush=True)
58
+
59
+
60
+ def evaluate(segment=None, name: str = "my-segmenter", predictions: str | Path | None = None, batch: bool = False,
61
+ save: bool = True) -> dict:
62
+ """Score `segment` (a function: text -> list of tokens; with batch=True: list of texts -> list of token lists),
63
+ or the files of a `predictions` folder (SET.jsonl, one {"id", "tokens"} per line of data/SET.jsonl)."""
64
+ if (segment is None) == (predictions is None):
65
+ raise ValueError("give either a segment function or a predictions folder")
66
+ prepare()
67
+ word, boundary = {}, {}
68
+ print(f"{BENCHMARK} {VERSION}: {name}\n", flush=True)
69
+ for s, label in SETS:
70
+ rows = load_set(s)
71
+ t0 = time.time()
72
+ if predictions is not None:
73
+ preds = read_jsonl(Path(predictions) / f"{s}.jsonl")
74
+ if [p["id"] for p in preds] != [r["id"] for r in rows]:
75
+ raise SystemExit(f"{predictions}/{s}.jsonl: one line per line of data/{s}.jsonl, in the same order")
76
+ else:
77
+ texts = [r["text"] for r in rows]
78
+ tokens = list(segment(texts)) if batch else [list(segment(t)) for t in texts]
79
+ preds = [{"id": r["id"], "tokens": list(t)} for r, t in zip(rows, tokens)]
80
+ word[s] = round(float(f1(word_counts(rows, preds).sum(0))), 2)
81
+ boundary[s] = round(float(f1(boundary_counts(rows, preds).sum(0))), 2)
82
+ print(f" {label:7s} word F1 {word[s]:6.2f} boundary F1 {boundary[s]:6.2f} ({len(rows):,} lines, "
83
+ f"{time.time() - t0:.0f} s)", flush=True)
84
+ average = round(sum(word.values()) / len(word), 2)
85
+ average_boundary = round(sum(boundary.values()) / len(boundary), 2)
86
+ print(f" {'Average':7s} word F1 {average:6.2f} boundary F1 {average_boundary:6.2f}\n")
87
+ expected = json.loads((HERE / "results" / "expected.json").read_text())
88
+ board = sorted([(v["avg7"], expected["boundary_f1_sets"][k]["avg7"], k) for k, v in expected["averages"].items()
89
+ if "(unit rule)" not in k] + [(average, average_boundary, f"> {name}")], reverse=True)
90
+ print("Average F1 with the published systems:\n word boundary")
91
+ for v, b, k in board:
92
+ print(f" {v:6.2f} {b:6.2f} {k}")
93
+ result = {"benchmark": BENCHMARK, "version": VERSION, "system": name, "date": time.strftime("%Y-%m-%d"),
94
+ "word_f1": word, "boundary_f1": boundary, "average_word_f1": average,
95
+ "average_boundary_f1": average_boundary}
96
+ if save:
97
+ path = Path(f"results-{name}.json")
98
+ path.write_text(json.dumps(result, indent=1, ensure_ascii=False) + "\n", encoding="utf-8")
99
+ print(f"\nSaved {path}")
100
+ return result
101
+
102
+
103
+ def main() -> None:
104
+ ap = argparse.ArgumentParser(description=__doc__.split("\n\n")[0])
105
+ src = ap.add_mutually_exclusive_group(required=True)
106
+ src.add_argument("--module", help="MODULE:FUNCTION, a function text -> list of tokens")
107
+ src.add_argument("--predictions", help="a folder with SET.jsonl files of {\"id\", \"tokens\"}")
108
+ ap.add_argument("--name", default="my-segmenter")
109
+ ap.add_argument("--batch", action="store_true", help="the function takes a list of texts")
110
+ args = ap.parse_args()
111
+ segment = None
112
+ if args.module:
113
+ mod, _, fn = args.module.partition(":")
114
+ segment = getattr(importlib.import_module(mod), fn)
115
+ evaluate(segment, name=args.name, predictions=args.predictions, batch=args.batch)
116
+
117
+
118
+ if __name__ == "__main__":
119
+ main()
build_testsets.py ADDED
@@ -0,0 +1,289 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ """Rebuild the Thai word-segmentation test sets from their original sources, byte for byte.
2
+
3
+ python fetch_sources.py # downloads the original files into raw/ (see sources.json)
4
+ python build_testsets.py # raw/ -> data/*.jsonl, then checks every file against checksums.json
5
+
6
+ Requirements: Python 3.9+, pandas with pyarrow (BEST is a parquet file). Every set is natural text: one row per
7
+ line, {"id", "text", "gold": [[start, end], ...]} where the gold words are the maximal non-whitespace runs of
8
+ the source's tokens (a space always separates words; whitespace is never scored). Splits of BEST and ORCHID,
9
+ which have no official test split with gold boundaries, are seeded 80/10/10 shuffles (seed 112, as LATTE).
10
+
11
+ Sets: tud, pud (+ pudA, pudB: halves for choosing a unit), ws1000, ws160, vistec, best, bestdd, bestdev, orchidnho,
12
+ orchidndd, orchiddev, khm, khmdd. `dd`: test lines that (nearly) repeat the training text are removed: equal
13
+ to a line of it after removing whitespace, or at least 80% of their 12-character shingles (every 4th) found in
14
+ it. bestdd and orchidndd against the train+dev part of their corpus, khmdd against khammoon's training text.
15
+ Exact string matching throughout, so the result is deterministic.
16
+ """
17
+ from __future__ import annotations
18
+
19
+ import argparse
20
+ import gzip
21
+ import hashlib
22
+ import json
23
+ import random
24
+ import re
25
+ from pathlib import Path
26
+
27
+ HERE = Path(__file__).resolve().parent
28
+ SEED = 112
29
+ K, STRIDE, NEAR = 12, 4, 0.8
30
+ WS = re.compile(r"\s+")
31
+ RUN = re.compile(r"\S+")
32
+ ORCHID_TAGS = {"<space>": " ", "<minus>": "-", "<full_stop>": ".", "<left_parenthesis>": "(",
33
+ "<right_parenthesis>": ")", "<quotation>": '"', "<comma>": ",", "<slash>": "/", "<equal>": "=",
34
+ "<colon>": ":", "<asterisk>": "*", "<plus>": "+", "<greater_than>": ">", "<semi_colon>": ";",
35
+ "<less_than>": "<", "<apostrophe>": "'", "<dollar>": "$", "<number>": "#", "<ampersand>": "&",
36
+ "<question_mark>": "?", "<circumflex_accent>": "^", "<left_curly_bracket>": "{",
37
+ "<right_curly_bracket>": "}", "<exclamation>": "!", "<at_mark>": "@"}
38
+
39
+
40
+ def spans_of(tokens: list[str]) -> tuple[str, list[list[int]]]:
41
+ """Text = the tokens joined; gold = every maximal non-whitespace run inside a token."""
42
+ text, gold, pos = "", [], 0
43
+ for tok in tokens:
44
+ gold.extend([pos + m.start(), pos + m.end()] for m in RUN.finditer(tok))
45
+ text += tok
46
+ pos += len(tok)
47
+ return text, gold
48
+
49
+
50
+ def shuffled(n: int) -> list[int]:
51
+ idx = list(range(n))
52
+ random.Random(SEED).shuffle(idx)
53
+ return idx
54
+
55
+
56
+ def repeats(tests: list[str], refs: list[str]) -> list[bool]:
57
+ """Which test texts (nearly) repeat the reference texts (whitespace ignored)."""
58
+ tests = [WS.sub("", t) for t in tests]
59
+ found = {t[i:i + K]: False for t in tests for i in range(0, len(t) - K + 1, STRIDE)}
60
+ lines = set(tests)
61
+ exact = set()
62
+ for r in refs:
63
+ r = WS.sub("", r)
64
+ if r in lines:
65
+ exact.add(r)
66
+ for i in range(len(r) - K + 1):
67
+ if r[i:i + K] in found:
68
+ found[r[i:i + K]] = True
69
+ out = []
70
+ for t in tests:
71
+ g = [t[i:i + K] for i in range(0, len(t) - K + 1, STRIDE)]
72
+ out.append(t in exact or (bool(g) and sum(found[x] for x in g) / len(g) >= NEAR))
73
+ return out
74
+
75
+
76
+ # ---------------------------------------------------------------- sources
77
+ def ud(path: Path) -> list[dict]:
78
+ """CoNLL-U -> natural text: a space follows a token unless SpaceAfter=No; gold = the word forms."""
79
+ sents, toks = [], []
80
+ for line in path.read_text(encoding="utf-8").splitlines():
81
+ if not line.strip():
82
+ if toks:
83
+ sents.append(toks)
84
+ toks = []
85
+ continue
86
+ if line.startswith("#"):
87
+ continue
88
+ cols = line.split("\t")
89
+ if "-" in cols[0] or "." in cols[0]:
90
+ continue
91
+ toks.append(cols[1])
92
+ if "SpaceAfter=No" not in cols[9]:
93
+ toks.append(" ")
94
+ if toks:
95
+ sents.append(toks)
96
+ rows = []
97
+ for toks in sents:
98
+ while toks and toks[-1] == " ":
99
+ toks.pop()
100
+ text, gold = spans_of(toks)
101
+ rows.append({"text": text, "gold": gold})
102
+ return rows
103
+
104
+
105
+ def wisesight(path: Path) -> list[dict]:
106
+ """'|'-separated tokens, space tokens included."""
107
+ opener = gzip.open if path.suffix == ".gz" else open
108
+ with opener(path, "rt", encoding="utf-8") as f:
109
+ return [dict(zip(("text", "gold"), spans_of(line.rstrip("\n").split("|")))) for line in f if line.strip()]
110
+
111
+
112
+ def vistec(path: Path) -> list[dict]:
113
+ """VISTEC-TP-TH-2021: '|'-separated tokens with markup (<ne>, <compound>, <msp value=...>): the markup is
114
+ removed and the written surface kept."""
115
+ tag = re.compile(r"<[^>]*>")
116
+ with open(path, encoding="utf-8") as f:
117
+ return [dict(zip(("text", "gold"), spans_of(tag.sub("", line.rstrip("\n")).split("|"))))
118
+ for line in f if line.strip()]
119
+
120
+
121
+ def best_lines(path: Path) -> list[list[str]]:
122
+ """BEST 2009 (nectec/best2009 train.parquet): characters with word-start flags -> words, spaces kept."""
123
+ import pandas as pd
124
+ df = pd.read_parquet(path)
125
+ lines = []
126
+ for chars, begins in zip(df["char"], df["is_beginning"]):
127
+ words, cur = [], ""
128
+ for ch, beg in zip(chars, begins):
129
+ if beg == 1 and cur:
130
+ words.append(cur)
131
+ cur = ""
132
+ cur += ch
133
+ if cur:
134
+ words.append(cur)
135
+ if any(w.strip() for w in words):
136
+ lines.append(words)
137
+ return lines
138
+
139
+
140
+ def orchid_sentences(path: Path) -> list[list[str]]:
141
+ """ORCHID 1997 (orchid97.crp.utf): every Thai sentence from its word lines, tags mapped back to their
142
+ characters and <space> to a space (English sentences, '%E ...', have no word lines and are skipped)."""
143
+ outside, raw_text, in_words = 0, 1, 2
144
+ out, raw, toks, state = [], "", [], outside
145
+ for line in path.open(encoding="utf-8"):
146
+ line = line.rstrip("\n")
147
+ if state == outside:
148
+ if line.startswith("#") and line[1:].isdigit():
149
+ state, raw = raw_text, ""
150
+ elif state == raw_text:
151
+ raw += line
152
+ if line.endswith("//"):
153
+ state, toks = (outside if raw.startswith("%E") else in_words), []
154
+ elif line == "//":
155
+ if any(w and w != "<space>" for w in (t.replace(" ", "") for t in toks)):
156
+ out.append([ORCHID_TAGS.get(t, t) for t in toks])
157
+ state = outside
158
+ else:
159
+ toks.append(line.rsplit("/", 1)[0])
160
+ return out
161
+
162
+
163
+ def khm(path: Path) -> list[dict]:
164
+ """khammoon's human test documents, split at newlines into lines (no gold word crosses a newline)."""
165
+ out = []
166
+ for d, line in enumerate(path.open(encoding="utf-8")):
167
+ r = json.loads(line)
168
+ text = r["text"]
169
+ assert "".join(r["tokens"]) == text
170
+ _, spans = spans_of(r["tokens"])
171
+ start, k = 0, 0
172
+ for part in text.split("\n"):
173
+ end = start + len(part)
174
+ gold = []
175
+ while k < len(spans) and spans[k][1] <= end:
176
+ gold.append([spans[k][0] - start, spans[k][1] - start])
177
+ k += 1
178
+ if part.strip():
179
+ out.append({"doc": d, "text": part, "gold": gold})
180
+ start = end + 1
181
+ assert k == len(spans)
182
+ return out
183
+
184
+
185
+ # ---------------------------------------------------------------- build
186
+ def build_open(raw: Path) -> dict[str, list[dict]]:
187
+ """The sets whose licence allows sharing them (they are in data/ already)."""
188
+ sets: dict[str, list[dict]] = {}
189
+ sets["tud"] = ud(raw / "ud/UD_Thai-TUD/th_tud-ud-test.conllu")
190
+ sets["pud"] = ud(raw / "ud/UD_Thai-PUD/th_pud-ud-test.conllu")
191
+ order = shuffled(len(sets["pud"]))
192
+ half = len(order) // 2
193
+ for name, part in (("pudA", sorted(order[:half])), ("pudB", sorted(order[half:]))):
194
+ sets[name] = [{"src_id": i, **sets["pud"][i]} for i in part]
195
+ sets["ws1000"] = wisesight(raw / "wisesight1000/wisesight-1000-samples-tokenised.label.gz")
196
+ sets["ws160"] = wisesight(raw / "wisesight1000/wisesight-160-samples-tokenised.label")
197
+ sets["vistec"] = vistec(raw / "vistec/test.txt")
198
+ return sets
199
+
200
+
201
+ def build_best(raw: Path) -> dict[str, list[dict]]:
202
+ lines = best_lines(raw / "best2009/train.parquet")
203
+ idx = shuffled(len(lines))
204
+ cut1, cut2 = int(0.8 * len(idx)), int(0.9 * len(idx))
205
+ rows = [dict(zip(("text", "gold"), spans_of(lines[i]))) for i in idx[cut2:]]
206
+ sets = {"best": rows, "bestdev": [dict(zip(("text", "gold"), spans_of(lines[i]))) for i in idx[cut1:cut2]]}
207
+ dup = repeats([r["text"] for r in rows], ["".join(lines[i]) for i in idx[:cut2]])
208
+ sets["bestdd"] = [{"src_id": k, **r} for k, (r, d) in enumerate(zip(rows, dup)) if not d]
209
+ return sets
210
+
211
+
212
+ def build_orchid(raw: Path) -> dict[str, list[dict]]:
213
+ sents = orchid_sentences(raw / "orchid/orchid97.crp.utf")
214
+ nat = [spans_of(t) for t in sents]
215
+ idx = shuffled(len(nat))
216
+ cut1, cut2 = int(0.8 * len(idx)), int(0.9 * len(idx))
217
+ test = idx[cut2:]
218
+ sets = {"orchidnho": [{"src_id": i, "text": nat[i][0], "gold": nat[i][1]} for i in test],
219
+ "orchiddev": [{"src_id": i, "text": nat[i][0], "gold": nat[i][1]} for i in idx[cut1:cut2]]}
220
+ dup = repeats([nat[i][0] for i in test], [nat[i][0] for i in idx[:cut2]])
221
+ sets["orchidndd"] = [{"src_id": k, "text": nat[i][0], "gold": nat[i][1]}
222
+ for k, (i, d) in enumerate(zip(test, dup)) if not d]
223
+ return sets
224
+
225
+
226
+ KHMDD_TEXT = ("llm_labelled/best2009.jsonl", "llm_labelled/wangchan.jsonl", "dev/best2009.jsonl")
227
+
228
+
229
+ def build_khm(raw: Path, dedup: bool | None = None) -> dict[str, list[dict]]:
230
+ """khammoon's test; khmdd too when its training text was fetched (780 MB, only needed for khmdd)."""
231
+ sets = {"khm": khm(raw / "khammoon/test/wangchan.jsonl")}
232
+ if dedup is None:
233
+ dedup = all((raw / "khammoon" / f).is_file() for f in KHMDD_TEXT)
234
+ if dedup:
235
+ training = []
236
+ for f in KHMDD_TEXT:
237
+ for line in (raw / "khammoon" / f).open(encoding="utf-8"):
238
+ training.extend(json.loads(line)["text"].split("\n"))
239
+ dup = repeats([r["text"] for r in sets["khm"]], training)
240
+ sets["khmdd"] = [{"src_id": k, **r} for k, (r, d) in enumerate(zip(sets["khm"], dup)) if not d]
241
+ return sets
242
+
243
+
244
+ def build(raw: Path) -> dict[str, list[dict]]:
245
+ """Every set (maintainers): the same files as before the split into groups."""
246
+ return {**build_open(raw), **build_best(raw), **build_orchid(raw), **build_khm(raw)}
247
+
248
+
249
+ KEY_ORDER = ("id", "src_id", "doc", "text", "gold")
250
+
251
+
252
+ def dump(rows: list[dict]) -> bytes:
253
+ out = []
254
+ for i, r in enumerate(rows):
255
+ r = {"id": i, **r}
256
+ out.append(json.dumps({k: r[k] for k in KEY_ORDER if k in r}, ensure_ascii=False) + "\n")
257
+ return "".join(out).encode("utf-8")
258
+
259
+
260
+ def main() -> None:
261
+ ap = argparse.ArgumentParser()
262
+ ap.add_argument("--raw", type=Path, default=HERE / "raw")
263
+ ap.add_argument("--out", type=Path, default=HERE / "data")
264
+ ap.add_argument("--write-checksums", action="store_true", help="record the built files' SHA-256 (maintainers)")
265
+ args = ap.parse_args()
266
+ sets = build(args.raw)
267
+ args.out.mkdir(parents=True, exist_ok=True)
268
+ sums = {}
269
+ for name, rows in sets.items():
270
+ blob = dump(rows)
271
+ (args.out / f"{name}.jsonl").write_bytes(blob)
272
+ sums[f"{name}.jsonl"] = hashlib.sha256(blob).hexdigest()
273
+ print(f"{name}: {len(rows):,} lines, {sum(len(r['gold']) for r in rows):,} gold words")
274
+ path = HERE / "checksums.json"
275
+ if args.write_checksums:
276
+ path.write_text(json.dumps(sums, indent=1) + "\n")
277
+ print(f"wrote {path.name}")
278
+ return
279
+ want = json.loads(path.read_text())
280
+ optional = {"khmdd.jsonl"} # built only when khammoon's training text was fetched
281
+ if "khmdd.jsonl" not in sums:
282
+ print("khmdd skipped: it needs khammoon's training text (python fetch_sources.py, without --no-khmdd)")
283
+ bad = [n for n in want if (n in sums or n not in optional) and sums.get(n) != want[n]]
284
+ print("all files identical to the published ones" if not bad else f"DIFFERENT from the published files: {bad}")
285
+ raise SystemExit(1 if bad else 0)
286
+
287
+
288
+ if __name__ == "__main__":
289
+ main()
checksums.json ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "tud.jsonl": "837ecf03a4ef55cb5c44bfdef794b1ba377656fccaf9446741431ad5fc93d6a3",
3
+ "pud.jsonl": "1deee272f6faa1e9428572e4f93000c7e874050f83d8217694f7e15a075284da",
4
+ "pudA.jsonl": "841f55f3cbbaf57dbf278d731da54ac85b23f06396626c0705cd83e4801ca96b",
5
+ "pudB.jsonl": "a00511f036443bb84c174cf11486d3c3f421a42063eb76060ef534636baac2d4",
6
+ "ws1000.jsonl": "330fb3cb6fda5b5ce5ca90ac3ea7e7145319d0987cd5defd54ed74d9e4b10abb",
7
+ "ws160.jsonl": "f16ea5744619dda6deec62a1976b2253db3a4255108c59d9d5ae98f1b303ed5c",
8
+ "vistec.jsonl": "673d4fc87c0491e4f5ec3668c9f2981f249bbab2edcf8345aff34907ad9bae8f",
9
+ "best.jsonl": "04422900726a655d9685fd2458ee152e44a88a46d39c15c68fd1ca0ef3e80f7f",
10
+ "bestdev.jsonl": "ee96e1697c36d1ae84a99a424077e404f284f9fe7b04b5958b912d61cf79cd7a",
11
+ "bestdd.jsonl": "6058a8cf99b2094c1fe1b6960ba099739a470a4679108c44f701961b9aadf225",
12
+ "orchidnho.jsonl": "37b25823578041f827ffed0e0d953ff0bb0938a54422dd1ffcc509dbb3f6023a",
13
+ "orchiddev.jsonl": "114c338304f29cc2df41ea094304f7bd32df6a217c074950d0cffe20e443f473",
14
+ "orchidndd.jsonl": "a865226fa2d5e300f1cd5550afba263afd1f1176e99b56f877c4bb0f1742aef3",
15
+ "khm.jsonl": "ee34c922705adb24155633670cfb03ca34f357d802d29f06a6be49fc04e22e03",
16
+ "khmdd.jsonl": "9441f9213640d7fb6fda9b575ba7de80240059cf9469ce454bbc79220fc90c81"
17
+ }
data/pud.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
data/pudA.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
data/pudB.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
data/tud.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
data/vistec.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:673d4fc87c0491e4f5ec3668c9f2981f249bbab2edcf8345aff34907ad9bae8f
3
+ size 15572765
data/ws160.jsonl ADDED
@@ -0,0 +1,160 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {"id": 0, "text": "Eucerin pro acne ค่ะ ใช้แล้วสิวขึ้นเพิ่มทุกวัน มาดูกันนะคะ ว่าจัดการปัญหาสิวใน7วันได้รึมั่ยยยยยยยย ล่าสุดไปล้างหน้…", "gold": [[0, 7], [8, 11], [12, 16], [17, 20], [21, 24], [24, 28], [28, 31], [31, 35], [35, 40], [40, 46], [47, 49], [49, 51], [51, 54], [54, 56], [56, 58], [59, 62], [62, 65], [65, 68], [68, 73], [73, 76], [76, 78], [78, 79], [79, 82], [82, 85], [85, 87], [87, 98], [99, 102], [102, 105], [105, 107], [107, 111], [111, 115]]}
2
+ {"id": 1, "text": "แพงเว่อร์ เบียร์ช้างต้นทุนขวดละไม่ถึง 50 ขาย 120 😰😰😰์", "gold": [[0, 3], [3, 9], [10, 16], [16, 20], [20, 26], [26, 29], [29, 31], [31, 34], [34, 37], [38, 40], [41, 44], [45, 48], [49, 50], [50, 51], [51, 53]]}
3
+ {"id": 2, "text": "ไม่อยากก้ไม่ต้องกินเพราะตังของคุนแต่จะมาพูดให้คนที่เขายังกินอยุ่รุ้สึกแย่กันเพื่อ??ต่างคนต่างมีดุลพินิจไม่เหมือนกันปะ", "gold": [[0, 3], [3, 7], [7, 9], [9, 12], [12, 16], [16, 19], [19, 24], [24, 27], [27, 30], [30, 33], [33, 36], [36, 38], [38, 40], [40, 43], [43, 46], [46, 48], [48, 51], [51, 54], [54, 57], [57, 60], [60, 64], [64, 67], [67, 70], [70, 73], [73, 76], [76, 81], [81, 82], [82, 83], [83, 87], [87, 89], [89, 93], [93, 95], [95, 98], [98, 103], [103, 106], [106, 112], [112, 115], [115, 117]]}
4
+ {"id": 3, "text": "กูเริ่มรำคาญโฆษณาเบียร์ช้างที่พี่เวียร์เช่นละ", "gold": [[0, 2], [2, 7], [7, 12], [12, 17], [17, 23], [23, 27], [27, 30], [30, 33], [33, 39], [39, 43], [43, 45]]}
5
+ {"id": 4, "text": "ตายๆHONDAทำไรไม่สนใจโลกเลยเขามา4วาวแล้วซีๆก็เยอะไม่เปลี่ยนโฉมสักทีเลย", "gold": [[0, 3], [3, 4], [4, 9], [9, 11], [11, 13], [13, 16], [16, 18], [18, 20], [20, 23], [23, 26], [26, 29], [29, 31], [31, 32], [32, 35], [35, 39], [39, 41], [41, 42], [42, 44], [44, 48], [48, 51], [51, 58], [58, 61], [61, 64], [64, 66], [66, 69]]}
6
+ {"id": 5, "text": "สำหรับอติล่าเป็นการขายโตโยต้าให้เหมือนขายเบนซ์ค่ะ ดูแพงดูเฉียบ! กรี้ดสุดดด #TheFaceAllStars", "gold": [[0, 6], [6, 12], [12, 16], [16, 19], [19, 22], [22, 29], [29, 32], [32, 38], [38, 41], [41, 46], [46, 49], [50, 52], [52, 55], [55, 57], [57, 62], [62, 63], [64, 69], [69, 74], [75, 76], [76, 79], [79, 83], [83, 86], [86, 91]]}
7
+ {"id": 6, "text": "เมื่อวานไปหาหมอด้วยความอยากดื่มกาแฟเย็นแต่เป็นหวัดเลยเดินไปสั่งลาเต้ร้อนระหว่างที่รอจ่ายเงินค่ายา พร้อมกับหยิบโดนัทและพิซซ่าเลยถามพนักงานว่าอันนี้ทำใหม่วันนี้หรือเปล่าครับ ถามอยู่สองครั้งแต่พนักงานคนนั้นไม่ได้ยินอาจเป็นเพราะว่าเราเจ็บคอเสียงแหบๆ แต่เธอเดินหนีแล้วบอกพนักงานอีกคนว่า มาคุยแทนพี่หน่อยพี่ไม่ได้ยินแล้วก็เดินหันหลังไปแบบไม่ใส่ใจทั้งๆที่ตัวเองเป็นพนักงานประจำอยู่แล้วใส่ฟอร์มพนักงานอยู่แต่ให้น้องผู้หญิงน่าจะเป็นน้องฝึกงานให้มาพูดกับเราแทน ผมรู้สึกเหมือนผมไปสร้างความรำคาญยังไงไม่รู้ สังเกตุหลายครั้งแล้วนะ สาขานี้บริการแบบไม่ค่อยเต็มใจ เวลาลูกค้าถามอะไรหน่อย เหมือนจะรำคาญ ไม่เหมือนที่รพ.พญาไท 1 แทบจะกราบลูกค้าเลย อยากได้อะไรจัดให้ บริการเลิศ หรือว่าสถานที่มันไม่ไฮโซหรือไง ทำไงก็ขายๆไป กาแฟแก้วละร้อยกว��าข้างนอกมีอีกตั้งหลายร้านถูกกว่าเกินครึ่ง ซื้อกาแฟซื้อเบเกอรี่รวมกันแต่ละครั้งสามสี่ร้อย แต่ก็รู้สึกรำคาญพนักงานที่นี่จริงๆ ถามอะไรต้องรู้สึกเกรงใจพนักงานทุกทีกลัวเขาจะรำคาญเรา เลิกซื้อเด็ดขาดแล้วครับ ขอบคุณที่ไม่เต็มใจบริการ ผมก็ไม่เต็มใจซื้อเหมือนกัน", "gold": [[0, 8], [8, 10], [10, 12], [12, 15], [15, 19], [19, 23], [23, 27], [27, 31], [31, 35], [35, 39], [39, 42], [42, 46], [46, 50], [50, 53], [53, 57], [57, 59], [59, 63], [63, 68], [68, 72], [72, 79], [79, 82], [82, 84], [84, 88], [88, 92], [92, 95], [95, 97], [98, 103], [103, 106], [106, 110], [110, 115], [115, 118], [118, 124], [124, 127], [127, 130], [130, 137], [137, 140], [140, 143], [143, 146], [146, 148], [148, 152], [152, 155], [155, 158], [158, 162], [162, 167], [167, 171], [172, 175], [175, 179], [179, 182], [182, 187], [187, 190], [190, 197], [197, 199], [199, 203], [203, 206], [206, 209], [209, 212], [212, 215], [215, 219], [219, 224], [224, 227], [227, 230], [230, 234], [234, 236], [236, 241], [241, 244], [244, 245], [246, 249], [249, 252], [252, 256], [256, 259], [259, 263], [263, 266], [266, 273], [273, 276], [276, 278], [278, 281], [282, 284], [284, 287], [287, 290], [290, 293], [293, 298], [298, 301], [301, 304], [304, 307], [307, 310], [310, 314], [314, 316], [316, 320], [320, 323], [323, 327], [327, 329], [329, 332], [332, 335], [335, 338], [338, 340], [340, 344], [344, 345], [345, 348], [348, 354], [354, 358], [358, 365], [365, 370], [370, 374], [374, 378], [378, 381], [381, 386], [386, 393], [393, 397], [397, 400], [400, 403], [403, 407], [407, 414], [414, 417], [417, 419], [419, 423], [423, 427], [427, 430], [430, 433], [433, 436], [436, 438], [438, 441], [441, 444], [444, 447], [447, 450], [451, 453], [453, 456], [456, 459], [459, 465], [465, 467], [467, 469], [469, 474], [474, 478], [478, 483], [483, 488], [488, 491], [491, 494], [495, 502], [502, 506], [506, 511], [511, 515], [515, 517], [518, 522], [522, 525], [525, 531], [531, 534], [534, 537], [537, 541], [541, 547], [548, 552], [552, 558], [558, 561], [561, 565], [565, 570], [571, 577], [577, 579], [579, 584], [585, 588], [588, 594], [594, 597], [597, 600], [600, 605], [606, 607], [608, 611], [611, 613], [613, 617], [617, 623], [623, 626], [627, 631], [631, 634], [634, 638], [638, 641], [641, 644], [645, 651], [651, 655], [656, 660], [660, 663], [663, 670], [670, 673], [673, 676], [676, 680], [680, 684], [684, 686], [687, 689], [689, 691], [691, 693], [693, 696], [696, 697], [697, 699], [700, 704], [704, 708], [708, 710], [710, 714], [714, 718], [718, 722], [722, 725], [725, 727], [727, 730], [730, 734], [734, 738], [738, 742], [742, 745], [745, 749], [749, 753], [753, 758], [759, 763], [763, 767], [767, 771], [771, 779], [779, 782], [782, 785], [785, 788], [788, 790], [790, 795], [795, 798], [798, 801], [801, 805], [806, 809], [809, 811], [811, 817], [817, 822], [822, 829], [829, 832], [832, 835], [835, 839], [839, 840], [841, 844], [844, 848], [848, 852], [852, 858], [858, 864], [864, 871], [871, 874], [874, 876], [876, 880], [880, 883], [883, 885], [885, 890], [890, 893], [894, 898], [898, 902], [902, 909], [909, 913], [913, 917], [918, 924], [924, 927], [927, 930], [930, 936], [936, 942], [943, 945], [945, 947], [947, 950], [950, 956], [956, 960], [960, 966], [966, 969]]}
8
+ {"id": 7, "text": "เคี้ยวไม่ไหว", "gold": [[0, 6], [6, 9], [9, 12]]}
9
+ {"id": 8, "text": "เบียร์ยูไม่อร่อยสัดๆๆๆๆๆฟๆ", "gold": [[0, 6], [6, 8], [8, 11], [11, 16], [16, 19], [19, 20], [20, 21], [21, 22], [22, 23], [23, 24], [24, 25], [25, 26]]}
10
+ {"id": 9, "text": "พะเยาไม่ร่วมรายการง่ะเธออ😭", "gold": [[0, 5], [5, 8], [8, 12], [12, 18], [18, 21], [21, 25], [25, 26]]}
11
+ {"id": 10, "text": "ผ่าม พ้าม #แวะมาตบเสียงมุข", "gold": [[0, 4], [5, 9], [10, 11], [11, 14], [14, 16], [16, 18], [18, 23], [23, 26]]}
12
+ {"id": 11, "text": "กวนตีนป่าวล่ะ ปีใหม่เก็บเต็มราคา พ่องงงงงง", "gold": [[0, 6], [6, 10], [10, 13], [14, 16], [16, 20], [20, 24], [24, 28], [28, 32], [33, 42]]}
13
+ {"id": 12, "text": "และ 25 ประเทศที่ The Face Shop เปิด ไม่มี Austria!!! เซงเบย T T #GOT7 #GOT7Official", "gold": [[0, 3], [4, 6], [7, 13], [13, 16], [17, 20], [21, 25], [26, 30], [31, 35], [36, 39], [39, 41], [42, 49], [49, 50], [50, 51], [51, 52], [53, 56], [56, 59], [60, 61], [62, 63], [64, 65], [65, 69], [70, 71], [71, 75], [75, 83]]}
14
+ {"id": 13, "text": "รถไฟฟ้���สร้างพร้อมกัน3-4สายแบบนี้คือวิกฤติรถติดเลยนะ แต่ข้อดีของการสร้างพร้อมกันคือเสร็จพร้อมกันเหมือนเสกไเลยล่ะ กทม.มีแผนีถไฟ13สาย เปิดแล้ว5 ถ้ามัวแต่ทำทีละสาย กว่าจะครบเราคงตายกันหมดแล้ว", "gold": [[0, 4], [4, 7], [7, 12], [12, 17], [17, 20], [20, 21], [21, 22], [22, 23], [23, 26], [26, 29], [29, 32], [32, 35], [35, 41], [41, 43], [43, 46], [46, 49], [49, 51], [52, 55], [55, 58], [58, 60], [60, 63], [63, 66], [66, 71], [71, 76], [76, 79], [79, 82], [82, 87], [87, 92], [92, 95], [95, 101], [101, 104], [104, 106], [106, 108], [108, 111], [112, 116], [116, 118], [118, 122], [122, 125], [125, 127], [127, 130], [131, 135], [135, 139], [139, 140], [141, 144], [144, 147], [147, 150], [150, 152], [152, 154], [154, 156], [156, 159], [160, 164], [164, 166], [166, 169], [169, 172], [172, 174], [174, 177], [177, 180], [180, 183], [183, 187]]}
15
+ {"id": 14, "text": "มาพนักงานแจ้งว่า ไม่มี", "gold": [[0, 2], [2, 9], [9, 13], [13, 16], [17, 20], [20, 22]]}
16
+ {"id": 15, "text": "บรัช 4u2 มันดูกระจุกๆอะพพอเอาไปใช้ที่แก้ม เกลี่ยยากเว่อ หรือแปรงนี่ไม่ดีเอง", "gold": [[0, 4], [5, 8], [9, 12], [12, 14], [14, 20], [20, 21], [21, 23], [23, 26], [26, 29], [29, 31], [31, 34], [34, 37], [37, 41], [42, 48], [48, 51], [51, 55], [56, 60], [60, 64], [64, 67], [67, 70], [70, 72], [72, 75]]}
17
+ {"id": 16, "text": "สาขาหาดใหญ่หายไปแล้ว....เมื่อไหร่จะกลับมาเปิดอีกคะ", "gold": [[0, 4], [4, 11], [11, 14], [14, 16], [16, 20], [20, 24], [24, 33], [33, 35], [35, 39], [39, 41], [41, 45], [45, 48], [48, 50]]}
18
+ {"id": 17, "text": "เวลาเห็นคนรุ่นใหม่สมัครลงเลือกตั้งเป็นตัวแทนพรรคพลังประชารัฐแล้วรู้สึกสิ้นหวังยังไงก็ไม่รู้ คือพวกคนแก่ๆที่อยู่ในพรรคก็ยังพอทำใจว่าแม่งเปลี่ยนยาก แต่คนรุ่นใหม่วัย30ที่เห็นดีเห็นงามกับวิธีการของพรรคนี้คือยังไงอ่ะ", "gold": [[0, 4], [4, 8], [8, 10], [10, 14], [14, 18], [18, 23], [23, 25], [25, 34], [34, 38], [38, 41], [41, 44], [44, 48], [48, 60], [60, 64], [64, 70], [70, 78], [78, 83], [83, 85], [85, 88], [88, 91], [92, 95], [95, 98], [98, 100], [100, 103], [103, 104], [104, 107], [107, 111], [111, 113], [113, 117], [117, 119], [119, 122], [122, 124], [124, 128], [128, 131], [131, 135], [135, 142], [142, 145], [146, 149], [149, 151], [151, 155], [155, 159], [159, 162], [162, 164], [164, 167], [167, 171], [171, 173], [173, 177], [177, 180], [180, 183], [183, 187], [187, 190], [190, 193], [193, 197], [197, 200], [200, 203], [203, 208], [208, 211]]}
19
+ {"id": 18, "text": "990,000 น่าจะขายดี ใช้เครื่องปัจจุบันของนาวาร่านี่แหละ Cost จะได้ต่ำ อุปกรณ์ที่ไม่จำเป็นไม่ต้องใส่มาให้ตัวรถมันแพง เด่วไปแต่งกันข้างนอกเอง มาทีหลังต้องมีทีเด็ด", "gold": [[0, 7], [8, 11], [11, 13], [13, 16], [16, 18], [19, 22], [22, 29], [29, 37], [37, 40], [40, 47], [47, 50], [50, 54], [55, 59], [60, 62], [62, 65], [65, 68], [69, 76], [76, 79], [79, 82], [82, 88], [88, 91], [91, 95], [95, 98], [98, 100], [100, 103], [103, 106], [106, 108], [108, 111], [111, 114], [115, 119], [119, 121], [121, 125], [125, 128], [128, 132], [132, 135], [135, 138], [139, 141], [141, 143], [143, 147], [147, 151], [151, 153], [153, 159]]}
20
+ {"id": 19, "text": "ไม่น่ากินเบียร์ช้างเลยจริงๆ", "gold": [[0, 3], [3, 6], [6, 9], [9, 15], [15, 19], [19, 22], [22, 26], [26, 27]]}
21
+ {"id": 20, "text": "จุกอก กันไป กลืนไม่เข้าคายไม่ออก", "gold": [[0, 3], [3, 5], [6, 9], [9, 11], [12, 16], [16, 19], [19, 23], [23, 26], [26, 29], [29, 32]]}
22
+ {"id": 21, "text": "ความร้อนไม่ทำให้เดือดร้อนสักเท่าไหร่ ถ้าใจเราร่มเย็น เข้าวัดเลยค่ะ เข้าโบสถ์ #สวยกล้าท้าทุกแสง #คู่หูท้าแดด #kmacosmetics", "gold": [[0, 4], [4, 8], [8, 11], [11, 13], [13, 16], [16, 25], [25, 28], [28, 32], [32, 36], [37, 40], [40, 42], [42, 45], [45, 52], [53, 57], [57, 60], [60, 63], [63, 66], [67, 71], [71, 76], [77, 78], [78, 81], [81, 85], [85, 88], [88, 91], [91, 94], [95, 96], [96, 101], [101, 104], [104, 107], [108, 109], [109, 112], [112, 121]]}
23
+ {"id": 22, "text": "เจ้าชู้ไม่ผิดแต่ติดตรงเมิงมีแฟนเป็นคนดังเคีลย์ยากอีโก้เยอะดราม่าไม่จบ เมียสาวโรงงานพาไปกินเอ็มเคเลี้ยงหนังช็อปปิ้วนิดหน่อย ก็หายแบ้ววว", "gold": [[0, 7], [7, 10], [10, 13], [13, 16], [16, 19], [19, 22], [22, 26], [26, 28], [28, 31], [31, 35], [35, 37], [37, 40], [40, 46], [46, 49], [49, 54], [54, 58], [58, 64], [64, 67], [67, 69], [70, 74], [74, 77], [77, 83], [83, 85], [85, 87], [87, 90], [90, 96], [96, 102], [102, 106], [106, 114], [114, 117], [117, 122], [123, 125], [125, 128], [128, 134]]}
24
+ {"id": 23, "text": "ใช่ๆๆ ลิ้นเเทบขาด", "gold": [[0, 3], [3, 4], [4, 5], [6, 10], [10, 14], [14, 17]]}
25
+ {"id": 24, "text": "ขอแชร์ประสบการณ์นะคะ ฝังยาคุมที่โรงพยาบาลกรุงเทพเมื่อห้าปีก่อน ราคาประมาณแปดพัน (แพงมากกกก) ฝังครบสามปี ตัวเราไม่มีผลข้างเคียงอะไรเลย ผ่าตัดเล็กเอาหลอดยาออกมีแผลนิดเดียว", "gold": [[0, 2], [2, 6], [6, 16], [16, 18], [18, 20], [21, 24], [24, 29], [29, 32], [32, 41], [41, 48], [48, 53], [53, 56], [56, 58], [58, 62], [63, 67], [67, 73], [73, 76], [76, 79], [80, 81], [81, 84], [84, 90], [90, 91], [92, 95], [95, 98], [98, 101], [101, 103], [104, 107], [107, 110], [110, 113], [113, 115], [115, 117], [117, 126], [126, 130], [130, 133], [134, 140], [140, 144], [144, 147], [147, 151], [151, 153], [153, 156], [156, 158], [158, 161], [161, 164], [164, 169]]}
26
+ {"id": 25, "text": "Honda city ก็เเจ่มนะ แต่แพงขึ้นอีก", "gold": [[0, 5], [6, 10], [11, 13], [13, 18], [18, 20], [21, 24], [24, 27], [27, 31], [31, 34]]}
27
+ {"id": 26, "text": "ตัวดีเซลคือไปออกฟอจูนเนอเหอะ crv ทุกตัวความเหี้ยคือแดกน้ำมัน", "gold": [[0, 3], [3, 8], [8, 11], [11, 13], [13, 16], [16, 24], [24, 28], [29, 32], [33, 36], [36, 39], [39, 43], [43, 48], [48, 51], [51, 54], [54, 60]]}
28
+ {"id": 27, "text": "ส่งไม่เห็นขึ้นเลยคะ ทาง inbox", "gold": [[0, 3], [3, 6], [6, 10], [10, 14], [14, 17], [17, 19], [20, 23], [24, 29]]}
29
+ {"id": 28, "text": "ขอบใจมากจ้า อีสัดดดด", "gold": [[0, 5], [5, 8], [8, 11], [12, 14], [14, 20]]}
30
+ {"id": 29, "text": "ข้าวผัดแฉะมากค่ะ ยังไม่โดน", "gold": [[0, 4], [4, 7], [7, 10], [10, 13], [13, 16], [17, 20], [20, 23], [23, 26]]}
31
+ {"id": 30, "text": "เรดส์ปลอมนี่สถุนกว่าเหล้าขาว 40 ดีกรีอีก", "gold": [[0, 5], [5, 9], [9, 12], [12, 16], [16, 20], [20, 25], [25, 28], [29, 31], [32, 37], [37, 40]]}
32
+ {"id": 31, "text": "งดกินไปก่อนนน", "gold": [[0, 2], [2, 5], [5, 7], [7, 13]]}
33
+ {"id": 32, "text": "ส่งให้เรียบร้อยครับ ไม่พอใจจริงๆครับครั้งนี้ พนักงานชุดนี้แย่จริงๆ", "gold": [[0, 3], [3, 6], [6, 15], [15, 19], [20, 23], [23, 27], [27, 31], [31, 32], [32, 36], [36, 41], [41, 44], [45, 52], [52, 55], [55, 58], [58, 61], [61, 65], [65, 66]]}
34
+ {"id": 33, "text": "โดน smirnoff เข้าไป เกือบเอาตัวไม่รอด", "gold": [[0, 3], [4, 12], [13, 17], [17, 19], [20, 25], [25, 28], [28, 31], [31, 34], [34, 37]]}
35
+ {"id": 34, "text": "เหมือนจะแพ้ Blemish Cleanser ของ Kiehl’s #เพิ่งซื้อมาเอง #ไว้อาลัย", "gold": [[0, 6], [6, 8], [8, 11], [12, 19], [20, 28], [29, 32], [33, 40], [41, 42], [42, 47], [47, 51], [51, 53], [53, 56], [57, 58], [58, 61], [61, 66]]}
36
+ {"id": 35, "text": "#ไม่แพงเอาเข้าจริงแล้ว เพราะสั่งอาหารอย่างละ2 และบองคิดเล่นๆ เบียร์ช้างขวดละ120 น้ำขวดละ50 มะพร้าวลูกละ60 ข้าวผัดทะเลโถละ100 ข้าวแค่โถละ60 ส่วนอาหารทะเลก็สมควรแกราคาถ้าไปกินที่ร้านก็ประมาณนี้แต่ทางร้านของป้า เขียนอาหารทะเล500 แต่ติด 700 ลูกค้าไม่ยอมน่าจะเป็นส่วนนี้ตามจริงแล้วทางร้านต้องแจ้งก่อน", "gold": [[0, 1], [1, 4], [4, 7], [7, 10], [10, 14], [14, 18], [18, 22], [23, 28], [28, 32], [32, 37], [37, 42], [42, 44], [44, 45], [46, 49], [49, 52], [52, 55], [55, 59], [59, 60], [61, 67], [67, 71], [71, 74], [74, 76], [76, 79], [80, 83], [83, 86], [86, 88], [88, 90], [91, 98], [98, 101], [101, 103], [103, 105], [106, 113], [113, 117], [117, 119], [119, 121], [121, 124], [125, 129], [129, 132], [132, 134], [134, 136], [136, 138], [139, 143], [143, 148], [148, 152], [152, 154], [154, 159], [159, 161], [161, 165], [165, 168], [168, 170], [170, 173], [173, 176], [176, 180], [180, 182], [182, 188], [188, 191], [191, 194], [194, 197], [197, 201], [201, 204], [204, 207], [208, 213], [213, 218], [218, 222], [222, 225], [226, 229], [229, 232], [233, 236], [237, 243], [243, 246], [246, 249], [249, 252], [252, 254], [254, 258], [258, 262], [262, 265], [265, 268], [268, 272], [272, 276], [276, 279], [279, 283], [283, 287], [287, 291], [291, 295]]}
37
+ {"id": 36, "text": "เห็นทุกค่ายเลย ลูกค้าร้องเรียนปัญหาดีดีไม่เคยรับผิดชอบเต็มที่ ใครจะอยากจ่ายเงินแล้วได้ของห่วยมาวะ ถ้ารับผิดชอบคงไม่ต้องขนาดนี้ ไม่ใช่แค่ยี่ห้อนี้นะ ไปจบที่ทุบรถโชว์ ถือป้ายประท้วงในงานมอเตอร์โชว์ นัดรวมตัวกันประท้วง ทุกที ถึงจะรับผิดชอบ นี่เค้าถึงขั้นไปกลางงาย นอกจากไม่รับผิดชอบแล้วยังฟ้องอีก ต่อไปคงจะมีคนซื้อรถมึงหรอก", "gold": [[0, 4], [4, 7], [7, 11], [11, 14], [15, 21], [21, 30], [30, 35], [35, 37], [37, 39], [39, 42], [42, 45], [45, 54], [54, 61], [62, 65], [65, 67], [67, 71], [71, 75], [75, 79], [79, 83], [83, 86], [86, 89], [89, 93], [93, 95], [95, 97], [98, 101], [101, 110], [110, 112], [112, 115], [115, 119], [119, 123], [123, 126], [127, 130], [130, 133], [133, 136], [136, 142], [142, 145], [145, 147], [148, 150], [150, 152], [152, 155], [155, 158], [158, 160], [160, 164], [165, 168], [168, 172], [172, 179], [179, 181], [181, 184], [184, 191], [191, 195], [196, 199], [199, 202], [202, 205], [205, 208], [208, 215], [216, 219], [219, 221], [222, 225], [225, 227], [227, 236], [237, 240], [240, 244], [244, 247], [247, 251], [251, 253], [253, 257], [257, 260], [261, 264], [264, 267], [267, 270], [270, 279], [279, 283], [283, 286], [286, 290], [290, 293], [294, 297], [297, 299], [299, 301], [301, 303], [303, 305], [305, 307], [307, 311], [311, 313], [313, 316], [316, 320]]}
38
+ {"id": 37, "text": "แต่ผมชอบกินช้างนะ เพราะลีโอมันออกเปรี้ยวๆ", "gold": [[0, 3], [3, 5], [5, 8], [8, 11], [11, 15], [15, 17], [18, 23], [23, 27], [27, 30], [30, 33], [33, 40], [40, 41]]}
39
+ {"id": 38, "text": "ขับอีโค่ของฮอนด้า แล้วไปขับกระบะของบริษัท เป็นดีแมกซ์พวงมาลัยพาวเวอร์แข็งมาก ช่วงยูเทิร์นหมุนจนปวดแขนเลยทีเดียว", "gold": [[0, 3], [3, 8], [8, 11], [11, 17], [18, 22], [22, 24], [24, 27], [27, 32], [32, 35], [35, 41], [42, 46], [46, 53], [53, 61], [61, 69], [69, 73], [73, 76], [77, 81], [81, 89], [89, 93], [93, 95], [95, 98], [98, 101], [101, 104], [104, 106], [106, 111]]}
40
+ {"id": 39, "text": "ช้าง...... กินแล้วอวกเกือบตาย...... เบียร์คนไทยทำเองๆๆๆๆ555", "gold": [[0, 4], [4, 10], [11, 14], [14, 18], [18, 21], [21, 26], [26, 29], [29, 35], [36, 42], [42, 44], [44, 47], [47, 49], [49, 52], [52, 53], [53, 54], [54, 55], [55, 56], [56, 59]]}
41
+ {"id": 40, "text": "ที่นี่ไม่มีความลับ : ‘ซีพี’ หวังกินรวบประเทศ! รุกอสังหาฯ-ขนส่งหนังสือพิมพ์ฐานเศรษฐกิจได้นำเสนอข่าวที่บริษัทยักษ์ใหญ่ด้านการเกษตรอย่าง “เครือเจริญโภคภัณฑ์” หันมาสนใจธุรกิจใหม่ 2 ด้านที่มีความสัมพันธ์ยึดโยงกัน คือด้านอสังหาริมทรัพย์และการขนส่งคือ รถไฟความเร็วสูงสายภาคตะวันออกที่เชื่อม 3 สนามบิน “ดอนเมือง-สุวรรณภูมิ-อู่ตะเภา” ระยะทางรวม 220 กม. ซึ่งไม่ผิดความคาดหมายเพราะก่อนหน้านี้เมื่อกลางปี 2560 คอลัมน์เอราวัณ ก็ได้เปิดประเด็นไปครั้งหนึ่งแล้วว่ารถไฟความเร็วสูงภาคตะวันออกเป็นเส้นทางที่ “บิ๊กซีพี” ขอจอง ขณะที่สายใต้ กรุงเทพฯ-หัวหิน เป็นการจองของ “เบียร์ช้าง” แต่การทำรถไฟความเร็วสูงหาจุดคุ้มทุนไม่ง่าย สิ่งที่ภาคเอกชนรวมทั้งบิ๊กในซีพี อย่างเจ้าสัวใหญ่-ธนินท์ เจียรวนนท์ คิดคือการทำกำไรในธุรกิจอสังหาริมทรัพย์ ในเส้นทางที่รถไฟความเร็วสูงผ่านและมีสถานีใกล้ที่ดินเหล่านั้น http://www.thansettakij.com/content/265699", "gold": [[0, 3], [3, 6], [6, 9], [9, 11], [11, 15], [15, 18], [19, 20], [21, 22], [22, 26], [26, 27], [28, 32], [32, 35], [35, 38], [38, 44], [44, 45], [46, 49], [49, 56], [56, 57], [57, 62], [62, 74], [74, 77], [77, 85], [85, 88], [88, 90], [90, 94], [94, 98], [98, 101], [101, 107], [107, 112], [112, 116], [116, 120], [120, 123], [123, 128], [128, 133], [134, 135], [135, 140], [140, 153], [153, 154], [155, 158], [158, 160], [160, 164], [164, 170], [170, 174], [175, 176], [177, 181], [181, 184], [184, 186], [186, 190], [190, 198], [198, 201], [201, 204], [204, 207], [208, 211], [211, 215], [215, 230], [230, 233], [233, 236], [236, 238], [238, 241], [241, 244], [245, 249], [249, 253], [253, 257], [257, 260], [260, 263], [263, 266], [266, 271], [271, 274], [274, 277], [277, 283], [284, 285], [286, 293], [294, 295], [295, 303], [303, 304], [304, 314], [314, 315], [315, 323], [323, 324], [325, 329], [329, 332], [332, 335], [336, 339], [340, 343], [344, 348], [348, 351], [351, 354], [354, 358], [358, 365], [365, 370], [370, 374], [374, 378], [378, 381], [381, 386], [386, 390], [390, 392], [393, 397], [398, 405], [405, 412], [413, 415], [415, 418], [418, 422], [422, 429], [429, 431], [431, 436], [436, 441], [441, 445], [445, 448], [448, 452], [452, 460], [460, 463], [463, 466], [466, 474], [474, 478], [478, 485], [485, 488], [489, 490], [490, 494], [494, 498], [498, 499], [500, 502], [502, 505], [506, 509], [509, 512], [512, 515], [515, 518], [519, 527], [527, 528], [528, 534], [535, 539], [539, 542], [542, 545], [545, 548], [549, 550], [550, 556], [556, 560], [560, 561], [562, 565], [565, 568], [568, 570], [570, 574], [574, 578], [578, 582], [582, 585], [585, 587], [587, 590], [590, 594], [594, 597], [597, 600], [600, 604], [605, 609], [609, 612], [612, 615], [615, 620], [620, 623], [623, 627], [627, 631], [631, 633], [633, 637], [638, 643], [643, 650], [650, 654], [654, 655], [655, 661], [662, 672], [673, 676], [676, 679], [679, 682], [682, 684], [684, 688], [688, 690], [690, 696], [696, 711], [712, 714], [714, 718], [718, 721], [721, 724], [724, 728], [728, 732], [732, 736], [736, 739], [739, 743], [743, 746], [746, 748], [748, 753], [753, 757], [757, 763], [763, 768], [768, 772], [773, 815]]}
42
+ {"id": 41, "text": "ไช่นายป่าวคริๆ", "gold": [[0, 3], [3, 6], [6, 10], [10, 13], [13, 14]]}
43
+ {"id": 42, "text": "5 สิ่ง...รู้เท่าทัน FORD Ranger Raptor", "gold": [[0, 1], [2, 6], [6, 9], [9, 12], [12, 16], [16, 19], [20, 24], [25, 31], [32, 38]]}
44
+ {"id": 43, "text": "ทุกสิ่งอย่างที่พวกเค้าต้านเรา ก็แค่ประโยคเดียวล่ะครับ \" ผลประโยชน์ \"", "gold": [[0, 3], [3, 7], [7, 12], [12, 15], [15, 18], [18, 22], [22, 26], [26, 29], [30, 32], [32, 35], [35, 41], [41, 46], [46, 49], [49, 53], [54, 55], [56, 58], [58, 66], [67, 68]]}
45
+ {"id": 44, "text": "นิสสัน เทอร์รา ใหม่…พัฒนาโดยฝ่ายโครงสร้างรถยนต์เพื่อการพาณิชย์ขนาดเล็กของนิสสัน เพื่อให้เป็นรถเอนกประสงค์สำหรับคนรุ่นใหม่อย่างแท้จริง", "gold": [[0, 6], [7, 14], [15, 19], [19, 20], [20, 25], [25, 28], [28, 32], [32, 41], [41, 47], [47, 52], [52, 55], [55, 62], [62, 66], [66, 70], [70, 73], [73, 79], [80, 85], [85, 88], [88, 92], [92, 94], [94, 105], [105, 111], [111, 113], [113, 117], [117, 121], [121, 126], [126, 129], [129, 133]]}
46
+ {"id": 45, "text": "ชุดทรูนี้คือต้องใช้เพราะรายเดือนหร่อค่ะหรือว่าแบบเติมเงินก็ได้ค่ะ", "gold": [[0, 3], [3, 6], [6, 9], [9, 12], [12, 16], [16, 19], [19, 24], [24, 27], [27, 32], [32, 36], [36, 39], [39, 43], [43, 46], [46, 49], [49, 53], [53, 57], [57, 59], [59, 62], [62, 65]]}
47
+ {"id": 46, "text": "ที่เมกะมีมั้ย", "gold": [[0, 3], [3, 7], [7, 9], [9, 13]]}
48
+ {"id": 47, "text": "ค้นพบถุงเก็บชาอันใหม่ งื้ออออ ชอบพกชา กาแฟ และเครื่องดื่มชนิดชงร้อนตั่งต่างอยู่แล้ว เป็นกระเป๋าที่เหมาะมักๆ เดี๋ยวจะใส่รูปพิทอมไว้หลัก Coffee or Tea or Tom?", "gold": [[0, 3], [3, 5], [5, 8], [8, 12], [12, 14], [14, 17], [17, 21], [22, 29], [30, 33], [33, 35], [35, 37], [38, 42], [43, 46], [46, 57], [57, 61], [61, 63], [63, 67], [67, 71], [71, 75], [75, 79], [79, 83], [84, 88], [88, 95], [95, 98], [98, 103], [103, 106], [106, 107], [108, 114], [114, 116], [116, 119], [119, 122], [122, 124], [124, 127], [127, 130], [130, 134], [135, 141], [142, 144], [145, 148], [149, 151], [152, 155], [155, 156]]}
49
+ {"id": 48, "text": "ตื่นๆ ไปกินข้าวกัน", "gold": [[0, 4], [4, 5], [6, 8], [8, 11], [11, 15], [15, 18]]}
50
+ {"id": 49, "text": "แปปปดูปฏิทิน", "gold": [[0, 4], [4, 6], [6, 12]]}
51
+ {"id": 50, "text": "เมื่อก่อนหน้าแห้งมากคะตั้งแต่มาใช้คู่นี้ Hada Labo หน้ามีความแววมากแต่ยังคงมีกระจางๆอยู่คะ", "gold": [[0, 5], [5, 9], [9, 13], [13, 17], [17, 20], [20, 22], [22, 26], [26, 29], [29, 31], [31, 34], [34, 37], [37, 40], [41, 45], [46, 50], [51, 55], [55, 57], [57, 61], [61, 64], [64, 67], [67, 70], [70, 73], [73, 75], [75, 77], [77, 83], [83, 84], [84, 88], [88, 90]]}
52
+ {"id": 51, "text": "นี่ขนาดในหมอชิตนะ รถที่จะรับควรมีการลงทะเบียนตรวจสอบได้และทำตามกฎน่าจะดีกว่า เราลงตรงนครชัยขึ้นรถตรงนั้น บางทีได้ขึ้นพรีอุสนครชัย รถเยอะไม่ต้องรอ", "gold": [[0, 3], [3, 7], [7, 9], [9, 15], [15, 17], [18, 20], [20, 23], [23, 25], [25, 28], [28, 31], [31, 33], [33, 36], [36, 45], [45, 49], [49, 52], [52, 55], [55, 58], [58, 60], [60, 63], [63, 65], [65, 68], [68, 70], [70, 72], [72, 76], [77, 80], [80, 82], [82, 85], [85, 91], [91, 95], [95, 97], [97, 100], [100, 104], [105, 108], [108, 110], [110, 113], [113, 117], [117, 123], [123, 129], [130, 132], [132, 136], [136, 139], [139, 143], [143, 145]]}
53
+ {"id": 52, "text": "หน้าตาเหมือนยาริส", "gold": [[0, 6], [6, 12], [12, 17]]}
54
+ {"id": 53, "text": "พบการกลับมาของ 2 วัฒนธรรมชื่อดังจากฝั่งอังกฤษ กลุ่มคนที่ต่างคลับ-ต่างขั้ว...จะมารวมตัวกันในงาน \" Mods 🇬🇧 & Rockers ♣️\" ......11 พฤศจิกายน 2017...... - 16.00 น. รวมตัวหน้าห้าง DD Mall ( สวนจตุจักร ) - 17.30 น. เคลื่อนขบวน \" MODS & ROCKERS \" สู่ \" RODS MOTO CAFE \" ตลาดนัดรถไฟรัชดา - 19.00 น. \" Free Bar \" ( สำหรับผู้ลงทะเบียนเข้างาน ) พบกิจกรรม สังสรรค์ กินดื่ม พร้อมชมมินิคอนเสิร์ต ...รวมตัวต้องมันส์...รักกันต้องมา... *** ขอความร่วมมือผู้ร่วมงานทุกท่านแต่งตัวเต็มยศทั้ง Mods และ Rockers ครับ ผู้สนับสนุนการปลุกตำนาน - ตลาดนัดรถไฟ - Rods Antique - DOGSKIN - Bell's Whisky", "gold": [[0, 2], [2, 5], [5, 9], [9, 11], [11, 14], [15, 16], [17, 25], [25, 29], [29, 32], [32, 35], [35, 39], [39, 45], [46, 51], [51, 53], [53, 56], [56, 60], [60, 64], [64, 65], [65, 69], [69, 73], [73, 76], [76, 78], [78, 80], [80, 83], [83, 86], [86, 89], [89, 91], [91, 94], [95, 96], [97, 101], [102, 104], [105, 106], [107, 114], [115, 117], [117, 118], [119, 125], [125, 127], [128, 137], [138, 142], [142, 148], [149, 150], [151, 156], [157, 159], [160, 163], [163, 166], [166, 170], [170, 174], [175, 177], [178, 182], [183, 184], [185, 195], [196, 197], [198, 199], [200, 205], [206, 208], [209, 216], [216, 220], [221, 222], [223, 227], [228, 229], [230, 237], [238, 239], [240, 243], [244, 245], [246, 250], [251, 255], [256, 260], [261, 262], [263, 270], [270, 274], [274, 279], [280, 281], [282, 287], [288, 290], [291, 292], [293, 297], [298, 301], [302, 303], [304, 305], [306, 312], [312, 315], [315, 317], [317, 324], [324, 328], [328, 331], [332, 333], [334, 336], [336, 343], [344, 352], [353, 356], [356, 360], [361, 366], [366, 368], [368, 372], [372, 381], [382, 385], [385, 391], [391, 395], [395, 400], [400, 403], [403, 406], [406, 409], [409, 413], [413, 415], [415, 418], [419, 422], [423, 425], [425, 429], [429, 436], [436, 439], [439, 443], [443, 446], [446, 449], [449, 453], [453, 457], [457, 460], [460, 464], [464, 466], [466, 470], [471, 475], [476, 479], [480, 487], [488, 492], [493, 496], [496, 504], [504, 507], [507, 511], [511, 516], [517, 518], [519, 530], [531, 532], [533, 537], [538, 545], [546, 547], [548, 555], [556, 557], [558, 564], [565, 571]]}
55
+ {"id": 54, "text": "นี่เป็นแค่โฆษณา ผ้าอนามัยนะ จะเอาอะไรนักหนา", "gold": [[0, 3], [3, 7], [7, 10], [10, 15], [16, 25], [25, 27], [28, 30], [30, 33], [33, 37], [37, 43]]}
56
+ {"id": 55, "text": "จ๋าาาจ้ะ", "gold": [[0, 5], [5, 8]]}
57
+ {"id": 56, "text": "‼️เเค่ชื่อก็มันเเล้ว 🎶 วงป๊อบร็อคหน้าหล่อ LIZM ร่วมกับ Chivas Regal Thailand เสนอคอนเสิร์ต POTATO 🙌 ศุกร์ที่ 6 ตุลาคม ที่ร้าน #Lizm ประชาชื่น เพลงช้าก็จี๊ด เพลงเร็วก็เต้นมันส์ มาเซลฟี่หน้าเวทีกัน พิเศษเฉพาะงานนี้ โปร Chivas Regal Extra โทรจองโต๊ะ หรือสอบถามได้ที่ 022222222 / 088-888-8888 #LizmBangkok #Potato #ChivasX #ShareChivas", "gold": [[0, 2], [2, 6], [6, 10], [10, 12], [12, 15], [15, 20], [21, 22], [23, 25], [25, 29], [29, 33], [33, 37], [37, 41], [42, 46], [47, 51], [51, 54], [55, 61], [62, 67], [68, 76], [77, 81], [81, 90], [91, 97], [98, 99], [100, 105], [105, 108], [109, 110], [111, 117], [118, 121], [121, 125], [126, 127], [127, 131], [132, 141], [142, 146], [146, 149], [149, 151], [151, 155], [156, 160], [160, 164], [164, 166], [166, 170], [170, 175], [176, 178], [178, 184], [184, 188], [188, 192], [192, 195], [196, 201], [201, 206], [206, 209], [209, 212], [213, 216], [217, 223], [224, 229], [230, 235], [236, 239], [239, 242], [242, 246], [247, 251], [251, 254], [254, 257], [257, 260], [260, 263], [264, 273], [274, 275], [276, 288], [289, 290], [290, 294], [294, 301], [302, 303], [303, 309], [310, 311], [311, 317], [317, 318], [319, 320], [320, 325], [325, 331]]}
58
+ {"id": 57, "text": "กินป๋องดำ แล้วเปิดไลฟ์ดู มันอยู่ในห้องก็ได้จบๆไป", "gold": [[0, 3], [3, 7], [7, 9], [10, 14], [14, 18], [18, 22], [22, 24], [25, 28], [28, 32], [32, 34], [34, 38], [38, 40], [40, 43], [43, 45], [45, 46], [46, 48]]}
59
+ {"id": 58, "text": "การจะเรียกโออิรันให้มารับใช้นั้น ลูกค้าต้องเรียกผ่านร้านน้ำชาหรือโรงน้ำชาที่เป็นตัวกลาง พวกเขาต้องพิสูจน์ว่าพวกเขามีเงินเพียงพอที่จะเล่นกับโออิรัน เมื่อพวกเขาขอเชิญโออิรันผ่านทางโรงน้ำชาเรียบร้อยแล้วก็ต้องฝากเงินก้อนไว้หนึ่งก้อน", "gold": [[0, 3], [3, 5], [5, 10], [10, 17], [17, 20], [20, 22], [22, 25], [25, 28], [28, 32], [33, 39], [39, 43], [43, 48], [48, 52], [52, 56], [56, 61], [61, 65], [65, 73], [73, 76], [76, 80], [80, 87], [88, 91], [91, 94], [94, 98], [98, 105], [105, 108], [108, 111], [111, 114], [114, 116], [116, 120], [120, 125], [125, 127], [127, 130], [130, 132], [132, 136], [136, 139], [139, 146], [147, 152], [152, 155], [155, 158], [158, 160], [160, 164], [164, 171], [171, 175], [175, 178], [178, 186], [186, 195], [195, 199], [199, 201], [201, 205], [205, 208], [208, 212], [212, 216], [216, 219], [219, 224], [224, 228]]}
60
+ {"id": 59, "text": "หาน้ำกิน #ข่าวว่าพรุ่งนี้ภาษีขึ้น #beer #ubeer #golf_eat_alot", "gold": [[0, 2], [2, 5], [5, 8], [9, 10], [10, 14], [14, 17], [17, 22], [22, 25], [25, 29], [29, 33], [34, 35], [35, 39], [40, 41], [41, 46], [47, 48], [48, 61]]}
61
+ {"id": 60, "text": "กอล์ฟเมากินช้างเยอะไป", "gold": [[0, 5], [5, 8], [8, 11], [11, 15], [15, 19], [19, 21]]}
62
+ {"id": 61, "text": "มาสด้า ซีเอ���กซ์-3 สนุกกันอีกสักดอก", "gold": [[0, 6], [7, 17], [18, 22], [22, 25], [25, 28], [28, 31], [31, 34]]}
63
+ {"id": 62, "text": "วันเกิดกูไงง ไปกัน", "gold": [[0, 7], [7, 9], [9, 12], [13, 15], [15, 18]]}
64
+ {"id": 63, "text": "รับเงินมาแล้วเผาadส่งไงฮะ ถถถถถถถถถถ", "gold": [[0, 3], [3, 7], [7, 9], [9, 13], [13, 16], [16, 18], [18, 21], [21, 23], [23, 25], [26, 36]]}
65
+ {"id": 64, "text": "เสี่ยเจริญ เป็นเจ้าของ แม่โขง แสงโสม เหล้าขาว เบียร์ช้าง จบ", "gold": [[0, 5], [5, 10], [11, 15], [15, 22], [23, 29], [30, 36], [37, 42], [42, 45], [46, 52], [52, 56], [57, 59]]}
66
+ {"id": 65, "text": "ตอนนี้ทาพอนแล้วก็กันแดดเคเอค่ะ", "gold": [[0, 3], [3, 6], [6, 8], [8, 11], [11, 15], [15, 17], [17, 23], [23, 27], [27, 30]]}
67
+ {"id": 66, "text": "“เรื่องราวที่คุณไม่เคยรู้ของพวกเธอทั้ง 6 คน” ไม่ว่าจะเป็น มารีญา, เจนนี่ ปาหนัน ,นุ่น นพลักษณ์, แก้ว MasterChef, กานดา แฟชั่นไอคอน และ ชมพู่ มิสวีลแชร์ 2012 ที่จะมาบอกเล่าเรื่องราว #TrueToMyShade ของแต่ละคน เพราะ รองพื้นเนื้อซาตินจากทรูแมช 13 เฉดสี เข้าใจและพร้อมสนับสนุนให้ทุกเพศ ทุกสีผิว ทุกช่วงอายุเป็นตัวของตัวเอง เพราะความสวยงามคือความหลากหลาย โทนผิวขาว N1 - สำหรับ ผิวขาวอมชมพู [ใหม่] N2 - สำหรับ ผิวขาว G1 - สำหรับ ผิวขาวอมเหลือง โทนกลาง G2 - สำหรับ ผิวสีเนื้ออมชมพู [ใหม่] G2.5 - สำหรับ ผิวสีเนื้อ N3 - สำหรับ ผิวสีเนื้อปานกลาง G3 - สำหรับ ผิวสีเนื้อโทนเหลือง G4 - ผิวสองสีโทนอ่อน โทนเข้ม [ใหม่] G4.5 สำหรับ ผิวสีน้ำผึ้ง G5 - สำหรับ ผิวสองสีปานกลาง (เฉพาะช่องทางออนไลน์) N7 - สำหรับ ผิวสองสีโทนเข้ม [ใหม่] G6 สำหรับ ผิวสีเข้ม [ใหม่] G7 สำหรับ ผิวสีแทน ช้อปออนไลน์ คลิก: http://bit.ly/2IHUe1H #LOrealParisTH #TrueMatch #TrueToMyShade #ลอรีอัลปารีสเมคอัพ", "gold": [[0, 1], [1, 10], [10, 13], [13, 16], [16, 19], [19, 22], [22, 25], [25, 28], [28, 31], [31, 34], [34, 38], [39, 40], [41, 43], [43, 44], [45, 48], [48, 51], [51, 53], [53, 57], [58, 64], [64, 65], [66, 72], [73, 79], [80, 81], [81, 85], [86, 94], [94, 95], [96, 100], [101, 111], [111, 112], [113, 118], [119, 130], [131, 134], [135, 140], [141, 151], [152, 156], [157, 160], [160, 162], [162, 164], [164, 167], [167, 171], [171, 180], [181, 182], [182, 186], [186, 188], [188, 190], [190, 195], [196, 199], [199, 202], [202, 204], [204, 206], [207, 212], [213, 220], [220, 225], [225, 230], [230, 233], [233, 239], [240, 242], [243, 246], [246, 248], [249, 255], [255, 258], [258, 263], [263, 271], [271, 274], [274, 277], [277, 280], [281, 284], [284, 286], [286, 289], [290, 293], [293, 297], [297, 301], [301, 305], [305, 308], [308, 311], [311, 317], [318, 323], [323, 327], [327, 330], [330, 333], [333, 336], [336, 340], [340, 344], [344, 348], [349, 352], [352, 355], [355, 358], [359, 361], [362, 363], [364, 370], [371, 374], [374, 377], [377, 379], [379, 381], [381, 383], [384, 385], [385, 389], [389, 390], [391, 393], [394, 395], [396, 402], [403, 409], [410, 412], [413, 414], [415, 421], [422, 425], [425, 428], [428, 430], [430, 436], [437, 440], [440, 444], [445, 447], [448, 449], [450, 456], [457, 460], [460, 462], [462, 467], [467, 469], [469, 473], [474, 475], [475, 479], [479, 480], [481, 485], [486, 487], [488, 494], [495, 498], [498, 500], [500, 505], [506, 508], [509, 510], [511, 517], [518, 521], [521, 523], [523, 528], [528, 531], [531, 535], [536, 538], [539, 540], [541, 547], [548, 551], [551, 553], [553, 558], [558, 561], [561, 567], [568, 570], [571, 572], [573, 576], [576, 579], [579, 581], [581, 584], [584, 588], [589, 592], [592, 596], [597, 598], [598, 602], [602, 603], [604, 608], [609, 615], [616, 619], [619, 621], [621, 628], [629, 631], [632, 633], [634, 640], [641, 644], [644, 647], [647, 649], [649, 652], [652, 656], [657, 658], [658, 663], [663, 667], [667, 670], [670, 677], [677, 678], [679, 681], [682, 683], [684, 690], [691, 694], [694, 697], [697, 699], [699, 702], [702, 706], [707, 708], [708, 712], [712, 713], [714, 716], [717, 723], [724, 727], [727, 729], [729, 733], [734, 735], [735, 739], [739, 740], [741, 743], [744, 750], [751, 754], [754, 756], [756, 759], [760, 764], [764, 771], [772, 776], [776, 777], [778, 799], [800, 801], [801, 807], [807, 812], [812, 814], [815, 816], [816, 820], [820, 825], [826, 827], [827, 831], [831, 833], [833, 835], [835, 840], [841, 860]]}
68
+ {"id": 67, "text": "มันคืออัลไลน้อ กินก้อนฟรี1เดือนหรือเปล่า", "gold": [[0, 3], [3, 6], [6, 11], [11, 14], [15, 18], [18, 22], [22, 25], [25, 26], [26, 31], [31, 35], [35, 40]]}
69
+ {"id": 68, "text": "เส้นทางจราจร ริมถนน เส้นทาง อ.เมือง เชื่อมมายัง อำเภอปากพนัง จ.นครศรีธรรมราช บางจุดต้นไม้หักโค่น ขณะที่ถนนติดริมทะเลตั้งเเต่อำเภอหัวไทร ไปจนถึง อำเภอปากพนัง น้ำทะเลทะลัก เข้าท่วมถนน เเละบ้านเรือนของชาวบ้านเเล้ว", "gold": [[0, 7], [7, 12], [13, 16], [16, 19], [20, 27], [28, 30], [30, 35], [36, 42], [42, 44], [44, 47], [48, 53], [53, 60], [61, 63], [63, 76], [77, 80], [80, 83], [83, 89], [89, 92], [92, 96], [97, 100], [100, 103], [103, 106], [106, 109], [109, 112], [112, 116], [116, 120], [120, 124], [124, 129], [129, 135], [136, 138], [138, 140], [140, 143], [144, 149], [149, 156], [157, 160], [160, 164], [164, 169], [170, 174], [174, 178], [178, 181], [182, 186], [186, 190], [190, 195], [195, 198], [198, 205], [205, 210]]}
70
+ {"id": 69, "text": "มา สมองตื่นแล้ว มารู้จักกับแนวางยุบรวม 3 กองทุนที่สปสช.กระทรวงสาธารณสุข กระทรวงแรงงาน และกรมบัญชีกลาง พยายามผลักดันมานานเกือบสิบปี และเริ่มลงมือใช้มาแล้วปีนี้ก็นับเป็นปีที่ 6ก่อนที่จะมีคนหรือพรรคการเมืองมาแอบอ้างผลงานตัวเองอีก ติดตามได้ใน #3กองทุน", "gold": [[0, 2], [3, 7], [7, 11], [11, 15], [16, 18], [18, 24], [24, 27], [27, 32], [32, 35], [35, 38], [39, 40], [41, 47], [47, 50], [50, 55], [55, 62], [62, 71], [72, 79], [79, 85], [86, 89], [89, 92], [92, 97], [97, 101], [102, 108], [108, 112], [112, 115], [115, 117], [117, 120], [120, 125], [125, 128], [128, 130], [131, 134], [134, 139], [139, 141], [141, 144], [144, 147], [147, 149], [149, 153], [153, 155], [155, 158], [158, 160], [160, 163], [163, 167], [167, 169], [169, 172], [173, 174], [174, 178], [178, 181], [181, 183], [183, 185], [185, 187], [187, 191], [191, 195], [195, 198], [198, 203], [203, 205], [205, 208], [208, 212], [212, 214], [214, 217], [217, 223], [223, 226], [227, 233], [233, 236], [236, 238], [239, 240], [240, 241], [241, 247]]}
71
+ {"id": 70, "text": "พกประตูวิเศษของโดเรมอนไปค่ะ จะได้แว๊บไปแว๊บมา ซื้อเบียร์ช้างมากินริมหาดด้วย เนอะๆๆ", "gold": [[0, 2], [2, 7], [7, 12], [12, 15], [15, 22], [22, 24], [24, 27], [28, 30], [30, 33], [33, 37], [37, 39], [39, 43], [43, 45], [46, 50], [50, 56], [56, 60], [60, 62], [62, 65], [65, 68], [68, 71], [71, 75], [76, 80], [80, 81], [81, 82]]}
72
+ {"id": 71, "text": "คุณลูกค้าสามารถซื้อผลิตภัณฑ์ออนไลน์ได้ทาง www.eucerin.co.th หรือทาง Lazada ที่ https://www.lazada.co.th/eucerin-official-store และสามารถเลือกชมโปรโมชั่นต่างๆ ได้ที่ https://www.eucerin.co.th/local/th/promotion/new-promotion และที่ร้านบูทส์ วัตสัน ร้านขายยาชั้นนำ โรงพยาบาลที่ร่วมรายการค่ะ", "gold": [[0, 3], [3, 9], [9, 15], [15, 19], [19, 28], [28, 35], [35, 38], [38, 41], [42, 59], [60, 64], [64, 67], [68, 74], [75, 78], [79, 126], [127, 130], [130, 136], [136, 141], [141, 143], [143, 152], [152, 156], [156, 157], [158, 164], [165, 223], [224, 227], [227, 230], [230, 239], [240, 246], [247, 256], [256, 262], [263, 272], [272, 275], [275, 279], [279, 285], [285, 288]]}
73
+ {"id": 72, "text": "ใช่ครับผม แต่จะมีบางสาขาเท่านั้นนะคร้าบต���นนี้ตอนนี้ ก้อนจัดโปรโมชั่นรีฟิล ทั้งหมด 10 สาขา รายละเอียดดังนี้ครับ 1) โปรโมชั่นรีฟิล ชุด Happy Refill 399.- และ Super Refill 499.- ที่สาขาเทอมินอล 21 และสาขาเดอะแพลทฟอร์ม วงเวียนใหญ่ เฉพาะลูกค้าที่ทำการจองผ่านระบบออนไลน์ของ App หรือ เว็ปไซต์ Hungry Hub เท่านั้น เริ่มตั้งแต่วันที่ 2 ธันวาคม’60 หมดเขต 31 มกราคม 2561 นี้คร้าบ 2) โปรโมชั่นรีฟิล ชุด Happy Refill 399.- และ Super Refill 499.- (ระยะเวลาในการทาน 1.30 ช.ม.) เริ่ม 7 ธันวาคม’60 จนถึง 31 มกราคม’61 สามารถ Walk in เข้าไปที่สาขาตามนี้ได้เลยนะคร้าบ - เซ็นทรัล บางนาชั้น6 - ดิอเวนิวรัชโยธิน - บิ๊กซี ลำพูน - โรบินสัน แม่สอด - โรบินสัน นครศรีธรรมราช - โลตัส บึงกาฬ - โลตัส จันทบุรี - เทอมินอลโคราช", "gold": [[0, 3], [3, 9], [10, 13], [13, 15], [15, 17], [17, 20], [20, 24], [24, 32], [32, 34], [34, 39], [39, 45], [45, 51], [52, 56], [56, 59], [59, 68], [68, 73], [74, 81], [82, 84], [85, 89], [90, 100], [100, 106], [106, 110], [111, 113], [114, 123], [123, 128], [129, 132], [133, 138], [139, 145], [146, 151], [152, 155], [156, 161], [162, 168], [169, 174], [175, 178], [178, 182], [182, 190], [191, 193], [194, 197], [197, 201], [201, 214], [215, 226], [227, 232], [232, 238], [238, 241], [241, 243], [243, 249], [249, 253], [253, 264], [264, 267], [268, 271], [272, 276], [277, 285], [286, 292], [293, 296], [297, 305], [306, 311], [311, 318], [318, 324], [325, 326], [327, 334], [334, 337], [338, 341], [341, 344], [345, 347], [348, 354], [355, 359], [360, 363], [363, 368], [369, 371], [372, 381], [381, 386], [387, 390], [391, 396], [397, 403], [404, 409], [410, 413], [414, 419], [420, 426], [427, 432], [433, 434], [434, 442], [442, 444], [444, 450], [451, 455], [456, 460], [460, 461], [462, 467], [468, 469], [470, 477], [477, 480], [481, 483], [483, 486], [487, 489], [490, 499], [500, 506], [507, 511], [512, 514], [515, 519], [519, 521], [521, 524], [524, 528], [528, 531], [531, 534], [534, 537], [537, 540], [540, 542], [542, 547], [548, 549], [550, 558], [559, 564], [564, 568], [568, 569], [570, 571], [572, 580], [580, 588], [589, 590], [591, 597], [598, 603], [604, 605], [606, 614], [615, 621], [622, 623], [624, 632], [633, 646], [647, 648], [649, 654], [655, 661], [662, 663], [664, 669], [670, 678], [679, 680], [681, 689], [689, 694]]}
74
+ {"id": 73, "text": "ใช่แล้วววววว", "gold": [[0, 3], [3, 12]]}
75
+ {"id": 74, "text": "ห่าน", "gold": [[0, 4]]}
76
+ {"id": 75, "text": "ทางแพทย์ผู้เชี่ยวชาญทางโรงพยาบาลพญาไทศรีราชาได้แจ้งรายละเอียดมาดังนี้นะคะ เนื่องจากอาการที่เป็นมีได้หลายสาเหตุด้วยกันค่ะ อาจเกิดจากความผิดปกติของฮอร์โมนโปรเจสเตอร์โรนในร่างกายที่สูงขึ้น ส่งผลให้มีอุณหภูมิกายสูงขึ้นตามไปด้วย (ระดับโปรเจสเตอร์โรนและอุณหภูมิกายจะคงอยู่สูงไปจนถึงวันที่ประจำเดือนมา) แนะนำเข้ามาพบแพทย์เพื่อตรวจร่างกายพร้อมกับซักประวัติเพิ่มเติมค่ะ", "gold": [[0, 3], [3, 8], [8, 20], [20, 23], [23, 32], [32, 44], [44, 47], [47, 51], [51, 61], [61, 63], [63, 69], [69, 71], [71, 73], [74, 80], [80, 83], [83, 88], [88, 91], [91, 95], [95, 97], [97, 100], [100, 104], [104, 110], [110, 114], [114, 117], [117, 120], [121, 124], [124, 128], [128, 131], [131, 138], [138, 142], [142, 145], [145, 152], [152, 166], [166, 168], [168, 175], [175, 178], [178, 181], [181, 185], [186, 189], [189, 191], [191, 194], [194, 196], [196, 207], [207, 210], [210, 214], [214, 217], [217, 219], [219, 223], [224, 225], [225, 230], [230, 244], [244, 247], [247, 258], [258, 260], [260, 262], [262, 266], [266, 269], [269, 271], [271, 273], [273, 276], [276, 279], [279, 282], [282, 292], [292, 294], [294, 295], [296, 301], [301, 305], [305, 307], [307, 309], [309, 314], [314, 319], [319, 323], [323, 330], [330, 335], [335, 338], [338, 341], [341, 348], [348, 357], [357, 360]]}
77
+ {"id": 76, "text": "เจอกันวันเกิด", "gold": [[0, 3], [3, 6], [6, 13]]}
78
+ {"id": 77, "text": "ซื้อปุ๊บ ได้เงินคืนปั๊บเลยค่ะ", "gold": [[0, 4], [4, 8], [9, 12], [12, 16], [16, 19], [19, 23], [23, 26], [26, 29]]}
79
+ {"id": 78, "text": "อย่างเช่น อะไร", "gold": [[0, 5], [5, 9], [10, 14]]}
80
+ {"id": 79, "text": "จิงป่าววว วันจันทร์-ศุกร์ เลือกมาเลย", "gold": [[0, 3], [3, 9], [10, 19], [19, 20], [20, 25], [26, 31], [31, 33], [33, 36]]}
81
+ {"id": 80, "text": "น้องชายเพิ่งจองมา ของแถมคุ้มเลยคะ ดอกเบี้ยถูกมาก ที่บ้านใช้ honda ทุกคันคะ", "gold": [[0, 7], [7, 12], [12, 15], [15, 17], [18, 21], [21, 24], [24, 28], [28, 31], [31, 33], [34, 42], [42, 45], [45, 48], [49, 52], [52, 56], [56, 59], [60, 65], [66, 69], [69, 72], [72, 74]]}
82
+ {"id": 81, "text": "เมื่อวานชานมไข่มุกก็อร่อย วันนี้รุ่นพี่ก็อร่อย //สูดกาววาดไม่ทันแล้ว #เต็นสิค", "gold": [[0, 5], [5, 8], [8, 12], [12, 18], [18, 20], [20, 25], [26, 32], [32, 36], [36, 39], [39, 41], [41, 46], [47, 49], [49, 52], [52, 55], [55, 58], [58, 61], [61, 64], [64, 68], [69, 70], [70, 77]]}
83
+ {"id": 82, "text": "นิสสันออกแต่ละรุ้นสุดยอดมากๆรุ่นนี้ดีมากๆๆประหยัด นม. สุดยอดมากๆเข้าไปดูด่วนๆ", "gold": [[0, 6], [6, 9], [9, 12], [12, 14], [14, 18], [18, 24], [24, 27], [27, 28], [28, 32], [32, 35], [35, 37], [37, 40], [40, 41], [41, 42], [42, 49], [50, 53], [54, 60], [60, 63], [63, 64], [64, 68], [68, 70], [70, 72], [72, 76], [76, 77]]}
84
+ {"id": 83, "text": "เคเอฟซี บาร์บีกอน", "gold": [[0, 7], [8, 17]]}
85
+ {"id": 84, "text": "ลุยๆ😚😚", "gold": [[0, 3], [3, 4], [4, 5], [5, 6]]}
86
+ {"id": 85, "text": "ก้อยากกกกกเป้นตาแซ่บบบ", "gold": [[0, 2], [2, 10], [10, 14], [14, 16], [16, 22]]}
87
+ {"id": 86, "text": "4หม้อเลยหน่าาาา", "gold": [[0, 1], [1, 5], [5, 8], [8, 15]]}
88
+ {"id": 87, "text": "หมดเขต ก่อนพอดี", "gold": [[0, 6], [7, 11], [11, 15]]}
89
+ {"id": 88, "text": "ชอบครับ รถในฝันเลยครับ Honda Accord ไปจองมาเรียบร้อยแล้วครับ", "gold": [[0, 3], [3, 7], [8, 10], [10, 12], [12, 15], [15, 18], [18, 22], [23, 28], [29, 35], [36, 38], [38, 41], [41, 43], [43, 52], [52, 56], [56, 60]]}
90
+ {"id": 89, "text": "น่าสนๆ^^", "gold": [[0, 3], [3, 5], [5, 6], [6, 8]]}
91
+ {"id": 90, "text": "นี่อยากไปพรุ่งนี้พุ่งตัวไปซีโฟร่าเลยหนาแต่ท่าพน.ต้องบีทีเอสเพราะไม่มีที่จอดหนาวันอาทิตย์ 555", "gold": [[0, 3], [3, 7], [7, 9], [9, 17], [17, 21], [21, 24], [24, 26], [26, 33], [33, 36], [36, 39], [39, 42], [42, 45], [45, 48], [48, 52], [52, 59], [59, 64], [64, 67], [67, 69], [69, 72], [72, 75], [75, 78], [78, 88], [89, 92]]}
92
+ {"id": 91, "text": "เห้ยน่าไปๆๆ จิบเบียร์ฟังดนตรี งานดีๆแบบนี้ต้องไม่พลาด #chang", "gold": [[0, 4], [4, 7], [7, 9], [9, 10], [10, 11], [12, 15], [15, 21], [21, 24], [24, 29], [30, 33], [33, 35], [35, 36], [36, 39], [39, 42], [42, 46], [46, 49], [49, 53], [54, 55], [55, 60]]}
93
+ {"id": 92, "text": "ไหวบ่ คนเยอะเด้ต้นเดือนวันหยุด วันหลังดีก่างับ", "gold": [[0, 3], [3, 5], [6, 8], [8, 12], [12, 15], [15, 23], [23, 30], [31, 38], [38, 40], [40, 43], [43, 46]]}
94
+ {"id": 93, "text": "ทักมาค่ะ", "gold": [[0, 3], [3, 5], [5, 8]]}
95
+ {"id": 94, "text": "พาไปหน่อยสิ อยากกิน", "gold": [[0, 2], [2, 4], [4, 9], [9, 11], [12, 16], [16, 19]]}
96
+ {"id": 95, "text": "เอาตัง800ไปสั่งเป็ดย่างเอ็มเคกิน ฟินกว่าเยอะ มีทอนอีกนะ", "gold": [[0, 3], [3, 6], [6, 9], [9, 11], [11, 15], [15, 19], [19, 23], [23, 29], [29, 32], [33, 36], [36, 40], [40, 44], [45, 47], [47, 50], [50, 53], [53, 55]]}
97
+ {"id": 96, "text": "ตกเป็นทาสการตลาด😑😑😓", "gold": [[0, 2], [2, 6], [6, 9], [9, 12], [12, 16], [16, 17], [17, 18], [18, 19]]}
98
+ {"id": 97, "text": "คันจริงมันสวยมากๆๆๆๆๆๆๆนะ เริ่มเห็นบ่อ��กว่า Toyota Noah Voxy ด้วย555 มีครบ รุ่นถูก แพง มาหมด #แต่NissanThaiทำอะไรอยู่", "gold": [[0, 3], [3, 7], [7, 10], [10, 13], [13, 16], [16, 17], [17, 18], [18, 19], [19, 20], [20, 21], [21, 22], [22, 23], [23, 25], [26, 31], [31, 35], [35, 39], [39, 43], [44, 50], [51, 55], [56, 60], [61, 65], [65, 68], [69, 71], [71, 74], [75, 79], [79, 82], [83, 86], [87, 89], [89, 92], [93, 94], [94, 97], [97, 103], [103, 107], [107, 109], [109, 113], [113, 117]]}
99
+ {"id": 98, "text": "Lamborghini Huracan คันนี้ไม่เดิม ซ่อนกำลัง 1,000 แรงม้า แหวกอากาศที่ความเร็วทะลุ 300 กม./ชม.", "gold": [[0, 11], [12, 19], [20, 23], [23, 26], [26, 29], [29, 33], [34, 38], [38, 43], [44, 49], [50, 56], [57, 61], [61, 66], [66, 69], [69, 77], [77, 81], [82, 85], [86, 93]]}
100
+ {"id": 99, "text": "นั่งอ่านทัศนคติของผู้บริหารแสนสิริแล้วชอบมาก อยากเข้าทำงานที่นี่", "gold": [[0, 4], [4, 8], [8, 15], [15, 18], [18, 27], [27, 34], [34, 38], [38, 41], [41, 44], [45, 49], [49, 53], [53, 58], [58, 61], [61, 64]]}
101
+ {"id": 100, "text": "ชอบ Honda Civic เพราะมันคล่องตัวและกว้างเหมาะกับการใช้งาน Eveeve Eve แกไปดูกะชั้นนะ #ฮอนด้าที่รัก #EnjoyHondaThailand", "gold": [[0, 3], [4, 9], [10, 15], [16, 21], [21, 24], [24, 32], [32, 35], [35, 40], [40, 45], [45, 48], [48, 51], [51, 54], [54, 57], [58, 64], [65, 68], [69, 71], [71, 73], [73, 75], [75, 77], [77, 81], [81, 83], [84, 85], [85, 91], [91, 97], [98, 99], [99, 104], [104, 109], [109, 117]]}
102
+ {"id": 101, "text": "ขอกราบตรงหัวใจผู้บริหารโตโยต้าทุกๆท่านด้วยความเต็มใจที่เห็นคุณค่าของคนไทยที่ทำชื่อเสียงให้กับประเทศชาติ", "gold": [[0, 2], [2, 6], [6, 9], [9, 14], [14, 23], [23, 30], [30, 33], [33, 34], [34, 38], [38, 42], [42, 46], [46, 50], [50, 52], [52, 55], [55, 59], [59, 65], [65, 68], [68, 73], [73, 76], [76, 78], [78, 82], [82, 87], [87, 90], [90, 93], [93, 103]]}
103
+ {"id": 102, "text": "หนวดน่ากินนมากค่าาาาา", "gold": [[0, 4], [4, 7], [7, 11], [11, 14], [14, 21]]}
104
+ {"id": 103, "text": "วีออสสวยคับ", "gold": [[0, 5], [5, 8], [8, 11]]}
105
+ {"id": 104, "text": "วันพุธนี้ไปกินกันมั้ย #เวสเกต", "gold": [[0, 3], [3, 6], [6, 9], [9, 11], [11, 14], [14, 17], [17, 21], [22, 29]]}
106
+ {"id": 105, "text": "ซื้อไว้มั้ยย", "gold": [[0, 4], [4, 7], [7, 12]]}
107
+ {"id": 106, "text": "ไปพรุ่งนี้ไหม🤔", "gold": [[0, 2], [2, 7], [7, 10], [10, 13], [13, 14]]}
108
+ {"id": 107, "text": "หิวละ. ปะฮอตพอต", "gold": [[0, 3], [3, 5], [5, 6], [7, 9], [9, 15]]}
109
+ {"id": 108, "text": "ลงชื่อด้วยคนคับ", "gold": [[0, 2], [2, 6], [6, 10], [10, 12], [12, 15]]}
110
+ {"id": 109, "text": "เหมือน", "gold": [[0, 6]]}
111
+ {"id": 110, "text": "ถ้าบุฟเฟ่ก็ต้องไปบาบิก้อน", "gold": [[0, 3], [3, 9], [9, 11], [11, 15], [15, 17], [17, 25]]}
112
+ {"id": 111, "text": "วันนี้ปะ555", "gold": [[0, 6], [6, 8], [8, 11]]}
113
+ {"id": 112, "text": "วรรณษา จรัลพรไพศาล มาลองใช้ลอรีเอะ ซอฟท์แอนด์เซฟ เอ็กซ์ตร้าโพเทคชั่น แอนตี้แบคทีเรียใส่ตอนนอนสิ แล้วจะหลงรัก แผ่นใหญ่ปกป้องดีเยี่ยม ลดการสะสมของแบคทีเรีย ไร้กลิ่นอับชื้นกวนใจไร้กังวล บล็อกหยุดการซึมเปื้อน หลับสบายไม่ต้องกังวลว่าจะเลอะที่นอน ตื่นมาสดชื่นนนนน", "gold": [[0, 6], [7, 18], [19, 21], [21, 24], [24, 27], [27, 34], [35, 40], [40, 45], [45, 48], [49, 59], [59, 68], [69, 75], [75, 84], [84, 87], [87, 90], [90, 93], [93, 95], [96, 100], [100, 102], [102, 108], [109, 113], [113, 117], [117, 123], [123, 125], [125, 131], [132, 134], [134, 137], [137, 141], [141, 144], [144, 153], [154, 157], [157, 162], [162, 165], [165, 169], [169, 174], [174, 177], [177, 182], [183, 188], [188, 192], [192, 195], [195, 198], [198, 204], [205, 209], [209, 213], [213, 216], [216, 220], [220, 225], [225, 228], [228, 230], [230, 234], [234, 237], [237, 240], [241, 245], [245, 247], [247, 249], [249, 257]]}
114
+ {"id": 113, "text": "Spyรสชาติองุ่นอร่อยซ่าไม่คลื่นไม่เมาบอกเลยว่าชอบมากๆ", "gold": [[0, 3], [3, 9], [9, 14], [14, 19], [19, 22], [22, 25], [25, 30], [30, 33], [33, 36], [36, 39], [39, 42], [42, 45], [45, 48], [48, 51], [51, 52]]}
115
+ {"id": 114, "text": "#ลอรีเอะเฟรชคอนโทรล #30 cmซื้อง่ายที่เซ่เว่นค่ะ #ล็อคกลิ่นมั่นใจไม่อับชื้น #นี่สิคอนโทรลได้ทั้งคืน", "gold": [[0, 8], [8, 12], [12, 19], [20, 23], [24, 26], [26, 30], [30, 34], [34, 37], [37, 44], [44, 47], [48, 53], [53, 58], [58, 62], [62, 64], [64, 67], [67, 74], [75, 79], [79, 81], [81, 88], [88, 91], [91, 98]]}
116
+ {"id": 115, "text": "กิน Mk ดีก่า", "gold": [[0, 3], [4, 6], [7, 9], [9, 12]]}
117
+ {"id": 116, "text": "แต่ก่อนช้าเด่วนี้เร็วข้นนะ", "gold": [[0, 3], [3, 7], [7, 10], [10, 14], [14, 17], [17, 21], [21, 24], [24, 26]]}
118
+ {"id": 117, "text": "ไป!", "gold": [[0, 2], [2, 3]]}
119
+ {"id": 118, "text": "ปะจัดเลย", "gold": [[0, 2], [2, 5], [5, 8]]}
120
+ {"id": 119, "text": "เดือนหน้าค่อยไป อิอิ", "gold": [[0, 5], [5, 9], [9, 13], [13, 15], [16, 20]]}
121
+ {"id": 120, "text": "ขอนุญาตินอกเรื่องกูข้องใจนานละ เมื่อไหร่sextoy บุหรี่ไฟฟ้าจะถูกกฏหมายแบบญี่ปุ่นหรือประเทศอื่น", "gold": [[0, 2], [2, 8], [8, 17], [17, 19], [19, 25], [25, 28], [28, 30], [31, 40], [40, 46], [47, 53], [53, 58], [58, 60], [60, 63], [63, 69], [69, 72], [72, 79], [79, 83], [83, 89], [89, 93]]}
122
+ {"id": 121, "text": "เพอริเพอร่าไปยังค่ะ", "gold": [[0, 11], [11, 13], [13, 16], [16, 19]]}
123
+ {"id": 122, "text": "ผมซื้อวีออส 2006 มือสองมา ต้องดูแลอะไรเป็นพิเศษมั้ยครับ วิ่งมาแสนเก้ากว่าๆครับ", "gold": [[0, 2], [2, 6], [6, 11], [12, 16], [17, 23], [23, 25], [26, 30], [30, 34], [34, 38], [38, 42], [42, 47], [47, 51], [51, 55], [56, 60], [60, 62], [62, 69], [69, 73], [73, 74], [74, 78]]}
124
+ {"id": 123, "text": "เอานิสสันมาคร์เทรินได้ไหมเหลืออีกแสนนิดๆ", "gold": [[0, 3], [3, 14], [14, 19], [19, 22], [22, 25], [25, 30], [30, 33], [33, 36], [36, 39], [39, 40]]}
125
+ {"id": 124, "text": "หุ้มพวงมาลัย มาสด้า 3 เท่าไรครับ", "gold": [[0, 4], [4, 12], [13, 19], [20, 21], [22, 28], [28, 32]]}
126
+ {"id": 125, "text": "ถามครับยาริสตัวใหม่ที่ผมใช้อยู่มีที่ปะยางแต่ท่าผมต้องการจะเพิ่มยางอะไหล่เขาไปมันจะ ส่ตรงที่เก็บที่ประยางได้มัยครับ..ขอคุณครับ", "gold": [[0, 3], [3, 7], [7, 12], [12, 19], [19, 22], [22, 24], [24, 27], [27, 31], [31, 33], [33, 36], [36, 41], [41, 44], [44, 47], [47, 49], [49, 53], [53, 56], [56, 58], [58, 63], [63, 66], [66, 72], [72, 75], [75, 77], [77, 80], [80, 82], [83, 85], [85, 88], [88, 91], [91, 95], [95, 98], [98, 104], [104, 107], [107, 110], [110, 114], [114, 116], [116, 118], [118, 121], [121, 125]]}
127
+ {"id": 126, "text": "ลิปetude ราคาเท่าไหร่ สีเบอร์อะไรคะ", "gold": [[0, 3], [3, 8], [9, 13], [13, 21], [22, 24], [24, 29], [29, 33], [33, 35]]}
128
+ {"id": 127, "text": "ดีมั้ยคะสาวกคิวเพลส", "gold": [[0, 2], [2, 6], [6, 8], [8, 12], [12, 19]]}
129
+ {"id": 128, "text": "สอบถามหน่อยนะคะ บลัชออนของ 3ce สีอะไรคะ", "gold": [[0, 3], [3, 6], [6, 11], [11, 13], [13, 15], [16, 23], [23, 26], [27, 30], [31, 33], [33, 37], [37, 39]]}
130
+ {"id": 129, "text": "ชุดนี้ใส่รีโว่4ประตูได้ป่ะคับ", "gold": [[0, 6], [6, 9], [9, 14], [14, 15], [15, 20], [20, 23], [23, 26], [26, 29]]}
131
+ {"id": 130, "text": "ปิคอัพ โตโยต้า ไทเกอร์ แคป ราคาเท่าไหร่คับ", "gold": [[0, 6], [7, 14], [15, 22], [23, 26], [27, 31], [31, 39], [39, 42]]}
132
+ {"id": 131, "text": "สาขาซีคอนสแควร์ศรีนครินทร์ปรับปรุงเสร็จรึยังครับ", "gold": [[0, 4], [4, 15], [15, 26], [26, 34], [34, 39], [39, 41], [41, 44], [44, 48]]}
133
+ {"id": 132, "text": "นาวาร่า 4D 2012 มีไหมครับ", "gold": [[0, 7], [8, 10], [11, 15], [16, 18], [18, 21], [21, 25]]}
134
+ {"id": 133, "text": "เปลี่ยนยาง ช่สงนี้เท่าไหร่คะ ฮอนด้าซิตี้ค่ะ", "gold": [[0, 7], [7, 10], [11, 15], [15, 18], [18, 26], [26, 28], [29, 40], [40, 43]]}
135
+ {"id": 134, "text": "Mazda cx8 มีโอกาสเข้ามาในไทยมั้ยครับ", "gold": [[0, 5], [6, 9], [10, 12], [12, 17], [17, 21], [21, 23], [23, 25], [25, 28], [28, 32], [32, 36]]}
136
+ {"id": 135, "text": "ถามหน่อยครับแล้วนิสสัน โน๊ตตัวไฟฟ้าไม่มาแล้วเหรอครับ", "gold": [[0, 3], [3, 8], [8, 12], [12, 16], [16, 22], [23, 27], [27, 30], [30, 35], [35, 38], [38, 40], [40, 44], [44, 48], [48, 52]]}
137
+ {"id": 136, "text": "ขอรายลพเอียดมาสด้า2ค่ะ", "gold": [[0, 2], [2, 5], [5, 7], [7, 12], [12, 18], [18, 19], [19, 22]]}
138
+ {"id": 137, "text": "มีใส่ mu-x บ้างไหมครับ", "gold": [[0, 2], [2, 5], [6, 10], [11, 15], [15, 18], [18, 22]]}
139
+ {"id": 138, "text": "ที่ไหน biothermกูละ", "gold": [[0, 3], [3, 6], [7, 15], [15, 17], [17, 19]]}
140
+ {"id": 139, "text": "นกแอร์กับแอร์เอเชียเที่ยวบินไประนองยกเลิกมั้ยค่ะ", "gold": [[0, 6], [6, 9], [9, 19], [19, 25], [25, 28], [28, 30], [30, 35], [35, 41], [41, 45], [45, 48]]}
141
+ {"id": 140, "text": "รบกวนสอบถามหน่อยครับ ถ้าซื้อรถมาสด้าจะมีประกันภัย MPI ให้ทุกคันไหมครับ?", "gold": [[0, 5], [5, 11], [11, 16], [16, 20], [21, 24], [24, 28], [28, 30], [30, 36], [36, 38], [38, 40], [40, 49], [50, 53], [54, 57], [57, 60], [60, 63], [63, 66], [66, 70], [70, 71]]}
142
+ {"id": 141, "text": "ที่สาขาโรบินสันสระบุรีไม่มีแล้วหรือคร้า", "gold": [[0, 3], [3, 7], [7, 15], [15, 22], [22, 25], [25, 27], [27, 31], [31, 35], [35, 39]]}
143
+ {"id": 142, "text": "1000 สิทธิ์ทั่วประเทศใช่ไหมครับ", "gold": [[0, 4], [5, 11], [11, 15], [15, 21], [21, 24], [24, 27], [27, 31]]}
144
+ {"id": 143, "text": "กันแดดดีมั้ยคะ", "gold": [[0, 3], [3, 6], [6, 8], [8, 12], [12, 14]]}
145
+ {"id": 144, "text": "หิ้วเข้ามาจะโดนไหมคะใช้เอง", "gold": [[0, 4], [4, 8], [8, 10], [10, 12], [12, 15], [15, 18], [18, 20], [20, 23], [23, 26]]}
146
+ {"id": 145, "text": "ยาริส 2014 กับ​สวิฟ 2014 แนะนำหน่อยครับ​ คันไหนดี", "gold": [[0, 5], [6, 10], [11, 14], [14, 19], [20, 24], [25, 30], [30, 35], [35, 39], [39, 40], [41, 44], [44, 47], [47, 49]]}
147
+ {"id": 146, "text": "แนะนำลิปหน่อยค่ะ ตอนนี้ลิป peripera หมด ขอเป็นโทนธรรมชาติทาไปเรียนนะคะ😙 #ไว้รีวิวห้ามขายของโว้ยย", "gold": [[0, 5], [5, 8], [8, 13], [13, 16], [17, 20], [20, 23], [23, 26], [27, 35], [36, 39], [40, 42], [42, 46], [46, 49], [49, 57], [57, 59], [59, 61], [61, 66], [66, 68], [68, 70], [70, 71], [72, 96]]}
148
+ {"id": 147, "text": "อยุ่ตรงไหนของโรบินสันสระบุรีค่ะ", "gold": [[0, 4], [4, 7], [7, 10], [10, 13], [13, 21], [21, 28], [28, 31]]}
149
+ {"id": 148, "text": "ช่วยรีวิว estee vs bobbi brown ค่ะ อยากทราบความแตกต่าง คุมมันมั้ยคะ ขอบคุณค่ะ", "gold": [[0, 4], [4, 9], [10, 15], [16, 18], [19, 24], [25, 30], [31, 34], [35, 39], [39, 43], [43, 47], [47, 54], [55, 58], [58, 61], [61, 65], [65, 67], [68, 74], [74, 77]]}
150
+ {"id": 149, "text": "เพชรบุรีทำไมปิด", "gold": [[0, 8], [8, 12], [12, 15]]}
151
+ {"id": 150, "text": "ลิป the face shop รุ่นอะไรค่ะ", "gold": [[0, 3], [4, 7], [8, 12], [13, 17], [18, 22], [22, 26], [26, 29]]}
152
+ {"id": 151, "text": "ตัวเตี้ยรีโว่จะอัพหน้า อัพเกียร์จริงไหม?", "gold": [[0, 8], [8, 13], [13, 15], [15, 18], [18, 22], [23, 26], [26, 32], [32, 36], [36, 39], [39, 40]]}
153
+ {"id": 152, "text": "ดาวยาริส 100000 ส่งเท่าไหร่", "gold": [[0, 3], [3, 8], [9, 15], [16, 19], [19, 27]]}
154
+ {"id": 153, "text": "แล้ว แอคคอดไม่มาโชว์รึ", "gold": [[0, 4], [5, 11], [11, 14], [14, 16], [16, 20], [20, 22]]}
155
+ {"id": 154, "text": "พี่ครับทดลองขับJazzแล้วสะเทือนแรงไหมครับ", "gold": [[0, 3], [3, 7], [7, 12], [12, 15], [15, 19], [19, 23], [23, 30], [30, 33], [33, 36], [36, 40]]}
156
+ {"id": 155, "text": "โรบินสันเชียงใหม่มีปิ้งใหม่ค่ะ", "gold": [[0, 8], [8, 17], [17, 19], [19, 23], [23, 27], [27, 30]]}
157
+ {"id": 156, "text": "ใช้สิทธิได้ครั้งเดียวหรือคะ", "gold": [[0, 3], [3, 8], [8, 11], [11, 16], [16, 21], [21, 25], [25, 27]]}
158
+ {"id": 157, "text": "บลัชthe saemที่เกาหรีเท่าไหร่คะพี่นน กี่วอน", "gold": [[0, 4], [4, 7], [8, 12], [12, 15], [15, 21], [21, 29], [29, 31], [31, 36], [37, 40], [40, 43]]}
159
+ {"id": 158, "text": "แป้งฝุ่นลอร่าราคาเท่าไหร่ค่ะ", "gold": [[0, 8], [8, 13], [13, 17], [17, 25], [25, 28]]}
160
+ {"id": 159, "text": "อยากกิงบาบีก้อนบุฟอีกอ่ะ ทำไมมันยังไม่บุฟ คิดถึงที่ปีที่แล้วไปกินกะมึงง่ะ", "gold": [[0, 4], [4, 7], [7, 15], [15, 18], [18, 21], [21, 24], [25, 29], [29, 32], [32, 35], [35, 38], [38, 41], [42, 48], [48, 51], [51, 53], [53, 56], [56, 60], [60, 62], [62, 65], [65, 67], [67, 70], [70, 73]]}
fetch_sources.py ADDED
@@ -0,0 +1,65 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ """Download the original files of every test set (sources.json) into raw/ and verify their SHA-256.
2
+
3
+ python fetch_sources.py # everything (about 830 MB, most of it khammoon's training text)
4
+ python fetch_sources.py --no-khmdd # without khammoon's training text: every set except khmdd
5
+
6
+ Files already present with the right SHA-256 are not downloaded again. Standard library only. `fetch(raw, sets)`
7
+ downloads only the files of the given test sets (benchmark.py uses it on its first run).
8
+ """
9
+ import argparse
10
+ import hashlib
11
+ import json
12
+ import sys
13
+ import urllib.request
14
+ from pathlib import Path
15
+
16
+ HERE = Path(__file__).resolve().parent
17
+
18
+
19
+ def sha256(path: Path) -> str:
20
+ h = hashlib.sha256()
21
+ with path.open("rb") as f:
22
+ for chunk in iter(lambda: f.read(1 << 20), b""):
23
+ h.update(chunk)
24
+ return h.hexdigest()
25
+
26
+
27
+ def fetch(raw: Path, sets: set[str] | None = None, skip_khmdd: bool = False, quiet: bool = False) -> list[str]:
28
+ """Download the original files (of the given test sets, or all) into raw/; return the files whose SHA-256 differs."""
29
+ files = json.loads((HERE / "sources.json").read_text())["files"]
30
+ bad = []
31
+ for rel, info in files.items():
32
+ if sets is not None and not set(info["sets"]) & sets:
33
+ continue
34
+ if skip_khmdd and info["sets"] == ["khmdd"]:
35
+ continue
36
+ dest = raw / rel
37
+ if dest.exists() and sha256(dest) == info["sha256"]:
38
+ if not quiet:
39
+ print(f"ok {rel}")
40
+ continue
41
+ dest.parent.mkdir(parents=True, exist_ok=True)
42
+ print(f"fetch {rel} ({info['licence']})", flush=True)
43
+ req = urllib.request.Request(info["url"], headers={"User-Agent": "thai-ws-bench"})
44
+ with urllib.request.urlopen(req) as r, dest.open("wb") as f:
45
+ while chunk := r.read(1 << 20):
46
+ f.write(chunk)
47
+ if sha256(dest) != info["sha256"]:
48
+ bad.append(rel)
49
+ print(f"WRONG SHA-256: {rel} (the upstream file changed?)")
50
+ return bad
51
+
52
+
53
+ def main() -> None:
54
+ ap = argparse.ArgumentParser()
55
+ ap.add_argument("--raw", type=Path, default=HERE / "raw")
56
+ ap.add_argument("--no-khmdd", action="store_true", help="skip khammoon's training text (only needed for khmdd)")
57
+ args = ap.parse_args()
58
+ bad = fetch(args.raw, skip_khmdd=args.no_khmdd)
59
+ if bad:
60
+ sys.exit(f"{len(bad)} file(s) differ from the versions the benchmark was built from: {bad}")
61
+ print("all sources present and verified")
62
+
63
+
64
+ if __name__ == "__main__":
65
+ main()
predictions/attacut-1.0.6-c/bestdd.words.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:14caf15b62583da4933444a078509631ac416a0618341d637de2d3a134f8143d
3
+ size 812117
predictions/attacut-1.0.6-c/khm.words.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:601c07794b95dd1b4c6822eb1623d8c9e9cb63d112c59d60c75e3c90689d16d8
3
+ size 694066
predictions/attacut-1.0.6-c/khmdd.words.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ca3d6ee588779dffbc08588b07d1e4d5a5bfc28760d4b6fdd745fd64624ee8df
3
+ size 633238
predictions/attacut-1.0.6-c/orchidndd.words.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4860c2f259ac1bc8eaf1a50a17c4539d3ff23e37dd4279244bb61d4f84d6af5f
3
+ size 43513
predictions/attacut-1.0.6-c/pud.tokens.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9ce6cf5eeaf25abc233baa801fc4becb1f6d050458f0136e8105dacc0bf7cebb
3
+ size 73207
predictions/attacut-1.0.6-c/pud.words.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b6e52892b4277b300d3b21c7dfceef7ee44577fcf91506e1e67da451310ecd3b
3
+ size 32110
predictions/attacut-1.0.6-c/tud.tokens.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2eb2e94982bcbda602bd5919bdb38593172bd6a0946b73f34e46d29a8f82837c
3
+ size 24447
predictions/attacut-1.0.6-c/tud.words.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:76118c9c0d98ae0f1d1c0f4e88bd13578eeeed32448cc47234918949982df06a
3
+ size 12568
predictions/attacut-1.0.6-c/vistec.tokens.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:44af51241d6f9a32c35904ede8c15a1b7ec15251bc12e2bb05b20670d20eb810
3
+ size 2106937
predictions/attacut-1.0.6-c/vistec.words.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a39a0a865cd4f34697752968bbece8fd04a4709eb9f330cadb1194056c90f7e6
3
+ size 1004918
predictions/attacut-1.0.6-c/ws160.tokens.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e7dbec48a9cf2278b861cfe50eb15d97f231fad7e3b9de3310c35e52592a0392
3
+ size 12124
predictions/attacut-1.0.6-c/ws160.words.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fa1aaac55c4873a5651d9ae71507db6ee60f7c0331e1a4a53402df1db97c57c2
3
+ size 8157
predictions/attacut-1.0.6-sc/bestdd.words.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1f8a7bad0dd4880169472ee2edfdf1761f1379630c1add901c1861df895ccb1a
3
+ size 807391
predictions/attacut-1.0.6-sc/khm.words.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0135c5a7504bc69ec5dd175a3960dd9a9537c8431123f0d82f0c774ecdde4f7a
3
+ size 692277
predictions/attacut-1.0.6-sc/khmdd.words.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:70a60e15e190b625d2736f2173607a7b54f1a1d5bcfdbb148efce3dd4d4b8bab
3
+ size 631417
predictions/attacut-1.0.6-sc/orchidndd.words.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:218173a01f17ce8aaf7ade045b8d5b1d401238520b45ce13e75f301a43187d13
3
+ size 43365
predictions/attacut-1.0.6-sc/pud.tokens.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5e7b7bd801318a440c21bfd52b839beddbc509524496e030470ad1577583b6c0
3
+ size 73015
predictions/attacut-1.0.6-sc/pud.words.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dc828ca5577c44cad3ac4d2b9d7b614719b4584963affc3bab7dff267fa831aa
3
+ size 31915
predictions/attacut-1.0.6-sc/tud.tokens.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:41aec6daede2aeb3aedd7e79c90d0f0617eb548846c8dfbf804f0762fb7a8886
3
+ size 24443
predictions/attacut-1.0.6-sc/tud.words.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c8b45fb6e61e8d348fbd585ff7030d7ab5fcdf20a7a5cf787050dd6921f9a8f1
3
+ size 12569
predictions/attacut-1.0.6-sc/vistec.tokens.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2dce871b5e9768e9c0db84ca67d0792aee4715b19681211d2f5d45ee6eecc5eb
3
+ size 2103019
predictions/attacut-1.0.6-sc/vistec.words.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7656554b944a6df3e4362405495c435b3c8282fee8d9fc625349970978b35c7d
3
+ size 997645
predictions/attacut-1.0.6-sc/ws160.tokens.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4d601513a3a9a3c184fdc7bff8e53d09b89dd592ba5ae62c109779e522a4fec9
3
+ size 12103
predictions/attacut-1.0.6-sc/ws160.words.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:133697bbffb771d07c4c29d9ef6cbaca70bce8f0bef9394eb8cfcc58acca44b6
3
+ size 8098
predictions/deepcut-0.7.0.0/bestdd.words.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d7d69de8c505b3833d15e793d317b171b13d600ebaf35918ac3f925c35ba4937
3
+ size 802926
predictions/deepcut-0.7.0.0/khm.words.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b7aca397b45270d0cf34bb1b52b4522d548543ecb7cac2ab452b7a815da3ca2b
3
+ size 693065
predictions/deepcut-0.7.0.0/khmdd.words.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:255665ed05fc094652d9f778c2ed6ce7512e67e6c91d3db066243a8b38586dad
3
+ size 632283
predictions/deepcut-0.7.0.0/orchidndd.words.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dd4945269174264a3f8f0a0d4783a672e8cdf903d70297c07930977fd824835e
3
+ size 43513
predictions/deepcut-0.7.0.0/pud.tokens.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:61f7a0e444e3406f0faa994e3813d8863f2b2666a72e7d681c827e899d80308b
3
+ size 73027
predictions/deepcut-0.7.0.0/pud.words.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e78f4ba9b0da1c3f4b86d105f62fbf8ca574a99cb7fbb6dd85796bc0ba7dbb2
3
+ size 31959
predictions/deepcut-0.7.0.0/tud.tokens.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:36583c434ae4c0f79acafd05aaab0f151c8a0f0e8fd075b5d401d4007ed25484
3
+ size 24421
predictions/deepcut-0.7.0.0/tud.words.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bede1bc4bbb1ab7407866fd792ba5a821226ff3293ad23891ba1935fecde33c1
3
+ size 12522
predictions/deepcut-0.7.0.0/vistec.tokens.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5bff7ea472527252aedfcc93d9a433d92eddc23f32998065b8fc65aa12ce1967
3
+ size 2102990
predictions/deepcut-0.7.0.0/vistec.words.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0e4ad1e627357a53195f67861c07e1d443bbf010e175d717c207c1f4293f2212
3
+ size 986278
predictions/deepcut-0.7.0.0/ws160.tokens.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7621d362301e452f1847ba1659c121a82be1bb353d2f76717b5b5b213bddb110
3
+ size 12081
predictions/deepcut-0.7.0.0/ws160.words.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:002d37bd7494c0b4b236c73add9078f7a2f1f9cdfd8dd2a655a08af0fbb0bfc0
3
+ size 8108