akshara-tokenizer 1.1.0__tar.gz
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- akshara_tokenizer-1.1.0/LICENSE +194 -0
- akshara_tokenizer-1.1.0/NOTICE +7 -0
- akshara_tokenizer-1.1.0/PKG-INFO +436 -0
- akshara_tokenizer-1.1.0/README.md +210 -0
- akshara_tokenizer-1.1.0/akshara_tokenizer/__init__.py +4 -0
- akshara_tokenizer-1.1.0/akshara_tokenizer/boundary.py +120 -0
- akshara_tokenizer-1.1.0/akshara_tokenizer/model/akshara_tokenizer_v1_1.model +0 -0
- akshara_tokenizer-1.1.0/akshara_tokenizer/model/akshara_tokenizer_v1_1.vocab +16000 -0
- akshara_tokenizer-1.1.0/akshara_tokenizer/unicode_tables.py +156 -0
- akshara_tokenizer-1.1.0/akshara_tokenizer.egg-info/PKG-INFO +436 -0
- akshara_tokenizer-1.1.0/akshara_tokenizer.egg-info/SOURCES.txt +15 -0
- akshara_tokenizer-1.1.0/akshara_tokenizer.egg-info/dependency_links.txt +1 -0
- akshara_tokenizer-1.1.0/akshara_tokenizer.egg-info/requires.txt +6 -0
- akshara_tokenizer-1.1.0/akshara_tokenizer.egg-info/top_level.txt +1 -0
- akshara_tokenizer-1.1.0/pyproject.toml +71 -0
- akshara_tokenizer-1.1.0/setup.cfg +4 -0
- akshara_tokenizer-1.1.0/tests/test_boundary.py +206 -0
|
@@ -0,0 +1,194 @@
|
|
|
1
|
+
Apache License
|
|
2
|
+
Version 2.0, January 2004
|
|
3
|
+
http://www.apache.org/licenses/
|
|
4
|
+
|
|
5
|
+
TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
|
|
6
|
+
|
|
7
|
+
1. Definitions.
|
|
8
|
+
|
|
9
|
+
"License" shall mean the terms and conditions for use, reproduction,
|
|
10
|
+
and distribution as defined by Sections 1 through 9 of this document.
|
|
11
|
+
|
|
12
|
+
"Licensor" shall mean the copyright owner or entity authorized by
|
|
13
|
+
the copyright owner that is granting the License.
|
|
14
|
+
|
|
15
|
+
"Legal Entity" shall mean the union of the acting entity and all
|
|
16
|
+
other entities that control, are controlled by, or are under common
|
|
17
|
+
control with that entity. For the purposes of this definition,
|
|
18
|
+
"control" means (i) the power, direct or indirect, to cause the
|
|
19
|
+
direction or management of such entity, whether by contract or
|
|
20
|
+
otherwise, or (ii) ownership of fifty percent (50%) or more of the
|
|
21
|
+
outstanding shares, or (iii) beneficial ownership of such entity.
|
|
22
|
+
|
|
23
|
+
"You" (or "Your") shall mean an individual or Legal Entity
|
|
24
|
+
exercising permissions granted by this License.
|
|
25
|
+
|
|
26
|
+
"Source" form shall mean the preferred form for making modifications,
|
|
27
|
+
including but not limited to software source code, documentation
|
|
28
|
+
source, and configuration files.
|
|
29
|
+
|
|
30
|
+
"Object" form shall mean any form resulting from mechanical
|
|
31
|
+
transformation or translation of a Source form, including but
|
|
32
|
+
not limited to compiled object code, generated documentation,
|
|
33
|
+
and conversions to other media types.
|
|
34
|
+
|
|
35
|
+
"Work" shall mean the work of authorship made available under
|
|
36
|
+
the License, as indicated by a copyright notice that is included in
|
|
37
|
+
or attached to the work (an example is provided in the Appendix below).
|
|
38
|
+
|
|
39
|
+
"Derivative Works" shall mean any work, whether in Source or Object
|
|
40
|
+
form, that is based on (or derived from) the Work and for which the
|
|
41
|
+
editorial revisions, annotations, elaborations, or other modifications
|
|
42
|
+
represent, as a whole, an original work of authorship. For the purposes
|
|
43
|
+
of this License, Derivative Works shall not include works that remain
|
|
44
|
+
separable from, or merely link (or bind by name) to the interfaces of,
|
|
45
|
+
the Work and Derivative Works thereof.
|
|
46
|
+
|
|
47
|
+
"Contribution" shall mean, as submitted to the Licensor for inclusion
|
|
48
|
+
in the Work by the copyright owner or by an individual or Legal Entity
|
|
49
|
+
authorized to submit on behalf of the copyright owner. For the purposes
|
|
50
|
+
of this definition, "submitted" means any form of electronic, verbal,
|
|
51
|
+
or written communication sent to the Licensor or its representatives,
|
|
52
|
+
including but not limited to communication on electronic mailing lists,
|
|
53
|
+
source code control systems, and issue tracking systems that are managed
|
|
54
|
+
by, or on behalf of, the Licensor for the purpose of discussing and
|
|
55
|
+
improving the Work, but excluding communication that is conspicuously
|
|
56
|
+
marked or designated in writing by the copyright owner as "Not a
|
|
57
|
+
Contribution."
|
|
58
|
+
|
|
59
|
+
"Contributor" shall mean Licensor and any Legal Entity on behalf of
|
|
60
|
+
whom a Contribution has been received by the Licensor and subsequently
|
|
61
|
+
incorporated within the Work.
|
|
62
|
+
|
|
63
|
+
2. Grant of Copyright License. Subject to the terms and conditions of
|
|
64
|
+
this License, each Contributor hereby grants to You a perpetual,
|
|
65
|
+
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
|
|
66
|
+
copyright license to reproduce, prepare Derivative Works of,
|
|
67
|
+
publicly display, publicly perform, sublicense, and distribute the
|
|
68
|
+
Work and such Derivative Works in Source or Object form.
|
|
69
|
+
|
|
70
|
+
3. Grant of Patent License. Subject to the terms and conditions of
|
|
71
|
+
this License, each Contributor hereby grants to You a perpetual,
|
|
72
|
+
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
|
|
73
|
+
(except as stated in this section) patent license to make, have made,
|
|
74
|
+
use, offer to sell, sell, import, and otherwise transfer the Work,
|
|
75
|
+
where such license applies only to those patent claims licensable
|
|
76
|
+
by such Contributor that are necessarily infringed by their
|
|
77
|
+
Contribution(s) alone or by combination of their Contribution(s)
|
|
78
|
+
with the Work to which such Contribution(s) was submitted. If You
|
|
79
|
+
institute patent litigation against any entity (including a cross-claim
|
|
80
|
+
or counterclaim in a lawsuit) alleging that the Work or any
|
|
81
|
+
Contribution embodied within the Work constitutes direct or
|
|
82
|
+
contributory patent infringement, then any patent licenses granted to
|
|
83
|
+
You under this License for that Work shall terminate as of the date
|
|
84
|
+
such litigation is filed.
|
|
85
|
+
|
|
86
|
+
4. Redistribution. You may reproduce and distribute copies of the
|
|
87
|
+
Work or Derivative Works thereof in any medium, with or without
|
|
88
|
+
modifications, and in Source or Object form, provided that You
|
|
89
|
+
meet the following conditions:
|
|
90
|
+
|
|
91
|
+
(a) You must give any other recipients of the Work or
|
|
92
|
+
Derivative Works a copy of this License; and
|
|
93
|
+
|
|
94
|
+
(b) You must cause any modified files to carry prominent notices
|
|
95
|
+
stating that You changed the files; and
|
|
96
|
+
|
|
97
|
+
(c) You must retain, in the Source form of any Derivative Works
|
|
98
|
+
that You distribute, all copyright, patent, trademark, and
|
|
99
|
+
attribution notices from the Source form of the Work,
|
|
100
|
+
excluding those notices that do not pertain to any part of
|
|
101
|
+
the Derivative Works; and
|
|
102
|
+
|
|
103
|
+
(d) If the Work includes a "NOTICE" text file as part of its
|
|
104
|
+
distribution, You must include a readable copy of the
|
|
105
|
+
attribution notices contained within such NOTICE file, in
|
|
106
|
+
at least one of the following places: within a NOTICE text
|
|
107
|
+
file distributed as part of the Derivative Works; within
|
|
108
|
+
the Source form or documentation, if provided along with the
|
|
109
|
+
Derivative Works; or, within a display generated by the
|
|
110
|
+
Derivative Works, if and wherever such third-party notices
|
|
111
|
+
normally appear. The contents of the NOTICE file are for
|
|
112
|
+
informational purposes only and do not modify the License.
|
|
113
|
+
You may add Your own attribution notices within Derivative
|
|
114
|
+
Works that You distribute, alongside or as an addendum to
|
|
115
|
+
the NOTICE text from the Work, provided that such additional
|
|
116
|
+
attribution notices cannot be construed as modifying the License.
|
|
117
|
+
|
|
118
|
+
You may add Your own license statement for Your modifications and
|
|
119
|
+
may provide additional grant of rights to use, reproduce, modify,
|
|
120
|
+
prepare Derivative Works of, distribute, and sublicense such
|
|
121
|
+
modifications and Derivative Works as a whole, provided Your use,
|
|
122
|
+
reproduction, and distribution of the Work otherwise complies with
|
|
123
|
+
the conditions stated in this License.
|
|
124
|
+
|
|
125
|
+
5. Submission of Contributions. Unless You explicitly state otherwise,
|
|
126
|
+
any Contribution intentionally submitted for inclusion in the Work
|
|
127
|
+
by You to the Licensor shall be under the terms and conditions of
|
|
128
|
+
this License, without any additional terms or conditions.
|
|
129
|
+
Notwithstanding the above, nothing herein shall supersede or modify
|
|
130
|
+
the terms of any separate license agreement you may have executed
|
|
131
|
+
with Licensor regarding such Contributions.
|
|
132
|
+
|
|
133
|
+
6. Trademarks. This License does not grant permission to use the trade
|
|
134
|
+
names, trademarks, service marks, or product names of the Licensor,
|
|
135
|
+
except as required for reasonable and customary use in describing the
|
|
136
|
+
origin of the Work and reproducing the content of the NOTICE file.
|
|
137
|
+
|
|
138
|
+
7. Disclaimer of Warranty. Unless required by applicable law or
|
|
139
|
+
agreed to in writing, Licensor provides the Work (and each
|
|
140
|
+
Contributor provides its Contributions) on an "AS IS" BASIS,
|
|
141
|
+
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
|
|
142
|
+
implied, including, without limitation, any warranties or conditions
|
|
143
|
+
of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
|
|
144
|
+
PARTICULAR PURPOSE. You are solely responsible for determining the
|
|
145
|
+
appropriateness of using or reproducing the Work and assume any
|
|
146
|
+
risks associated with Your exercise of permissions under this License.
|
|
147
|
+
|
|
148
|
+
8. Limitation of Liability. In no event and under no legal theory,
|
|
149
|
+
whether in tort (including negligence), contract, or otherwise,
|
|
150
|
+
unless required by applicable law (such as deliberate and grossly
|
|
151
|
+
negligent acts) or agreed to in writing, shall any Contributor be
|
|
152
|
+
liable to You for damages, including any direct, indirect, special,
|
|
153
|
+
incidental, or exemplary damages of any character arising as a
|
|
154
|
+
result of this License or out of the use or inability to use the
|
|
155
|
+
Work (including but not limited to damages for loss of goodwill,
|
|
156
|
+
work stoppage, computer failure or malfunction, or all other
|
|
157
|
+
commercial damages or losses), even if such Contributor has been
|
|
158
|
+
advised of the possibility of such damages.
|
|
159
|
+
|
|
160
|
+
9. Accepting Warranty or Additional Liability. While redistributing
|
|
161
|
+
the Work or Derivative Works thereof, You may choose to offer, and
|
|
162
|
+
charge a fee for, acceptance of support, warranty, indemnity,
|
|
163
|
+
or other liability obligations and/or rights consistent with this
|
|
164
|
+
License. However, in accepting such obligations, You may act only
|
|
165
|
+
on Your own behalf and on Your sole responsibility, not on behalf
|
|
166
|
+
of any other Contributor, and only if You agree to indemnify,
|
|
167
|
+
defend, and hold each Contributor harmless for any liability
|
|
168
|
+
incurred by, or claims asserted against, such Contributor by reason
|
|
169
|
+
of your accepting any such warranty or additional liability.
|
|
170
|
+
|
|
171
|
+
END OF TERMS AND CONDITIONS
|
|
172
|
+
|
|
173
|
+
APPENDIX: How to apply the Apache License to your work.
|
|
174
|
+
|
|
175
|
+
To apply the Apache License to your work, attach the following
|
|
176
|
+
boilerplate notice, with the fields enclosed by brackets "[]"
|
|
177
|
+
replaced with your own identifying information. (Don't include
|
|
178
|
+
the brackets!) The text should be enclosed in the appropriate
|
|
179
|
+
comment syntax for the file format in question. Please also include
|
|
180
|
+
a "NOTICE" file with your name and affiliation.
|
|
181
|
+
|
|
182
|
+
Copyright 2026 Gursimran Singh
|
|
183
|
+
|
|
184
|
+
Licensed under the Apache License, Version 2.0 (the "License");
|
|
185
|
+
you may not use this file except in compliance with the License.
|
|
186
|
+
You may obtain a copy of the License at
|
|
187
|
+
|
|
188
|
+
http://www.apache.org/licenses/LICENSE-2.0
|
|
189
|
+
|
|
190
|
+
Unless required by applicable law or agreed to in writing, software
|
|
191
|
+
distributed under the License is distributed on an "AS IS" BASIS,
|
|
192
|
+
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
|
|
193
|
+
implied. See the License for the specific language governing
|
|
194
|
+
permissions and limitations under the License.
|
|
@@ -0,0 +1,7 @@
|
|
|
1
|
+
IYRA Akshara Tokenizer
|
|
2
|
+
Copyright 2026 Gursimran Singh
|
|
3
|
+
|
|
4
|
+
This product is licensed under the Apache License, Version 2.0.
|
|
5
|
+
You may obtain a copy of the License at http://www.apache.org/licenses/LICENSE-2.0
|
|
6
|
+
|
|
7
|
+
Patent pending: Indian provisional patent application 202611071450.
|
|
@@ -0,0 +1,436 @@
|
|
|
1
|
+
Metadata-Version: 2.4
|
|
2
|
+
Name: akshara-tokenizer
|
|
3
|
+
Version: 1.1.0
|
|
4
|
+
Summary: Linguistically-correct Akshara boundary tokenizer for Brahmic scripts
|
|
5
|
+
Author-email: Gursimran Singh <gursimranbasra1322@gmail.com>
|
|
6
|
+
License: Apache License
|
|
7
|
+
Version 2.0, January 2004
|
|
8
|
+
http://www.apache.org/licenses/
|
|
9
|
+
|
|
10
|
+
TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
|
|
11
|
+
|
|
12
|
+
1. Definitions.
|
|
13
|
+
|
|
14
|
+
"License" shall mean the terms and conditions for use, reproduction,
|
|
15
|
+
and distribution as defined by Sections 1 through 9 of this document.
|
|
16
|
+
|
|
17
|
+
"Licensor" shall mean the copyright owner or entity authorized by
|
|
18
|
+
the copyright owner that is granting the License.
|
|
19
|
+
|
|
20
|
+
"Legal Entity" shall mean the union of the acting entity and all
|
|
21
|
+
other entities that control, are controlled by, or are under common
|
|
22
|
+
control with that entity. For the purposes of this definition,
|
|
23
|
+
"control" means (i) the power, direct or indirect, to cause the
|
|
24
|
+
direction or management of such entity, whether by contract or
|
|
25
|
+
otherwise, or (ii) ownership of fifty percent (50%) or more of the
|
|
26
|
+
outstanding shares, or (iii) beneficial ownership of such entity.
|
|
27
|
+
|
|
28
|
+
"You" (or "Your") shall mean an individual or Legal Entity
|
|
29
|
+
exercising permissions granted by this License.
|
|
30
|
+
|
|
31
|
+
"Source" form shall mean the preferred form for making modifications,
|
|
32
|
+
including but not limited to software source code, documentation
|
|
33
|
+
source, and configuration files.
|
|
34
|
+
|
|
35
|
+
"Object" form shall mean any form resulting from mechanical
|
|
36
|
+
transformation or translation of a Source form, including but
|
|
37
|
+
not limited to compiled object code, generated documentation,
|
|
38
|
+
and conversions to other media types.
|
|
39
|
+
|
|
40
|
+
"Work" shall mean the work of authorship made available under
|
|
41
|
+
the License, as indicated by a copyright notice that is included in
|
|
42
|
+
or attached to the work (an example is provided in the Appendix below).
|
|
43
|
+
|
|
44
|
+
"Derivative Works" shall mean any work, whether in Source or Object
|
|
45
|
+
form, that is based on (or derived from) the Work and for which the
|
|
46
|
+
editorial revisions, annotations, elaborations, or other modifications
|
|
47
|
+
represent, as a whole, an original work of authorship. For the purposes
|
|
48
|
+
of this License, Derivative Works shall not include works that remain
|
|
49
|
+
separable from, or merely link (or bind by name) to the interfaces of,
|
|
50
|
+
the Work and Derivative Works thereof.
|
|
51
|
+
|
|
52
|
+
"Contribution" shall mean, as submitted to the Licensor for inclusion
|
|
53
|
+
in the Work by the copyright owner or by an individual or Legal Entity
|
|
54
|
+
authorized to submit on behalf of the copyright owner. For the purposes
|
|
55
|
+
of this definition, "submitted" means any form of electronic, verbal,
|
|
56
|
+
or written communication sent to the Licensor or its representatives,
|
|
57
|
+
including but not limited to communication on electronic mailing lists,
|
|
58
|
+
source code control systems, and issue tracking systems that are managed
|
|
59
|
+
by, or on behalf of, the Licensor for the purpose of discussing and
|
|
60
|
+
improving the Work, but excluding communication that is conspicuously
|
|
61
|
+
marked or designated in writing by the copyright owner as "Not a
|
|
62
|
+
Contribution."
|
|
63
|
+
|
|
64
|
+
"Contributor" shall mean Licensor and any Legal Entity on behalf of
|
|
65
|
+
whom a Contribution has been received by the Licensor and subsequently
|
|
66
|
+
incorporated within the Work.
|
|
67
|
+
|
|
68
|
+
2. Grant of Copyright License. Subject to the terms and conditions of
|
|
69
|
+
this License, each Contributor hereby grants to You a perpetual,
|
|
70
|
+
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
|
|
71
|
+
copyright license to reproduce, prepare Derivative Works of,
|
|
72
|
+
publicly display, publicly perform, sublicense, and distribute the
|
|
73
|
+
Work and such Derivative Works in Source or Object form.
|
|
74
|
+
|
|
75
|
+
3. Grant of Patent License. Subject to the terms and conditions of
|
|
76
|
+
this License, each Contributor hereby grants to You a perpetual,
|
|
77
|
+
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
|
|
78
|
+
(except as stated in this section) patent license to make, have made,
|
|
79
|
+
use, offer to sell, sell, import, and otherwise transfer the Work,
|
|
80
|
+
where such license applies only to those patent claims licensable
|
|
81
|
+
by such Contributor that are necessarily infringed by their
|
|
82
|
+
Contribution(s) alone or by combination of their Contribution(s)
|
|
83
|
+
with the Work to which such Contribution(s) was submitted. If You
|
|
84
|
+
institute patent litigation against any entity (including a cross-claim
|
|
85
|
+
or counterclaim in a lawsuit) alleging that the Work or any
|
|
86
|
+
Contribution embodied within the Work constitutes direct or
|
|
87
|
+
contributory patent infringement, then any patent licenses granted to
|
|
88
|
+
You under this License for that Work shall terminate as of the date
|
|
89
|
+
such litigation is filed.
|
|
90
|
+
|
|
91
|
+
4. Redistribution. You may reproduce and distribute copies of the
|
|
92
|
+
Work or Derivative Works thereof in any medium, with or without
|
|
93
|
+
modifications, and in Source or Object form, provided that You
|
|
94
|
+
meet the following conditions:
|
|
95
|
+
|
|
96
|
+
(a) You must give any other recipients of the Work or
|
|
97
|
+
Derivative Works a copy of this License; and
|
|
98
|
+
|
|
99
|
+
(b) You must cause any modified files to carry prominent notices
|
|
100
|
+
stating that You changed the files; and
|
|
101
|
+
|
|
102
|
+
(c) You must retain, in the Source form of any Derivative Works
|
|
103
|
+
that You distribute, all copyright, patent, trademark, and
|
|
104
|
+
attribution notices from the Source form of the Work,
|
|
105
|
+
excluding those notices that do not pertain to any part of
|
|
106
|
+
the Derivative Works; and
|
|
107
|
+
|
|
108
|
+
(d) If the Work includes a "NOTICE" text file as part of its
|
|
109
|
+
distribution, You must include a readable copy of the
|
|
110
|
+
attribution notices contained within such NOTICE file, in
|
|
111
|
+
at least one of the following places: within a NOTICE text
|
|
112
|
+
file distributed as part of the Derivative Works; within
|
|
113
|
+
the Source form or documentation, if provided along with the
|
|
114
|
+
Derivative Works; or, within a display generated by the
|
|
115
|
+
Derivative Works, if and wherever such third-party notices
|
|
116
|
+
normally appear. The contents of the NOTICE file are for
|
|
117
|
+
informational purposes only and do not modify the License.
|
|
118
|
+
You may add Your own attribution notices within Derivative
|
|
119
|
+
Works that You distribute, alongside or as an addendum to
|
|
120
|
+
the NOTICE text from the Work, provided that such additional
|
|
121
|
+
attribution notices cannot be construed as modifying the License.
|
|
122
|
+
|
|
123
|
+
You may add Your own license statement for Your modifications and
|
|
124
|
+
may provide additional grant of rights to use, reproduce, modify,
|
|
125
|
+
prepare Derivative Works of, distribute, and sublicense such
|
|
126
|
+
modifications and Derivative Works as a whole, provided Your use,
|
|
127
|
+
reproduction, and distribution of the Work otherwise complies with
|
|
128
|
+
the conditions stated in this License.
|
|
129
|
+
|
|
130
|
+
5. Submission of Contributions. Unless You explicitly state otherwise,
|
|
131
|
+
any Contribution intentionally submitted for inclusion in the Work
|
|
132
|
+
by You to the Licensor shall be under the terms and conditions of
|
|
133
|
+
this License, without any additional terms or conditions.
|
|
134
|
+
Notwithstanding the above, nothing herein shall supersede or modify
|
|
135
|
+
the terms of any separate license agreement you may have executed
|
|
136
|
+
with Licensor regarding such Contributions.
|
|
137
|
+
|
|
138
|
+
6. Trademarks. This License does not grant permission to use the trade
|
|
139
|
+
names, trademarks, service marks, or product names of the Licensor,
|
|
140
|
+
except as required for reasonable and customary use in describing the
|
|
141
|
+
origin of the Work and reproducing the content of the NOTICE file.
|
|
142
|
+
|
|
143
|
+
7. Disclaimer of Warranty. Unless required by applicable law or
|
|
144
|
+
agreed to in writing, Licensor provides the Work (and each
|
|
145
|
+
Contributor provides its Contributions) on an "AS IS" BASIS,
|
|
146
|
+
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
|
|
147
|
+
implied, including, without limitation, any warranties or conditions
|
|
148
|
+
of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
|
|
149
|
+
PARTICULAR PURPOSE. You are solely responsible for determining the
|
|
150
|
+
appropriateness of using or reproducing the Work and assume any
|
|
151
|
+
risks associated with Your exercise of permissions under this License.
|
|
152
|
+
|
|
153
|
+
8. Limitation of Liability. In no event and under no legal theory,
|
|
154
|
+
whether in tort (including negligence), contract, or otherwise,
|
|
155
|
+
unless required by applicable law (such as deliberate and grossly
|
|
156
|
+
negligent acts) or agreed to in writing, shall any Contributor be
|
|
157
|
+
liable to You for damages, including any direct, indirect, special,
|
|
158
|
+
incidental, or exemplary damages of any character arising as a
|
|
159
|
+
result of this License or out of the use or inability to use the
|
|
160
|
+
Work (including but not limited to damages for loss of goodwill,
|
|
161
|
+
work stoppage, computer failure or malfunction, or all other
|
|
162
|
+
commercial damages or losses), even if such Contributor has been
|
|
163
|
+
advised of the possibility of such damages.
|
|
164
|
+
|
|
165
|
+
9. Accepting Warranty or Additional Liability. While redistributing
|
|
166
|
+
the Work or Derivative Works thereof, You may choose to offer, and
|
|
167
|
+
charge a fee for, acceptance of support, warranty, indemnity,
|
|
168
|
+
or other liability obligations and/or rights consistent with this
|
|
169
|
+
License. However, in accepting such obligations, You may act only
|
|
170
|
+
on Your own behalf and on Your sole responsibility, not on behalf
|
|
171
|
+
of any other Contributor, and only if You agree to indemnify,
|
|
172
|
+
defend, and hold each Contributor harmless for any liability
|
|
173
|
+
incurred by, or claims asserted against, such Contributor by reason
|
|
174
|
+
of your accepting any such warranty or additional liability.
|
|
175
|
+
|
|
176
|
+
END OF TERMS AND CONDITIONS
|
|
177
|
+
|
|
178
|
+
APPENDIX: How to apply the Apache License to your work.
|
|
179
|
+
|
|
180
|
+
To apply the Apache License to your work, attach the following
|
|
181
|
+
boilerplate notice, with the fields enclosed by brackets "[]"
|
|
182
|
+
replaced with your own identifying information. (Don't include
|
|
183
|
+
the brackets!) The text should be enclosed in the appropriate
|
|
184
|
+
comment syntax for the file format in question. Please also include
|
|
185
|
+
a "NOTICE" file with your name and affiliation.
|
|
186
|
+
|
|
187
|
+
Copyright 2026 Gursimran Singh
|
|
188
|
+
|
|
189
|
+
Licensed under the Apache License, Version 2.0 (the "License");
|
|
190
|
+
you may not use this file except in compliance with the License.
|
|
191
|
+
You may obtain a copy of the License at
|
|
192
|
+
|
|
193
|
+
http://www.apache.org/licenses/LICENSE-2.0
|
|
194
|
+
|
|
195
|
+
Unless required by applicable law or agreed to in writing, software
|
|
196
|
+
distributed under the License is distributed on an "AS IS" BASIS,
|
|
197
|
+
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
|
|
198
|
+
implied. See the License for the specific language governing
|
|
199
|
+
permissions and limitations under the License.
|
|
200
|
+
|
|
201
|
+
Project-URL: Homepage, https://github.com/1322Guru/akshara-tokenizer
|
|
202
|
+
Project-URL: Repository, https://github.com/1322Guru/akshara-tokenizer
|
|
203
|
+
Project-URL: Bug Tracker, https://github.com/1322Guru/akshara-tokenizer/issues
|
|
204
|
+
Project-URL: HuggingFace, https://huggingface.co/GursimranSinghBasra/akshara-tokenizer
|
|
205
|
+
Keywords: nlp,tokenizer,indic,brahmic,hindi,punjabi,tamil,telugu,bengali,kannada,akshara,sentencepiece
|
|
206
|
+
Classifier: Development Status :: 3 - Alpha
|
|
207
|
+
Classifier: Intended Audience :: Science/Research
|
|
208
|
+
Classifier: License :: OSI Approved :: Apache Software License
|
|
209
|
+
Classifier: Programming Language :: Python :: 3
|
|
210
|
+
Classifier: Programming Language :: Python :: 3.10
|
|
211
|
+
Classifier: Programming Language :: Python :: 3.11
|
|
212
|
+
Classifier: Programming Language :: Python :: 3.12
|
|
213
|
+
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
|
|
214
|
+
Classifier: Topic :: Text Processing :: Linguistic
|
|
215
|
+
Classifier: Natural Language :: Hindi
|
|
216
|
+
Classifier: Natural Language :: Tamil
|
|
217
|
+
Requires-Python: >=3.10
|
|
218
|
+
Description-Content-Type: text/markdown
|
|
219
|
+
License-File: LICENSE
|
|
220
|
+
License-File: NOTICE
|
|
221
|
+
Provides-Extra: dev
|
|
222
|
+
Requires-Dist: pytest>=7.0; extra == "dev"
|
|
223
|
+
Provides-Extra: model
|
|
224
|
+
Requires-Dist: sentencepiece>=0.2.0; extra == "model"
|
|
225
|
+
Dynamic: license-file
|
|
226
|
+
|
|
227
|
+
# IYRA Akshara Tokenizer (IAT)
|
|
228
|
+
|
|
229
|
+
Linguistically-correct tokenization for Brahmic (Indic) scripts. It has two layers:
|
|
230
|
+
a rule-based Akshara segmenter that never splits an orthographic syllable, and a
|
|
231
|
+
trained SentencePiece model on top of it.
|
|
232
|
+
|
|
233
|
+
## Supported Scripts
|
|
234
|
+
|
|
235
|
+
Rule-based segmentation covers six scripts:
|
|
236
|
+
|
|
237
|
+
| Script | Languages |
|
|
238
|
+
|------------|--------------------------|
|
|
239
|
+
| Devanagari | Hindi, Sanskrit, Marathi |
|
|
240
|
+
| Gurmukhi | Punjabi |
|
|
241
|
+
| Tamil | Tamil |
|
|
242
|
+
| Telugu | Telugu |
|
|
243
|
+
| Bengali | Bengali |
|
|
244
|
+
| Kannada | Kannada |
|
|
245
|
+
|
|
246
|
+
## Why Akshara instead of BPE?
|
|
247
|
+
|
|
248
|
+
Byte-level BPE tokenizers split Brahmic text mid-character. A conjunct such as
|
|
249
|
+
ज्ञा, or a consonant plus vowel sign such as ਪੰ, is one orthographic syllable
|
|
250
|
+
(one Akshara), but BPE sees only UTF-8 bytes and cuts wherever its merges fall.
|
|
251
|
+
Captured output below: the same strings through the Qwen3-14B tokenizer and
|
|
252
|
+
through this tokenizer's segmenter. Each BPE token is decoded independently; a
|
|
253
|
+
`�` is a token that holds only a fragment of one character's bytes.
|
|
254
|
+
|
|
255
|
+
| Input | Qwen3-14B BPE tokens | Akshara segments |
|
|
256
|
+
|-------|----------------------|------------------|
|
|
257
|
+
| ज्ञान (Hindi) | `['ज', '्�', '�', 'ा�', '�']` (5 tokens) | `['ज्ञा', 'न']` (2 units) |
|
|
258
|
+
| ਪੰਜਾਬ (Punjabi) | `['ਪ', '�', '�', 'ਜ', '�', '�', 'ਬ']` (7 tokens) | `['ਪੰ', 'ਜਾ', 'ਬ']` (3 units) |
|
|
259
|
+
| நியாயம் (Tamil) | `['�', '�', 'ி�', '�', '�', '�', 'ய', 'ம', '்']` (9 tokens) | `['நி', 'யா', 'ய', 'ம்']` (4 units) |
|
|
260
|
+
|
|
261
|
+
The rule-based segmenter never splits an Akshara: that guarantee is exact, for
|
|
262
|
+
every input. At the SentencePiece layer the guarantee is statistical, not
|
|
263
|
+
absolute: aksharas too rare for the 16,000-piece vocabulary decompose into
|
|
264
|
+
sub-pieces or byte-fallback pieces, measured at 1.86 percent of aksharas overall
|
|
265
|
+
on FLORES-200 devtest (0.15 to 3.65 percent by script; see
|
|
266
|
+
`benchmark_2026_07/results_v1_1.md`). The fertility tables below quantify the
|
|
267
|
+
efficiency effect corpus-wide.
|
|
268
|
+
|
|
269
|
+
## Pipeline
|
|
270
|
+
|
|
271
|
+
Two layers: a rule-based Akshara segmenter (pure Unicode rules, no model), then
|
|
272
|
+
a trained SentencePiece model over the segmented sequence. One real example end
|
|
273
|
+
to end, with the actual ids from the shipped 16,000-piece v1.1 model:
|
|
274
|
+
|
|
275
|
+
```
|
|
276
|
+
Unicode text "न्याय दर्शन"
|
|
277
|
+
|
|
|
278
|
+
rule-based akshara segmenter
|
|
279
|
+
|
|
|
280
|
+
akshara sequence ['न्या', 'य', ' ', 'द', 'र्श', 'न']
|
|
281
|
+
|
|
|
282
|
+
SentencePiece (16k vocab, byte_fallback)
|
|
283
|
+
|
|
|
284
|
+
pieces ['▁न्या', '▁य', '▁द', '▁र्श', '▁न']
|
|
285
|
+
token ids [1734, 338, 364, 1398, 270]
|
|
286
|
+
```
|
|
287
|
+
|
|
288
|
+
## Model
|
|
289
|
+
|
|
290
|
+
This release ships one trained SentencePiece model (byte_fallback enabled):
|
|
291
|
+
|
|
292
|
+
| Model | Vocab | Trained scripts | File |
|
|
293
|
+
|-------|------:|-----------------|------|
|
|
294
|
+
| v1.1 (six-script) | 16,000 | all six, with the anusvara fix | `akshara_tokenizer/model/akshara_tokenizer_v1_1.model` |
|
|
295
|
+
|
|
296
|
+
An earlier 7,000-vocab model trained on Hindi, Punjabi, and Tamil was the originally
|
|
297
|
+
filed artifact and is not distributed in this release.
|
|
298
|
+
|
|
299
|
+
## Fertility (verified)
|
|
300
|
+
|
|
301
|
+
FLORES-200 devtest, tokens per whitespace word (lower is better). Numbers from
|
|
302
|
+
`benchmark_2026_07/results_v1_1.md`.
|
|
303
|
+
|
|
304
|
+
| Script (language) | v1.1 (16k) | Qwen3-14B |
|
|
305
|
+
|--------------------|-----------:|----------:|
|
|
306
|
+
| Devanagari (Hindi) | 2.45 | 4.77 |
|
|
307
|
+
| Gurmukhi (Punjabi) | 2.68 | 7.80 |
|
|
308
|
+
| Tamil (Tamil) | 5.01 | 10.06 |
|
|
309
|
+
| Telugu (Telugu) | 3.71 | 11.41 |
|
|
310
|
+
| Bengali (Bengali) | 3.38 | 7.16 |
|
|
311
|
+
| Kannada (Kannada) | 4.17 | 11.88 |
|
|
312
|
+
| English (info) | 5.08 | 1.26 |
|
|
313
|
+
|
|
314
|
+
|
|
315
|
+
|
|
316
|
+
Byte-fallback rate, percent (lower is better):
|
|
317
|
+
|
|
318
|
+
| Script | v1.1 (16k) |
|
|
319
|
+
|---------|-----------:|
|
|
320
|
+
| Hindi | 0.77 |
|
|
321
|
+
| Punjabi | 0.73 |
|
|
322
|
+
| Tamil | 0.02 |
|
|
323
|
+
| Telugu | 0.08 |
|
|
324
|
+
| Bengali | 0.09 |
|
|
325
|
+
| Kannada | 0.36 |
|
|
326
|
+
|
|
327
|
+
|
|
328
|
+
|
|
329
|
+
v1.1 fertility is below Qwen3-14B on all six Indic scripts, and its Telugu, Bengali,
|
|
330
|
+
and Kannada byte-fallback is under 0.4 percent. On English, Qwen3 is more efficient,
|
|
331
|
+
as expected for an English-centric byte-level BPE.
|
|
332
|
+
|
|
333
|
+
Both charts are generated from `benchmark_2026_07/results_v1_1.csv` by
|
|
334
|
+
`benchmark_2026_07/make_charts_v1_1.py`, so they can be reproduced from the
|
|
335
|
+
shipped data.
|
|
336
|
+
|
|
337
|
+
## Install
|
|
338
|
+
|
|
339
|
+
Requires Python >=3.10 (as declared in `pyproject.toml`).
|
|
340
|
+
|
|
341
|
+
```bash
|
|
342
|
+
git clone https://github.com/1322Guru/akshara-tokenizer
|
|
343
|
+
cd akshara-tokenizer
|
|
344
|
+
pip install . # rule-based segmenter only
|
|
345
|
+
pip install ".[model]" # includes sentencepiece, needed for the trained model
|
|
346
|
+
```
|
|
347
|
+
|
|
348
|
+
The base install is all you need for `segment_aksharas` and `count_aksharas`.
|
|
349
|
+
The `model` extra is needed to encode with the trained SentencePiece model.
|
|
350
|
+
|
|
351
|
+
## Usage
|
|
352
|
+
|
|
353
|
+
Rule-based segmentation (no model needed):
|
|
354
|
+
|
|
355
|
+
```python
|
|
356
|
+
from akshara_tokenizer import segment_aksharas, count_aksharas
|
|
357
|
+
|
|
358
|
+
segment_aksharas("ਨਿਆਯ ਦਰਸ਼ਨ")
|
|
359
|
+
# ["ਨਿ", "ਆ", "ਯ", " ", "ਦ", "ਰ", "ਸ਼", "ਨ"]
|
|
360
|
+
|
|
361
|
+
segment_aksharas("ज्ञान")
|
|
362
|
+
# ["ज्ञा", "न"]
|
|
363
|
+
|
|
364
|
+
count_aksharas("ਪੰਜਾਬ")
|
|
365
|
+
# 3
|
|
366
|
+
```
|
|
367
|
+
|
|
368
|
+
Encoding with the trained model (segment, then SentencePiece):
|
|
369
|
+
|
|
370
|
+
```python
|
|
371
|
+
import sentencepiece as spm
|
|
372
|
+
from importlib.resources import files, as_file
|
|
373
|
+
from akshara_tokenizer import segment_aksharas
|
|
374
|
+
|
|
375
|
+
# Load the model from the installed package (works after pip install, no repo path needed)
|
|
376
|
+
model_res = files("akshara_tokenizer").joinpath("model", "akshara_tokenizer_v1_1.model")
|
|
377
|
+
with as_file(model_res) as model_path:
|
|
378
|
+
sp = spm.SentencePieceProcessor(model_file=str(model_path))
|
|
379
|
+
|
|
380
|
+
ids = sp.encode(" ".join(segment_aksharas("न्याय दर्शन")))
|
|
381
|
+
```
|
|
382
|
+
|
|
383
|
+
## Akshara Formation Rules
|
|
384
|
+
|
|
385
|
+
1. C + virama + C forms a single conjunct (Devanagari, Telugu, Bengali, Kannada).
|
|
386
|
+
2. C + vowel sign forms one Akshara.
|
|
387
|
+
3. C + virama alone is a dead consonant and a complete Akshara.
|
|
388
|
+
4. An independent vowel is its own Akshara.
|
|
389
|
+
5. Anusvara, visarga, nukta, addak, tippi, bindi, and candrabindu attach to the
|
|
390
|
+
preceding Akshara.
|
|
391
|
+
6. Latin letters, digits, and punctuation are one token each.
|
|
392
|
+
|
|
393
|
+
## Limitations (honest)
|
|
394
|
+
|
|
395
|
+
- Round-trip is not lossless. The tokenizer is encode-optimized. `decode(encode(text))`
|
|
396
|
+
is not guaranteed to be byte-identical, because Akshara segmentation joins units with
|
|
397
|
+
spaces and SentencePiece normalizes whitespace. Use it for tokenization and fertility,
|
|
398
|
+
not for lossless reconstruction. A detokenizer is planned. (Verified: round-trip is not
|
|
399
|
+
byte-identical for any script in results_v1_1.md.)
|
|
400
|
+
- Script coverage. Telugu, Bengali, and Kannada are well-supported from v1.1 onward; the
|
|
401
|
+
earlier 7,000-vocab model covered only Hindi, Punjabi, and Tamil and is not shipped here.
|
|
402
|
+
|
|
403
|
+
## Run Tests
|
|
404
|
+
|
|
405
|
+
pytest is not part of the base install; get it via the dev extra, then run the suite:
|
|
406
|
+
|
|
407
|
+
```bash
|
|
408
|
+
pip install ".[dev]"
|
|
409
|
+
python -m pytest tests/ -v
|
|
410
|
+
```
|
|
411
|
+
|
|
412
|
+
## Script-level Notes
|
|
413
|
+
|
|
414
|
+
| Script | Conjuncts | Notes |
|
|
415
|
+
|------------|-----------|-------|
|
|
416
|
+
| Devanagari | Yes | Standard C + virama + C conjuncts |
|
|
417
|
+
| Gurmukhi | No | Virama creates dead consonants; addak doubles the next consonant |
|
|
418
|
+
| Tamil | No | Pulli always creates a dead consonant |
|
|
419
|
+
| Telugu | Yes | Conjuncts via virama |
|
|
420
|
+
| Bengali | Yes | Hasanta conjuncts |
|
|
421
|
+
| Kannada | Yes | Conjuncts via virama |
|
|
422
|
+
|
|
423
|
+
## Direction
|
|
424
|
+
|
|
425
|
+
Akshara-based segmentation is one instance of a more general idea: tokenizing
|
|
426
|
+
on the orthographic units a writing system is built from, rather than on bytes.
|
|
427
|
+
The same principle extends to other syllable-block writing systems.
|
|
428
|
+
|
|
429
|
+
## Patent and Copyright
|
|
430
|
+
|
|
431
|
+
Patent pending: Indian provisional patent application 202611071450.
|
|
432
|
+
Copyright 2026 Gursimran Singh.
|
|
433
|
+
|
|
434
|
+
## License
|
|
435
|
+
|
|
436
|
+
Apache License 2.0. See `LICENSE` and `NOTICE`.
|