interscript-maps 3.3.1 → 3.3.2

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.
@@ -0,0 +1,239 @@
1
+ system "ISO:prs-Arab:Latn:233-3-2026" {
2
+
3
+ metadata {
4
+ authority_id iso
5
+ id 233-3-2026
6
+ language iso-639-3:prs
7
+ source_script Arab
8
+ destination_script Latn
9
+ name "ISO 233-3:2026 Information and documentation -- Transliteration of Perso-Arabic characters into Latin characters -- Part 3: Persian language"
10
+ url https://www.iso.org/committee/48820.html
11
+ # Source document: interscript-private-references/iso/ISO-233-3-2026.pdf
12
+ # (purchased published edition, ISO Store order OP-1130631, downloaded
13
+ # 2026-11-19; verification-grade provenance).
14
+ creation_date 2026
15
+ adoption_date 2026
16
+ description { Third edition (cancels and replaces ISO 233-3:2023 as a minor
17
+ revision; the corpus's legacy map implements the 1999 first edition).
18
+ Encodes the strict transliteration of clause 4, the one-to-one
19
+ fully reversible system: Table 1 consonants (32 rows), Table 2
20
+ vowels (alef madda and the three written diacritical vowel signs),
21
+ Table 3 conventional signs (tashdid and the three tanvin marks),
22
+ 4.4.3 ta marbuta per ISO 233, 4.4.4 the Allah word, Table 4 hamze
23
+ forms, Table 5 the ezafe he-plus-hamze form, Table 1 note d's
24
+ strict provision for the yeh-plus-superscript-alef clarity variant,
25
+ Table 6 punctuation and Table 7 numerals. The clause 5 modified
26
+ transliteration (Tables 8-11) is not reversible - it depends on
27
+ word position and on whether a letter functions as consonant or
28
+ vowel - and is not encoded.
29
+ }
30
+
31
+ notes {
32
+ note "Encoded from the purchased published edition (ISO Store order OP-1130631, downloaded 2026-11-19); the PDF text layer is clean and the hex column is authoritative over the printed glyphs. Extracted 2026-11-19."
33
+ note "NFC quirk preserved exactly as the published standard prints it: the zal row (ذ, hex 1E95) and the zain row (ظ, hex 1E93) use precomposed codepoints, while the thal row (ث, hex 0073+0331) and the dad row (ض, hex 007A+0324) are printed decomposed; no precomposed forms exist for those two. Note that the decomposed zal (U+007A+U+0331) is canonically equivalent to the precomposed U+1E95, so NFC input affects which of the two Arabic letters a normalized application resolves; 4.1.3 requires systems to treat precomposed and decomposed characters equally on input."
34
+ note "Clause 4.1.3: a system transliterating from Latin into Persian script shall be case insensitive, and precomposed/decomposed input must be treated equally; both are input-normalization concerns left to the implementing engine, not rows of the map."
35
+ note "Tolerance notes of the source, flagged and not encoded as rows: Table 1 note b (U+2018 may be encountered for the ayn), note c (Arabic yeh U+064A may be encountered in existing text; no strict value is given for it, so bare U+064A passes through unromanized), Table 1 note d and Table 5 note b (U+06C0 may be encountered; it is not canonically equivalent to the encoded he-plus-hamze sequence U+0647+U+0654 and is not permitted in ISIRI 6219), Table 4 note a (U+2019 may be encountered for the hamze apostrophe). Bare alef maqsur U+0649 likewise has no explicit strict row: the source says Persian usually writes it as U+06CC or, for clarity, U+06CC+U+0670 (the latter is the encoded ý row)."
36
+ note "Strict alef rule (Table 1 note a and 4.2.1): alef is always transliterated as a-macron, even as the bearer of a written short vowel (alef + pish = a-macron + o, e.g. اُمید = āomyd); this falls out of the character-by-character rows."
37
+ note "The Allah word (4.4.4, hex 0627 0644 0644 0651 0670 0647) is encoded as one strict multi-character row producing Allāh (capital A as printed); the superscript alef U+0670 occurs in no other row. Reverse-direction limitation, flagged: the transliteration engine inverts rows with a single-character source or target, and this row is multi-character on both sides, so a reversed run does not reconstruct the six-character source from Allāh (the capital A, the shadda and the superscript alef are the unrecoverable parts). The forward direction - the bibliographic direction of the standard - is fully faithful."
38
+ note "Ezafe in the strict system (4.6): the written zir-e ezafe is treated as zir (e) by the character rows; he plus hamze above (hex 0647+0654) is the encoded h-apostrophe row (Table 5) as a pair row; after final alef or vav the ezafe ye is plain yeh (y) per 4.6.2. The clause 5 modified system's -e/-ye hyphenation and the tashdid double-letter U+035F convention are part of the non-reversible system and are not encoded."
39
+ note "Forward collision inherited from the source, flagged: he followed by the standalone hamza letter (hex 0647, 0621, Table 4 row 1 + Table 1 row 31) and he with hamze above (hex 0647+0654, Table 5) both romanize h-apostrophe, and the document provides no disambiguator. The reverse direction resolves h-apostrophe to the standalone form (he plus hamza) because the ezafe pair row does not invert; the choice is recorded here rather than hacked around."
40
+ note "Clause 4.4.1: an Arabic element present in Persian text but not covered by this document follows ISO 233-1; no ISO 233-1 rows are added here."
41
+ note "Provenance lineage: the legacy corpus map iso-prs-Arab-Latn-233-3-1999.isc implements the 1999 first edition (a simplified system); neither the 1999 nor the 2023 second edition document is held, and the legacy map is untouched. This map is the 2026 third edition strict system."
42
+ }
43
+ }
44
+
45
+ tests {
46
+ "آذَر" -> "âẕar"
47
+ "آذر" -> "âẕr"
48
+ "سَم" -> "sam"
49
+ "سم" -> "sm"
50
+ "پُر" -> "por"
51
+ "پر" -> "pr"
52
+ "سِپَر" -> "separ"
53
+ "سپر" -> "spr"
54
+ "اُمید" -> "āomyd"
55
+ "امید" -> "āmyd"
56
+ "مثلاً" -> "ms̱lāã"
57
+ "مربّع" -> "mrbʺʻ"
58
+ "جزء" -> "jzʼ"
59
+ "رأس" -> "rāʼs"
60
+ "سؤال" -> "svʼāl"
61
+ "پائین" -> "pāyʼyn"
62
+ "اللّٰه" -> "Allāh"
63
+ "هٔ" -> "hʼ"
64
+ "یٰ" -> "ý"
65
+ "ة" -> "ẗ"
66
+ "،" -> ","
67
+ "؛" -> ";"
68
+ "؟" -> "?"
69
+ "۱۲۳" -> "123"
70
+ # Exhaustive per-row tests for iso-prs-Arab-Latn-233-3-2026, generated from the source
71
+ # document's table.
72
+ # Source: ISO 233-3:2026(en), third edition, purchased published edition
73
+ # (ISO Store order OP-1130631, downloaded 2026-11-19; verification-grade
74
+ # provenance). STRICT transliteration only (clause 4, the fully reversible
75
+ # system); the clause 5 modified transliteration is not reversible and is
76
+ # not encoded. Rows in document order: Table 1 rows 1-32 (consonants; NFC
77
+ # quirk preserved as printed: the zal row uses precomposed U+1E95 and the
78
+ # zain row precomposed U+1E93, while the sin-e-thalathah row is the
79
+ # decomposed U+0073+U+0331 and the dad row the decomposed U+007A+U+0324 - no
80
+ # precomposed forms exist for these two), Table 2 rows 1-4 (vowels:
81
+ # alef-madda as a-circumflex, zebar a, pish o, zir e), Table 3 (tashdid
82
+ # double-prime U+02BA and the three tanvin rows), 4.4.3 (ta marbuta to
83
+ # t-with-line-below U+1E97 per ISO 233), 4.4.4 (the Allah word as one strict
84
+ # row over hex 0627,0644,0644,0651,0670,0647), Table 4 rows 1-4 (hamze
85
+ # forms; apostrophe U+02BC), Table 5 row 1 (ezafe he plus hamze above h with
86
+ # following U+02BC), Table 1 note d strict provision (yeh plus superscript
87
+ # alef clarity variant to y-acute U+00FD), Table 6 (three punctuation
88
+ # marks), Table 7 (ten numerals).
89
+ # Provenance artifact: scripts/row-sources/iso-prs-Arab-Latn-233-3-2026.json (61 rows).
90
+ # generated-by: maps/scripts/gen-row-tests.mjs (2026-11-19); regenerate from the artifact, do not hand-edit.
91
+ "ا" -> "ā"
92
+ "ب" -> "b"
93
+ "پ" -> "p"
94
+ "ت" -> "t"
95
+ "ث" -> "s̱"
96
+ "ج" -> "j"
97
+ "چ" -> "č"
98
+ "ح" -> "ḥ"
99
+ "خ" -> "x"
100
+ "د" -> "d"
101
+ "ذ" -> "ẕ"
102
+ "ر" -> "r"
103
+ "ز" -> "z"
104
+ "ژ" -> "ž"
105
+ "س" -> "s"
106
+ "ش" -> "š"
107
+ "ص" -> "ṣ"
108
+ "ض" -> "z̤"
109
+ "ط" -> "ṭ"
110
+ "ظ" -> "ẓ"
111
+ "ع" -> "ʻ"
112
+ "غ" -> "ġ"
113
+ "ف" -> "f"
114
+ "ق" -> "q"
115
+ "ک" -> "k"
116
+ "گ" -> "g"
117
+ "ل" -> "l"
118
+ "م" -> "m"
119
+ "ن" -> "n"
120
+ "و" -> "v"
121
+ "ه" -> "h"
122
+ "ی" -> "y"
123
+ "آ" -> "â"
124
+ "َ" -> "a"
125
+ "ُ" -> "o"
126
+ "ِ" -> "e"
127
+ "ّ" -> "ʺ"
128
+ "ً" -> "ã"
129
+ "ٍ" -> "ẽ"
130
+ "ٌ" -> "õ"
131
+ "ة" -> "ẗ"
132
+ "اللّٰه" -> "Allāh"
133
+ "ء" -> "ʼ"
134
+ "أ" -> "āʼ"
135
+ "ؤ" -> "vʼ"
136
+ "ئ" -> "yʼ"
137
+ "هٔ" -> "hʼ"
138
+ "یٰ" -> "ý"
139
+ "،" -> ","
140
+ "؛" -> ";"
141
+ "؟" -> "?"
142
+ "۰" -> "0"
143
+ "۱" -> "1"
144
+ "۲" -> "2"
145
+ "۳" -> "3"
146
+ "۴" -> "4"
147
+ "۵" -> "5"
148
+ "۶" -> "6"
149
+ "۷" -> "7"
150
+ "۸" -> "8"
151
+ "۹" -> "9"
152
+ }
153
+
154
+ stage main {
155
+ # 4.4.4 the Allah word: one strict row over hex 0627 0644 0644 0651 0670 0647
156
+ sub "اللّٰه" "Allāh"
157
+ # Table 5 row 1: ezafe, he plus hamze above (hex 0647+0654)
158
+ sub "هٔ" "hʼ"
159
+ # Table 1 note d: the clarity variant, yeh plus superscript alef
160
+ # (hex 06CC+0670), is transliterated y-acute (hex 00FD) in the strict system
161
+ sub "یٰ" "ý"
162
+
163
+ # Table 1 consonants (rows 1-32)
164
+ parallel {
165
+ sub "ا" "ā"
166
+ sub "ب" "b"
167
+ sub "پ" "p"
168
+ sub "ت" "t"
169
+ sub "ث" "s̱"
170
+ sub "ج" "j"
171
+ sub "چ" "č"
172
+ sub "ح" "ḥ"
173
+ sub "خ" "x"
174
+ sub "د" "d"
175
+ sub "ذ" "ẕ"
176
+ sub "ر" "r"
177
+ sub "ز" "z"
178
+ sub "ژ" "ž"
179
+ sub "س" "s"
180
+ sub "ش" "š"
181
+ sub "ص" "ṣ"
182
+ sub "ض" "z̤"
183
+ sub "ط" "ṭ"
184
+ sub "ظ" "ẓ"
185
+ sub "ع" "ʻ"
186
+ sub "غ" "ġ"
187
+ sub "ف" "f"
188
+ sub "ق" "q"
189
+ sub "ک" "k"
190
+ sub "گ" "g"
191
+ sub "ل" "l"
192
+ sub "م" "m"
193
+ sub "ن" "n"
194
+ sub "و" "v"
195
+ sub "ه" "h"
196
+ sub "ی" "y"
197
+
198
+ # Table 2 vowels for the fully reversible system (rows 1-4)
199
+ sub "آ" "â"
200
+ sub "َ" "a"
201
+ sub "ُ" "o"
202
+ sub "ِ" "e"
203
+
204
+ # Table 3 conventional signs: tashdid and the three tanvin marks
205
+ sub "ّ" "ʺ"
206
+ sub "ً" "ã"
207
+ sub "ٍ" "ẽ"
208
+ sub "ٌ" "õ"
209
+
210
+ # 4.4.3 ta marbuta per ISO 233 (Arabic)
211
+ sub "ة" "ẗ"
212
+
213
+ # Table 4 hamze forms (rows 1-4)
214
+ sub "ء" "ʼ"
215
+ sub "أ" "āʼ"
216
+ sub "ؤ" "vʼ"
217
+ sub "ئ" "yʼ"
218
+
219
+ # Table 6 punctuation marks (rows 1-3)
220
+ sub "،" ","
221
+ sub "؛" ";"
222
+ sub "؟" "?"
223
+
224
+ # Table 7 Persian numerals (rows 1-10)
225
+ sub "۰" "0"
226
+ sub "۱" "1"
227
+ sub "۲" "2"
228
+ sub "۳" "3"
229
+ sub "۴" "4"
230
+ sub "۵" "5"
231
+ sub "۶" "6"
232
+ sub "۷" "7"
233
+ sub "۸" "8"
234
+ sub "۹" "9"
235
+ }
236
+
237
+ }
238
+
239
+ }