@kensio/pinyinjs 1.7.5 → 1.7.7
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.
- package/README.md +128 -123
- package/data/full.entries +2637 -2637
- package/data/manifest.json +9 -9
- package/data/script.map +168 -3
- package/data/standard.entries +8 -8
- package/dist/decode/chang-complement.d.ts +19 -0
- package/dist/decode/chang-complement.d.ts.map +1 -0
- package/dist/decode/chang-complement.js +113 -0
- package/dist/decode/chang-complement.js.map +1 -0
- package/dist/decode/chang-context.d.ts +29 -0
- package/dist/decode/chang-context.d.ts.map +1 -0
- package/dist/decode/chang-context.js +99 -0
- package/dist/decode/chang-context.js.map +1 -0
- package/dist/decode/chang-rule.d.ts +47 -0
- package/dist/decode/chang-rule.d.ts.map +1 -0
- package/dist/decode/chang-rule.js +188 -0
- package/dist/decode/chang-rule.js.map +1 -0
- package/dist/decode/confidence.d.ts +11 -3
- package/dist/decode/confidence.d.ts.map +1 -1
- package/dist/decode/confidence.js +11 -3
- package/dist/decode/confidence.js.map +1 -1
- package/dist/decode/convert-pipeline.d.ts.map +1 -1
- package/dist/decode/convert-pipeline.js +5 -4
- package/dist/decode/convert-pipeline.js.map +1 -1
- package/dist/decode/de-rule.d.ts +74 -0
- package/dist/decode/de-rule.d.ts.map +1 -0
- package/dist/decode/de-rule.js +199 -0
- package/dist/decode/de-rule.js.map +1 -0
- package/dist/decode/decode.d.ts +12 -25
- package/dist/decode/decode.d.ts.map +1 -1
- package/dist/decode/decode.js +10 -47
- package/dist/decode/decode.js.map +1 -1
- package/dist/decode/decoders.d.ts +9 -4
- package/dist/decode/decoders.d.ts.map +1 -1
- package/dist/decode/decoders.js +7 -7
- package/dist/decode/decoders.js.map +1 -1
- package/dist/decode/edge-context.d.ts +19 -0
- package/dist/decode/edge-context.d.ts.map +1 -1
- package/dist/decode/edge-context.js +41 -0
- package/dist/decode/edge-context.js.map +1 -1
- package/dist/decode/hint-lattice.d.ts.map +1 -1
- package/dist/decode/hint-lattice.js +17 -3
- package/dist/decode/hint-lattice.js.map +1 -1
- package/dist/decode/jiao-rule.d.ts +9 -3
- package/dist/decode/jiao-rule.d.ts.map +1 -1
- package/dist/decode/jiao-rule.js +44 -5
- package/dist/decode/jiao-rule.js.map +1 -1
- package/dist/decode/lattice-decode.d.ts +28 -0
- package/dist/decode/lattice-decode.d.ts.map +1 -0
- package/dist/decode/lattice-decode.js +48 -0
- package/dist/decode/lattice-decode.js.map +1 -0
- package/dist/decode/lattice-slice.d.ts +22 -6
- package/dist/decode/lattice-slice.d.ts.map +1 -1
- package/dist/decode/lattice-slice.js +24 -18
- package/dist/decode/lattice-slice.js.map +1 -1
- package/dist/decode/lattice-types.d.ts +27 -8
- package/dist/decode/lattice-types.d.ts.map +1 -1
- package/dist/decode/lattice-types.js +27 -8
- package/dist/decode/lattice-types.js.map +1 -1
- package/dist/decode/numbers.d.ts +0 -8
- package/dist/decode/numbers.d.ts.map +1 -1
- package/dist/decode/numbers.js +0 -10
- package/dist/decode/numbers.js.map +1 -1
- package/dist/decode/polyphone-rules.d.ts +7 -27
- package/dist/decode/polyphone-rules.d.ts.map +1 -1
- package/dist/decode/polyphone-rules.js +7 -155
- package/dist/decode/polyphone-rules.js.map +1 -1
- package/dist/decode/reading-rules.d.ts +1 -29
- package/dist/decode/reading-rules.d.ts.map +1 -1
- package/dist/decode/reading-rules.js +5 -69
- package/dist/decode/reading-rules.js.map +1 -1
- package/dist/decode/rules.d.ts +1 -1
- package/dist/decode/rules.d.ts.map +1 -1
- package/dist/decode/rules.js +1 -1
- package/dist/decode/rules.js.map +1 -1
- package/dist/decode/run-context.d.ts +33 -0
- package/dist/decode/run-context.d.ts.map +1 -0
- package/dist/decode/run-context.js +47 -0
- package/dist/decode/run-context.js.map +1 -0
- package/dist/decode/run-lattice.d.ts +13 -10
- package/dist/decode/run-lattice.d.ts.map +1 -1
- package/dist/decode/run-lattice.js +15 -12
- package/dist/decode/run-lattice.js.map +1 -1
- package/dist/decode/run-words.d.ts +17 -0
- package/dist/decode/run-words.d.ts.map +1 -0
- package/dist/decode/run-words.js +29 -0
- package/dist/decode/run-words.js.map +1 -0
- package/dist/decode/sandhi-input.d.ts +30 -0
- package/dist/decode/sandhi-input.d.ts.map +1 -0
- package/dist/decode/sandhi-input.js +35 -0
- package/dist/decode/sandhi-input.js.map +1 -0
- package/dist/decode/sandhi-tones.d.ts +33 -5
- package/dist/decode/sandhi-tones.d.ts.map +1 -1
- package/dist/decode/sandhi-tones.js +59 -8
- package/dist/decode/sandhi-tones.js.map +1 -1
- package/dist/decode/sandhi.d.ts +9 -6
- package/dist/decode/sandhi.d.ts.map +1 -1
- package/dist/decode/sandhi.js +10 -7
- package/dist/decode/sandhi.js.map +1 -1
- package/dist/decode/script-region.d.ts +15 -0
- package/dist/decode/script-region.d.ts.map +1 -0
- package/dist/decode/script-region.js +61 -0
- package/dist/decode/script-region.js.map +1 -0
- package/dist/decode/script-types.d.ts +11 -2
- package/dist/decode/script-types.d.ts.map +1 -1
- package/dist/decode/script-types.js +2 -1
- package/dist/decode/script-types.js.map +1 -1
- package/dist/decode/script-words.d.ts +7 -5
- package/dist/decode/script-words.d.ts.map +1 -1
- package/dist/decode/script-words.js +18 -25
- package/dist/decode/script-words.js.map +1 -1
- package/dist/decode/script.d.ts.map +1 -1
- package/dist/decode/script.js +2 -1
- package/dist/decode/script.js.map +1 -1
- package/dist/decode/write-run.d.ts +0 -7
- package/dist/decode/write-run.d.ts.map +1 -1
- package/dist/decode/write-run.js +4 -39
- package/dist/decode/write-run.js.map +1 -1
- package/dist/dictionary/cedict-senses.d.ts +7 -0
- package/dist/dictionary/cedict-senses.d.ts.map +1 -1
- package/dist/dictionary/cedict-senses.js +12 -0
- package/dist/dictionary/cedict-senses.js.map +1 -1
- package/dist/dictionary/constituent-repair.d.ts +36 -0
- package/dist/dictionary/constituent-repair.d.ts.map +1 -0
- package/dist/dictionary/constituent-repair.js +117 -0
- package/dist/dictionary/constituent-repair.js.map +1 -0
- package/dist/dictionary/measure-tone.d.ts +36 -0
- package/dist/dictionary/measure-tone.d.ts.map +1 -0
- package/dist/dictionary/measure-tone.js +139 -0
- package/dist/dictionary/measure-tone.js.map +1 -0
- package/dist/dictionary/merge-types.d.ts +13 -2
- package/dist/dictionary/merge-types.d.ts.map +1 -1
- package/dist/dictionary/merge-types.js.map +1 -1
- package/dist/dictionary/merge-walk.d.ts +27 -0
- package/dist/dictionary/merge-walk.d.ts.map +1 -0
- package/dist/dictionary/merge-walk.js +49 -0
- package/dist/dictionary/merge-walk.js.map +1 -0
- package/dist/dictionary/merge.d.ts +3 -3
- package/dist/dictionary/merge.d.ts.map +1 -1
- package/dist/dictionary/merge.js +24 -28
- package/dist/dictionary/merge.js.map +1 -1
- package/dist/dictionary/overrides.d.ts +5 -0
- package/dist/dictionary/overrides.d.ts.map +1 -1
- package/dist/dictionary/overrides.js +5 -5
- package/dist/dictionary/overrides.js.map +1 -1
- package/dist/dictionary/script-tables.js +91 -13
- package/dist/dictionary/script-tables.js.map +1 -1
- package/dist/dictionary/word-readings.d.ts.map +1 -1
- package/dist/dictionary/word-readings.js +6 -0
- package/dist/dictionary/word-readings.js.map +1 -1
- package/dist/index.d.ts +1 -1
- package/dist/index.d.ts.map +1 -1
- package/dist/index.js +1 -1
- package/dist/index.js.map +1 -1
- package/dist/script/glyphs.d.ts +9 -8
- package/dist/script/glyphs.d.ts.map +1 -1
- package/dist/script/glyphs.js +21 -48
- package/dist/script/glyphs.js.map +1 -1
- package/dist/script/reading-splits.d.ts +41 -0
- package/dist/script/reading-splits.d.ts.map +1 -0
- package/dist/script/reading-splits.js +98 -0
- package/dist/script/reading-splits.js.map +1 -0
- package/package.json +2 -2
|
@@ -1,31 +1,11 @@
|
|
|
1
|
-
export { PLAYING_TAN } from "./tan-rule.js";
|
|
2
|
-
import { type EdgeRule } from "./rules.js";
|
|
3
1
|
/**
|
|
4
|
-
*
|
|
5
|
-
*
|
|
6
|
-
* The character is stored `zhǎng` with `cháng` as an alternate, and the data
|
|
7
|
-
* says so about the character in isolation: Unihan ranks the readings
|
|
8
|
-
* `zhǎng(1879) cháng(1179)` by corpus count and names `zhǎng` in `kMandarin`.
|
|
9
|
-
* That ranking is doing real work — 署长, 团长, 公安局长, 总会长, 审理长 and
|
|
10
|
-
* the rest reach a bare 长 at the end of a title and read it correctly — so the
|
|
11
|
-
* default is not the thing to change. What is left over is the 长 that is an
|
|
12
|
-
* adjective, and nothing in the cost model can prefer an alternate at a
|
|
13
|
-
* position no word covers: 这篇文章不太长 came out `bú tài zhǎng`.
|
|
14
|
-
*
|
|
15
|
-
* The context that separates them is one-sided, as it was for the modal 得. A
|
|
16
|
-
* 长 that is *growing* is a verb, and a degree adverb cannot modify one — there
|
|
17
|
-
* is no 很长 meaning it grew a lot — so 很, 太, 最, 多 and their kin in front of
|
|
18
|
-
* a bare 长 make it the adjective. The other side carries almost nothing: what
|
|
19
|
-
* follows an adjectival 长 is a noun (很长时间), a particle (最长的大桥), or the
|
|
20
|
-
* end of the sentence (他的腿很长), which is also what follows half the verbs.
|
|
2
|
+
* The rules that settle a polyphone from the words around it.
|
|
21
3
|
*
|
|
22
|
-
*
|
|
23
|
-
*
|
|
24
|
-
*
|
|
25
|
-
*
|
|
26
|
-
*
|
|
27
|
-
* The other half of the rule is {@link GROWN_INTO}, which pushes the opposite
|
|
28
|
-
* way on the one shape the sources get wrong.
|
|
4
|
+
* 弹 is `tán` or `dàn`, and which one a text means is a question about the
|
|
5
|
+
* neighbours rather than about the character. 长 is the same question and has
|
|
6
|
+
* a file of its own. The other rules in `reading-rules.ts` are about particles
|
|
7
|
+
* and 儿化, which is a different kind of evidence.
|
|
29
8
|
*/
|
|
30
|
-
export
|
|
9
|
+
export { ADJECTIVAL_CHANG } from "./chang-rule.js";
|
|
10
|
+
export { PLAYING_TAN } from "./tan-rule.js";
|
|
31
11
|
//# sourceMappingURL=polyphone-rules.d.ts.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"polyphone-rules.d.ts","sourceRoot":"","sources":["../../src/decode/polyphone-rules.ts"],"names":[],"mappings":"
|
|
1
|
+
{"version":3,"file":"polyphone-rules.d.ts","sourceRoot":"","sources":["../../src/decode/polyphone-rules.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AACH,OAAO,EAAE,gBAAgB,EAAE,MAAM,iBAAiB,CAAC;AACnD,OAAO,EAAE,WAAW,EAAE,MAAM,eAAe,CAAC"}
|
|
@@ -1,159 +1,11 @@
|
|
|
1
|
-
export { PLAYING_TAN } from "./tan-rule.js";
|
|
2
|
-
import { wordEndingAt } from "./rules.js";
|
|
3
|
-
/**
|
|
4
|
-
* Whether a syllable is `cháng`, the stative reading of 长.
|
|
5
|
-
*/
|
|
6
|
-
function isChang(syllable) {
|
|
7
|
-
return (syllable?.initial === "ch" &&
|
|
8
|
-
syllable.final === "ang" &&
|
|
9
|
-
syllable.tone === 2);
|
|
10
|
-
}
|
|
11
|
-
/**
|
|
12
|
-
* Adverbs that measure a quality rather than modify an action.
|
|
13
|
-
*
|
|
14
|
-
* The discriminating side of the 长 context, and the reason the set is written
|
|
15
|
-
* out rather than taken from a tag: jieba calls 很 `zg` and 多 `m`, so the two
|
|
16
|
-
* commonest members are not adverbs as far as the dictionary is concerned,
|
|
17
|
-
* while 已经, 也, 却 and 就 — which sit in front of the *verb* 长 — are `d`
|
|
18
|
-
* along with the ones wanted here. The tag names the wrong set in both
|
|
19
|
-
* directions, so the set is named directly.
|
|
20
|
-
*
|
|
21
|
-
* Everything here is degree or comparison. 还 and 不 are deliberately absent
|
|
22
|
-
* even though 距离还长 and 那座桥不长 are both `cháng`: they modify verbs as
|
|
23
|
-
* readily as qualities, and 还长有毛 and 胡子不长在前额上 are the same two
|
|
24
|
-
* characters around the verb.
|
|
25
|
-
*/
|
|
26
|
-
const DEGREE = new Set([
|
|
27
|
-
"一样",
|
|
28
|
-
"十分",
|
|
29
|
-
"多",
|
|
30
|
-
"多么",
|
|
31
|
-
"好",
|
|
32
|
-
"如此",
|
|
33
|
-
"尽量",
|
|
34
|
-
"很",
|
|
35
|
-
"格外",
|
|
36
|
-
"极",
|
|
37
|
-
"极其",
|
|
38
|
-
"比较",
|
|
39
|
-
"略",
|
|
40
|
-
"相当",
|
|
41
|
-
"真",
|
|
42
|
-
"稍",
|
|
43
|
-
"稍微",
|
|
44
|
-
"越",
|
|
45
|
-
"越来越",
|
|
46
|
-
"这么",
|
|
47
|
-
"这样",
|
|
48
|
-
"那么",
|
|
49
|
-
"那样",
|
|
50
|
-
"更",
|
|
51
|
-
"最",
|
|
52
|
-
"有多",
|
|
53
|
-
"特别",
|
|
54
|
-
"老",
|
|
55
|
-
"够",
|
|
56
|
-
"挺",
|
|
57
|
-
"太",
|
|
58
|
-
"非常",
|
|
59
|
-
]);
|
|
60
|
-
/**
|
|
61
|
-
* What shows a 长 to be growing even with a degree word in front of it.
|
|
62
|
-
*
|
|
63
|
-
* {@link DEGREE} already excludes nearly everything that precedes the verb, so
|
|
64
|
-
* this only has to settle the three shapes reachable from both sets: 真长得很快
|
|
65
|
-
* and 越长得快, where 得 marks the complement; 长着, where the aspect does; and
|
|
66
|
-
* 越长越高, where the correlative makes the 长 the thing that increases.
|
|
67
|
-
*
|
|
68
|
-
* 了 and 的 are pointedly *not* here, and that is the whole use of conditioning
|
|
69
|
-
* on the adverb first. Following a bare 长 either would suggest the verb — 长了
|
|
70
|
-
* 一个大包, 长的很像 — but following an adverb and a 长 they are the sentence
|
|
71
|
-
* particle and the attributive: 时间太长了 and 很长的道路 are both `cháng`, and
|
|
72
|
-
* an earlier draft of this rule that guarded on them read both as `zhǎng`.
|
|
73
|
-
*/
|
|
74
|
-
const GROWING = new Set(["得", "着", "越"]);
|
|
75
1
|
/**
|
|
76
|
-
*
|
|
2
|
+
* The rules that settle a polyphone from the words around it.
|
|
77
3
|
*
|
|
78
|
-
*
|
|
79
|
-
*
|
|
80
|
-
*
|
|
81
|
-
*
|
|
4
|
+
* 弹 is `tán` or `dàn`, and which one a text means is a question about the
|
|
5
|
+
* neighbours rather than about the character. 长 is the same question and has
|
|
6
|
+
* a file of its own. The other rules in `reading-rules.ts` are about particles
|
|
7
|
+
* and 儿化, which is a different kind of evidence.
|
|
82
8
|
*/
|
|
83
|
-
|
|
84
|
-
|
|
85
|
-
* Whether an edge is the 越长 of a 越长越X saying something grew.
|
|
86
|
-
*
|
|
87
|
-
* This exists because of a defect in the sources rather than a gap in them.
|
|
88
|
-
* 越长 is a key, read `yuè cháng`, and it is the *only* one of its shape: 越大,
|
|
89
|
-
* 越高, 越好, 越快 and 越多 are all absent, so 越高越好 decodes as two words and
|
|
90
|
-
* 越长越高 reaches for a word nothing else in the paradigm has. It carries no
|
|
91
|
-
* part of speech, which is how the dictionary holds a reading somebody asserted
|
|
92
|
-
* rather than a word anybody counted — the same category {@link TEACHING_JIAO}
|
|
93
|
-
* declines to trust — and it comes from one source, `large_pinyin.txt`, with
|
|
94
|
-
* CC-CEDICT holding no such entry.
|
|
95
|
-
*
|
|
96
|
-
* Taking the edge away leaves the character's own default, which is `zhǎng`,
|
|
97
|
-
* and leaves `cháng` standing as a rival about one bucket dearer. That last
|
|
98
|
-
* part is the point of forbidding rather than forcing: 越长越X is genuinely
|
|
99
|
-
* ambiguous — 孩子越长越漂亮 grows and 头发越长越漂亮 lengthens, and only the
|
|
100
|
-
* subject says which — so the decode should answer with the likelier reading
|
|
101
|
-
* and still report itself as guessing. A forced edge could not.
|
|
102
|
-
*
|
|
103
|
-
* Unlike the rest of this file the shape has no corpus behind it: 越长 occurs
|
|
104
|
-
* three times in the 88,866 lines, and all three are 越来越长 or 说的越长. The
|
|
105
|
-
* word list is therefore small on purpose, and everything outside it is left as
|
|
106
|
-
* the sources have it.
|
|
107
|
-
*/
|
|
108
|
-
function isGrowthCorrelative(context) {
|
|
109
|
-
const { characters, edge } = context;
|
|
110
|
-
return (edge.text === "越长" &&
|
|
111
|
-
edge.partOfSpeech === "" &&
|
|
112
|
-
characters[edge.to] === "越" &&
|
|
113
|
-
GROWN_INTO.has(characters[edge.to + 1] ?? ""));
|
|
114
|
-
}
|
|
115
|
-
/**
|
|
116
|
-
* 长 read as `cháng` where an adverb of degree measures it.
|
|
117
|
-
*
|
|
118
|
-
* The character is stored `zhǎng` with `cháng` as an alternate, and the data
|
|
119
|
-
* says so about the character in isolation: Unihan ranks the readings
|
|
120
|
-
* `zhǎng(1879) cháng(1179)` by corpus count and names `zhǎng` in `kMandarin`.
|
|
121
|
-
* That ranking is doing real work — 署长, 团长, 公安局长, 总会长, 审理长 and
|
|
122
|
-
* the rest reach a bare 长 at the end of a title and read it correctly — so the
|
|
123
|
-
* default is not the thing to change. What is left over is the 长 that is an
|
|
124
|
-
* adjective, and nothing in the cost model can prefer an alternate at a
|
|
125
|
-
* position no word covers: 这篇文章不太长 came out `bú tài zhǎng`.
|
|
126
|
-
*
|
|
127
|
-
* The context that separates them is one-sided, as it was for the modal 得. A
|
|
128
|
-
* 长 that is *growing* is a verb, and a degree adverb cannot modify one — there
|
|
129
|
-
* is no 很长 meaning it grew a lot — so 很, 太, 最, 多 and their kin in front of
|
|
130
|
-
* a bare 长 make it the adjective. The other side carries almost nothing: what
|
|
131
|
-
* follows an adjectival 长 is a noun (很长时间), a particle (最长的大桥), or the
|
|
132
|
-
* end of the sentence (他的腿很长), which is also what follows half the verbs.
|
|
133
|
-
*
|
|
134
|
-
* Measured over the same 88,866 lines of Tatoeba and zh.wikipedia the other
|
|
135
|
-
* rules were sized against, 260 长 decode as a word of their own and this moves
|
|
136
|
-
* 75 of them to `cháng`, all 75 correctly. On CPP's 40 hand-labelled 长 it takes
|
|
137
|
-
* the character from 85.00% to 87.50%, fixing 时间最长 and breaking nothing.
|
|
138
|
-
*
|
|
139
|
-
* The other half of the rule is {@link GROWN_INTO}, which pushes the opposite
|
|
140
|
-
* way on the one shape the sources get wrong.
|
|
141
|
-
*/
|
|
142
|
-
export const ADJECTIVAL_CHANG = {
|
|
143
|
-
name: "adjectival-chang",
|
|
144
|
-
verdictFor: (context) => {
|
|
145
|
-
const { characters, edge } = context;
|
|
146
|
-
if (isGrowthCorrelative(context)) {
|
|
147
|
-
return "forbid";
|
|
148
|
-
}
|
|
149
|
-
if (edge.text !== "长" || !isChang(edge.reading[0])) {
|
|
150
|
-
return "keep";
|
|
151
|
-
}
|
|
152
|
-
if (GROWING.has(characters[edge.to] ?? "")) {
|
|
153
|
-
return "keep";
|
|
154
|
-
}
|
|
155
|
-
const before = wordEndingAt(context, edge.from);
|
|
156
|
-
return before !== undefined && DEGREE.has(before) ? "force" : "keep";
|
|
157
|
-
},
|
|
158
|
-
};
|
|
9
|
+
export { ADJECTIVAL_CHANG } from "./chang-rule.js";
|
|
10
|
+
export { PLAYING_TAN } from "./tan-rule.js";
|
|
159
11
|
//# sourceMappingURL=polyphone-rules.js.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"polyphone-rules.js","sourceRoot":"","sources":["../../src/decode/polyphone-rules.ts"],"names":[],"mappings":"
|
|
1
|
+
{"version":3,"file":"polyphone-rules.js","sourceRoot":"","sources":["../../src/decode/polyphone-rules.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AACH,OAAO,EAAE,gBAAgB,EAAE,MAAM,iBAAiB,CAAC;AACnD,OAAO,EAAE,WAAW,EAAE,MAAM,eAAe,CAAC"}
|
|
@@ -1,34 +1,6 @@
|
|
|
1
|
+
export { MODAL_DE, PARTICLE_DE, POTENTIAL_DE } from "./de-rule.js";
|
|
1
2
|
export { TEACHING_JIAO } from "./jiao-rule.js";
|
|
2
3
|
import { type EdgeRule } from "./rules.js";
|
|
3
|
-
/**
|
|
4
|
-
* 得 read as `děi` where it is a modal rather than a particle.
|
|
5
|
-
*
|
|
6
|
-
* One character, three readings, and the dictionary can only carry a default:
|
|
7
|
-
* 得 is stored `de` — the structural particle, which is far the commonest — with
|
|
8
|
-
* `dé` and `děi` as alternates. Nothing in the cost model can prefer an
|
|
9
|
-
* alternate at a position no word covers, so 我得走了 came out `wǒ de zǒule`
|
|
10
|
-
* and every other modal 得 with it. This is the case ROADMAP.md was waiting for
|
|
11
|
-
* before building stage 4: neither harness could see it, since the gold corpus
|
|
12
|
-
* had only 觉得 and CPP's 40 得 cases are labelled `dé` or `de` and never `děi`.
|
|
13
|
-
*
|
|
14
|
-
* The context is what separates the three, and only one side of it carries
|
|
15
|
-
* information. The particle 得 always follows the verb or adjective it attaches
|
|
16
|
-
* to (说得好, 跑得快, 累得要命), so a 得 with a *pronoun, adverb or time word*
|
|
17
|
-
* in front of it is not that particle. Requiring a verb phrase after it as well
|
|
18
|
-
* keeps 得了感冒 and 得了奖 — `dé`, and a real reading of the same character —
|
|
19
|
-
* out of it.
|
|
20
|
-
*
|
|
21
|
-
* Conditioning on what precedes rather than on what follows was measured, not
|
|
22
|
-
* assumed: "the word before is not a verb" fires 309 times over 88,866 lines of
|
|
23
|
-
* Tatoeba and zh.wikipedia and is wrong about four times in five, because
|
|
24
|
-
* jieba tags 下 as `f`, 学 as `n` and 飞 as nothing at all, and 雨下得很大 is a
|
|
25
|
-
* particle whatever the tag says. Naming the small closed set of things that
|
|
26
|
-
* can precede a modal is the same rule from the other side, and over the same
|
|
27
|
-
* text it fires 126 times and is right 122: the four misses are literary 得
|
|
28
|
-
* read `dé` after an adverb — 未必得喜悦, 暂得于己, 仅得约, 未得迁还家乡 —
|
|
29
|
-
* which nothing short of semantics separates from 必须得承认.
|
|
30
|
-
*/
|
|
31
|
-
export declare const MODAL_DE: EdgeRule;
|
|
32
4
|
/**
|
|
33
5
|
* 儿 does not stand on its own where the dictionary attests the 儿化.
|
|
34
6
|
*
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"reading-rules.d.ts","sourceRoot":"","sources":["../../src/decode/reading-rules.ts"],"names":[],"mappings":"
|
|
1
|
+
{"version":3,"file":"reading-rules.d.ts","sourceRoot":"","sources":["../../src/decode/reading-rules.ts"],"names":[],"mappings":"AAIA,OAAO,EAAE,QAAQ,EAAE,WAAW,EAAE,YAAY,EAAE,MAAM,cAAc,CAAC;AACnE,OAAO,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAE/C,OAAO,EAEL,KAAK,QAAQ,EAGd,MAAM,YAAY,CAAC;AAEpB;;;;;;;;;;;;;;;;;;;GAmBG;AACH,eAAO,MAAM,cAAc,EAAE,QAY5B,CAAC;AAsCF;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AACH,eAAO,MAAM,eAAe,EAAE,QAuB7B,CAAC;AAIF,OAAO,EAAE,gBAAgB,EAAE,WAAW,EAAE,MAAM,sBAAsB,CAAC;AAErE;;GAEG;AACH,eAAO,MAAM,aAAa,EAAE,SAAS,QAAQ,EAS5C,CAAC"}
|
|
@@ -1,76 +1,10 @@
|
|
|
1
1
|
import { QUANTITY_CHARACTERS } from "../numerals/characters.js";
|
|
2
|
+
import { MODAL_DE, PARTICLE_DE, POTENTIAL_DE } from "./de-rule.js";
|
|
2
3
|
import { TEACHING_JIAO } from "./jiao-rule.js";
|
|
4
|
+
export { MODAL_DE, PARTICLE_DE, POTENTIAL_DE } from "./de-rule.js";
|
|
3
5
|
export { TEACHING_JIAO } from "./jiao-rule.js";
|
|
4
6
|
import { toCharacters } from "../script/characters.js";
|
|
5
|
-
import { tagOf, wordEndingAt,
|
|
6
|
-
/**
|
|
7
|
-
* Whether a syllable is `děi`, the modal reading of 得.
|
|
8
|
-
*
|
|
9
|
-
* Compared structurally rather than by spelling it out, since a syllable is
|
|
10
|
-
* stored as its parts and writing one back is the formatting stage's job.
|
|
11
|
-
*/
|
|
12
|
-
function isDei(syllable) {
|
|
13
|
-
return (syllable?.initial === "d" && syllable.final === "ei" && syllable.tone === 3);
|
|
14
|
-
}
|
|
15
|
-
/**
|
|
16
|
-
* What stands in front of a modal 得: a subject, an adverb, or a time word.
|
|
17
|
-
*
|
|
18
|
-
* 我得走, 你就得去, 我今天得走. The particle 得 cannot follow any of these,
|
|
19
|
-
* because it attaches to the predicate it modifies — which is exactly what
|
|
20
|
-
* makes this the discriminating side of the context.
|
|
21
|
-
*/
|
|
22
|
-
const SUBJECT_TAGS = new Set(["r", "d", "t"]);
|
|
23
|
-
/**
|
|
24
|
-
* What follows a modal 得: the verb phrase it governs.
|
|
25
|
-
*
|
|
26
|
-
* Prepositions and adverbs are in because 我得在八月, 我得先走 and 我得快点 are
|
|
27
|
-
* all modal, and what follows the 得 in each is the front of a verb phrase
|
|
28
|
-
* rather than a verb.
|
|
29
|
-
*/
|
|
30
|
-
const GOVERNED_TAGS = /^(?:v|p|d|a)/u;
|
|
31
|
-
/**
|
|
32
|
-
* 得 read as `děi` where it is a modal rather than a particle.
|
|
33
|
-
*
|
|
34
|
-
* One character, three readings, and the dictionary can only carry a default:
|
|
35
|
-
* 得 is stored `de` — the structural particle, which is far the commonest — with
|
|
36
|
-
* `dé` and `děi` as alternates. Nothing in the cost model can prefer an
|
|
37
|
-
* alternate at a position no word covers, so 我得走了 came out `wǒ de zǒule`
|
|
38
|
-
* and every other modal 得 with it. This is the case ROADMAP.md was waiting for
|
|
39
|
-
* before building stage 4: neither harness could see it, since the gold corpus
|
|
40
|
-
* had only 觉得 and CPP's 40 得 cases are labelled `dé` or `de` and never `děi`.
|
|
41
|
-
*
|
|
42
|
-
* The context is what separates the three, and only one side of it carries
|
|
43
|
-
* information. The particle 得 always follows the verb or adjective it attaches
|
|
44
|
-
* to (说得好, 跑得快, 累得要命), so a 得 with a *pronoun, adverb or time word*
|
|
45
|
-
* in front of it is not that particle. Requiring a verb phrase after it as well
|
|
46
|
-
* keeps 得了感冒 and 得了奖 — `dé`, and a real reading of the same character —
|
|
47
|
-
* out of it.
|
|
48
|
-
*
|
|
49
|
-
* Conditioning on what precedes rather than on what follows was measured, not
|
|
50
|
-
* assumed: "the word before is not a verb" fires 309 times over 88,866 lines of
|
|
51
|
-
* Tatoeba and zh.wikipedia and is wrong about four times in five, because
|
|
52
|
-
* jieba tags 下 as `f`, 学 as `n` and 飞 as nothing at all, and 雨下得很大 is a
|
|
53
|
-
* particle whatever the tag says. Naming the small closed set of things that
|
|
54
|
-
* can precede a modal is the same rule from the other side, and over the same
|
|
55
|
-
* text it fires 126 times and is right 122: the four misses are literary 得
|
|
56
|
-
* read `dé` after an adverb — 未必得喜悦, 暂得于己, 仅得约, 未得迁还家乡 —
|
|
57
|
-
* which nothing short of semantics separates from 必须得承认.
|
|
58
|
-
*/
|
|
59
|
-
export const MODAL_DE = {
|
|
60
|
-
name: "modal-de",
|
|
61
|
-
verdictFor: (context) => {
|
|
62
|
-
const { edge } = context;
|
|
63
|
-
if (edge.text !== "得" || !isDei(edge.reading[0])) {
|
|
64
|
-
return "keep";
|
|
65
|
-
}
|
|
66
|
-
const before = wordEndingAt(context, edge.from);
|
|
67
|
-
const after = wordStartingAt(context, edge.to);
|
|
68
|
-
return SUBJECT_TAGS.has(tagOf(context, before)) &&
|
|
69
|
-
GOVERNED_TAGS.test(tagOf(context, after))
|
|
70
|
-
? "force"
|
|
71
|
-
: "keep";
|
|
72
|
-
},
|
|
73
|
-
};
|
|
7
|
+
import { tagOf, wordEndingAt, } from "./rules.js";
|
|
74
8
|
/**
|
|
75
9
|
* 儿 does not stand on its own where the dictionary attests the 儿化.
|
|
76
10
|
*
|
|
@@ -194,6 +128,8 @@ export { ADJECTIVAL_CHANG, PLAYING_TAN } from "./polyphone-rules.js";
|
|
|
194
128
|
*/
|
|
195
129
|
export const READING_RULES = [
|
|
196
130
|
MODAL_DE,
|
|
131
|
+
PARTICLE_DE,
|
|
132
|
+
POTENTIAL_DE,
|
|
197
133
|
TEACHING_JIAO,
|
|
198
134
|
ATTESTED_ERHUA,
|
|
199
135
|
COUNTED_MEASURE,
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"reading-rules.js","sourceRoot":"","sources":["../../src/decode/reading-rules.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,mBAAmB,EAAE,MAAM,2BAA2B,CAAC;AAChE,OAAO,EAAE,
|
|
1
|
+
{"version":3,"file":"reading-rules.js","sourceRoot":"","sources":["../../src/decode/reading-rules.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,mBAAmB,EAAE,MAAM,2BAA2B,CAAC;AAChE,OAAO,EAAE,QAAQ,EAAE,WAAW,EAAE,YAAY,EAAE,MAAM,cAAc,CAAC;AACnE,OAAO,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAE/C,OAAO,EAAE,QAAQ,EAAE,WAAW,EAAE,YAAY,EAAE,MAAM,cAAc,CAAC;AACnE,OAAO,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAC/C,OAAO,EAAE,YAAY,EAAE,MAAM,yBAAyB,CAAC;AACvD,OAAO,EAGL,KAAK,EACL,YAAY,GACb,MAAM,YAAY,CAAC;AAEpB;;;;;;;;;;;;;;;;;;;GAmBG;AACH,MAAM,CAAC,MAAM,cAAc,GAAa;IACtC,IAAI,EAAE,gBAAgB;IACtB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,UAAU,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACrC,IAAI,IAAI,CAAC,IAAI,KAAK,GAAG,IAAI,IAAI,CAAC,IAAI,KAAK,CAAC,EAAE,CAAC;YACzC,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,wEAAwE;QACxE,MAAM,QAAQ,GAAG,GAAG,UAAU,CAAC,IAAI,CAAC,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,GAAG,CAAC;QACvD,MAAM,OAAO,GAAG,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,QAAQ,CAAC,EAAE,OAAO,CAAC;QAC7D,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,KAAK,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC;IAC/D,CAAC;CACF,CAAC;AAEF;;GAEG;AACH,MAAM,OAAO,GAAG,GAAG,CAAC;AAEpB;;;;;GAKG;AACH,MAAM,OAAO,GAAG,GAAG,CAAC;AAEpB;;;;;;;;GAQG;AACH,SAAS,WAAW,CAAC,OAAoB,EAAE,EAAU;IACnD,MAAM,EAAE,UAAU,EAAE,GAAG,OAAO,CAAC;IAC/B,IAAI,IAAI,GAAG,EAAE,CAAC;IACd,OAAO,IAAI,GAAG,CAAC,IAAI,mBAAmB,CAAC,GAAG,CAAC,UAAU,CAAC,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,EAAE,CAAC;QACvE,IAAI,EAAE,CAAC;IACT,CAAC;IACD,IAAI,IAAI,KAAK,EAAE,IAAI,UAAU,CAAC,IAAI,GAAG,CAAC,CAAC,KAAK,OAAO,EAAE,CAAC;QACpD,OAAO,SAAS,CAAC;IACnB,CAAC;IACD,MAAM,MAAM,GAAG,YAAY,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC;IACzC,MAAM,IAAI,GAAG,MAAM,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,GAAG,YAAY,CAAC,MAAM,CAAC,CAAC,MAAM,CAAC;IAC1E,OAAO,IAAI,GAAG,IAAI,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC;AACxC,CAAC;AAED;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AACH,MAAM,CAAC,MAAM,eAAe,GAAa;IACvC,IAAI,EAAE,iBAAiB;IACvB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,UAAU,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACrC,yEAAyE;QACzE,yEAAyE;QACzE,yEAAyE;QACzE,WAAW;QACX,IACE,IAAI,CAAC,EAAE,GAAG,IAAI,CAAC,IAAI,GAAG,CAAC;YACvB,IAAI,CAAC,OAAO,CAAC,MAAM,KAAK,IAAI,CAAC,EAAE,GAAG,IAAI,CAAC,IAAI,EAC3C,CAAC;YACD,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,MAAM,OAAO,GAAG,WAAW,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,CAAC,CAAC;QAChD,IAAI,OAAO,KAAK,SAAS,EAAE,CAAC;YAC1B,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,OAAO,KAAK,CAAC,OAAO,EAAE,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,KAAK,OAAO;YACtD,IAAI,CAAC,YAAY,KAAK,OAAO;YAC7B,CAAC,CAAC,QAAQ;YACV,CAAC,CAAC,MAAM,CAAC;IACb,CAAC;CACF,CAAC;AAEF,OAAO,EAAE,gBAAgB,EAAE,WAAW,EAAE,MAAM,sBAAsB,CAAC;AAErE,OAAO,EAAE,gBAAgB,EAAE,WAAW,EAAE,MAAM,sBAAsB,CAAC;AAErE;;GAEG;AACH,MAAM,CAAC,MAAM,aAAa,GAAwB;IAChD,QAAQ;IACR,WAAW;IACX,YAAY;IACZ,aAAa;IACb,cAAc;IACd,eAAe;IACf,gBAAgB;IAChB,WAAW;CACZ,CAAC"}
|
package/dist/decode/rules.d.ts
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
import type { Dictionary } from "../dictionary/dictionary.js";
|
|
2
2
|
import type { Lattice, LatticeEdge } from "./lattice.js";
|
|
3
|
-
export { ASPECT, PARTICLE_TAG, tagOf, wordEndingAt, wordStartingAt, } from "./edge-context.js";
|
|
3
|
+
export { ASPECT, PARTICLE_TAG, tagOf, wordEndingAt, wordsEndingAt, wordStartingAt, wordsStartingAt, } from "./edge-context.js";
|
|
4
4
|
/**
|
|
5
5
|
* What a rule can say about one edge.
|
|
6
6
|
*
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"rules.d.ts","sourceRoot":"","sources":["../../src/decode/rules.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,6BAA6B,CAAC;AAC9D,OAAO,KAAK,EAAE,OAAO,EAAE,WAAW,EAAE,MAAM,cAAc,CAAC;AAEzD,OAAO,EACL,MAAM,EACN,YAAY,EACZ,KAAK,EACL,YAAY,EACZ,cAAc,
|
|
1
|
+
{"version":3,"file":"rules.d.ts","sourceRoot":"","sources":["../../src/decode/rules.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,6BAA6B,CAAC;AAC9D,OAAO,KAAK,EAAE,OAAO,EAAE,WAAW,EAAE,MAAM,cAAc,CAAC;AAEzD,OAAO,EACL,MAAM,EACN,YAAY,EACZ,KAAK,EACL,YAAY,EACZ,aAAa,EACb,cAAc,EACd,eAAe,GAChB,MAAM,mBAAmB,CAAC;AAE3B;;;;;;;GAOG;AACH,MAAM,MAAM,WAAW,GAAG,MAAM,GAAG,QAAQ,GAAG,OAAO,CAAC;AAEtD;;;;;;GAMG;AACH,MAAM,WAAW,WAAW;IAC1B,QAAQ,CAAC,UAAU,EAAE,UAAU,CAAC;IAChC,QAAQ,CAAC,UAAU,EAAE,SAAS,MAAM,EAAE,CAAC;IACvC,QAAQ,CAAC,IAAI,EAAE,WAAW,CAAC;CAC5B;AAED;;;;;;;GAOG;AACH,MAAM,WAAW,QAAQ;IACvB,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB,QAAQ,CAAC,UAAU,EAAE,CAAC,OAAO,EAAE,WAAW,KAAK,WAAW,CAAC;CAC5D;AA+DD;;;;;;;;;;;GAWG;AACH,wBAAgB,cAAc,CAC5B,OAAO,EAAE,OAAO,EAChB,UAAU,EAAE,UAAU,EACtB,KAAK,EAAE,SAAS,QAAQ,EAAE,GACzB,OAAO,CAsCT"}
|
package/dist/decode/rules.js
CHANGED
|
@@ -1,4 +1,4 @@
|
|
|
1
|
-
export { ASPECT, PARTICLE_TAG, tagOf, wordEndingAt, wordStartingAt, } from "./edge-context.js";
|
|
1
|
+
export { ASPECT, PARTICLE_TAG, tagOf, wordEndingAt, wordsEndingAt, wordStartingAt, wordsStartingAt, } from "./edge-context.js";
|
|
2
2
|
/**
|
|
3
3
|
* Whether every position can still be left, and the end still reached.
|
|
4
4
|
*
|
package/dist/decode/rules.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"rules.js","sourceRoot":"","sources":["../../src/decode/rules.ts"],"names":[],"mappings":"AAGA,OAAO,EACL,MAAM,EACN,YAAY,EACZ,KAAK,EACL,YAAY,EACZ,cAAc,
|
|
1
|
+
{"version":3,"file":"rules.js","sourceRoot":"","sources":["../../src/decode/rules.ts"],"names":[],"mappings":"AAGA,OAAO,EACL,MAAM,EACN,YAAY,EACZ,KAAK,EACL,YAAY,EACZ,aAAa,EACb,cAAc,EACd,eAAe,GAChB,MAAM,mBAAmB,CAAC;AAsC3B;;;;;;;GAOG;AACH,SAAS,aAAa,CAAC,OAAgB;IACrC,MAAM,MAAM,GAAG,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC;IACzC,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,MAAM,GAAG,CAAC,EAAE,EAAE,GAAG,EAAE,CAAC,KAAK,CAAC,CAAC;IAChE,OAAO,CAAC,MAAM,CAAC,GAAG,IAAI,CAAC;IACvB,KAAK,IAAI,EAAE,GAAG,MAAM,GAAG,CAAC,EAAE,EAAE,IAAI,CAAC,EAAE,EAAE,EAAE,EAAE,CAAC;QACxC,0EAA0E;QAC1E,OAAO,CAAC,EAAE,CAAC,GAAG,CAAC,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,CAC1C,CAAC,IAAI,EAAE,EAAE,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,KAAK,CACpC,CAAC;IACJ,CAAC;IACD,qDAAqD;IACrD,OAAO,OAAO,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC;AAC7B,CAAC;AAED;;GAEG;AACH,SAAS,UAAU,CACjB,KAA0B,EAC1B,OAAoB;IAEpB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,OAAO,GAAG,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC,CAAC;QACzC,IAAI,OAAO,KAAK,MAAM,EAAE,CAAC;YACvB,OAAO,OAAO,CAAC;QACjB,CAAC;IACH,CAAC;IACD,OAAO,MAAM,CAAC;AAChB,CAAC;AAED;;;;;;;GAOG;AACH,SAAS,YAAY,CACnB,KAA6B,EAC7B,MAAgC,EAChC,SAA2B;IAE3B,MAAM,OAAO,GAAG,IAAI,GAAG,CACrB,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,CAChE,CAAC;IACF,KAAK,MAAM,KAAK,IAAI,KAAK,EAAE,CAAC;QAC1B,IAAI,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC;YAChD,SAAS,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;QACvB,CAAC;IACH,CAAC;AACH,CAAC;AAED;;;;;;;;;;;GAWG;AACH,MAAM,UAAU,cAAc,CAC5B,OAAgB,EAChB,UAAsB,EACtB,KAA0B;IAE1B,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACvB,OAAO,OAAO,CAAC;IACjB,CAAC;IACD,MAAM,SAAS,GAAG,IAAI,GAAG,EAAe,CAAC;IACzC,MAAM,MAAM,GAAG,IAAI,GAAG,EAAe,CAAC;IAEtC,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,KAAK,EAAE,CAAC;QAClC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YACzB,MAAM,OAAO,GAAG,UAAU,CAAC,KAAK,EAAE;gBAChC,UAAU;gBACV,UAAU,EAAE,OAAO,CAAC,UAAU;gBAC9B,IAAI;aACL,CAAC,CAAC;YACH,IAAI,OAAO,KAAK,QAAQ,EAAE,CAAC;gBACzB,SAAS,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;YACtB,CAAC;iBAAM,IAAI,OAAO,KAAK,OAAO,EAAE,CAAC;gBAC/B,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;YACnB,CAAC;QACH,CAAC;IACH,CAAC;IAED,4EAA4E;IAC5E,kEAAkE;IAClE,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,KAAK,EAAE,CAAC;QAClC,YAAY,CAAC,KAAK,EAAE,MAAM,EAAE,SAAS,CAAC,CAAC;IACzC,CAAC;IAED,IAAI,SAAS,CAAC,IAAI,KAAK,CAAC,EAAE,CAAC;QACzB,OAAO,OAAO,CAAC;IACjB,CAAC;IACD,MAAM,KAAK,GAAY;QACrB,UAAU,EAAE,OAAO,CAAC,UAAU;QAC9B,KAAK,EAAE,OAAO,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CACjC,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,SAAS,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAC7C;KACF,CAAC;IACF,OAAO,aAAa,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,OAAO,CAAC;AAChD,CAAC"}
|
|
@@ -0,0 +1,33 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* What a Han run is handed about the numbers around it.
|
|
3
|
+
*
|
|
4
|
+
* A run is decoded on its own, so the digits either side of it are invisible to
|
|
5
|
+
* it unless something puts them there. This is what puts them there, as the
|
|
6
|
+
* 汉字 they would have been written with.
|
|
7
|
+
*/
|
|
8
|
+
import type { NumeralSegment } from "../numerals/text.js";
|
|
9
|
+
import type { DecodeContext } from "./lattice-slice.js";
|
|
10
|
+
import type { TextRun } from "./runs.js";
|
|
11
|
+
/**
|
|
12
|
+
* The 汉字 a number in front of a Han run stands for, for that run's decode.
|
|
13
|
+
*
|
|
14
|
+
* Only the last segment of the run before, because only that one touches the
|
|
15
|
+
* Han: the D of 3D银行 comes between them, and a decode of 银行 that saw 三
|
|
16
|
+
* beside it would be reading a text nobody wrote.
|
|
17
|
+
*/
|
|
18
|
+
export declare function numeralBefore(segments: readonly NumeralSegment[]): string;
|
|
19
|
+
/**
|
|
20
|
+
* The same, for the number after a Han run: only the first segment of it.
|
|
21
|
+
*/
|
|
22
|
+
export declare function numeralAfter(segments: readonly NumeralSegment[]): string;
|
|
23
|
+
/**
|
|
24
|
+
* The 汉字 either side of a Han run, once the digits around it have been read.
|
|
25
|
+
*
|
|
26
|
+
* 那条河长300公里 is four runs, and the 长 ends the first of them, so a rule
|
|
27
|
+
* asking what the 长 is measured in used to see nothing at all: the numeral is
|
|
28
|
+
* one run away and the 公里 two. This hands the decode what the sentence would
|
|
29
|
+
* have said spelled out, which is the number in front, and the number behind
|
|
30
|
+
* with the Han that carries on from it.
|
|
31
|
+
*/
|
|
32
|
+
export declare function contextAround(runs: readonly TextRun[], read: readonly (readonly NumeralSegment[])[], at: number): DecodeContext;
|
|
33
|
+
//# sourceMappingURL=run-context.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"run-context.d.ts","sourceRoot":"","sources":["../../src/decode/run-context.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AACH,OAAO,KAAK,EAAE,cAAc,EAAE,MAAM,qBAAqB,CAAC;AAE1D,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,oBAAoB,CAAC;AACxD,OAAO,KAAK,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AAazC;;;;;;GAMG;AACH,wBAAgB,aAAa,CAAC,QAAQ,EAAE,SAAS,cAAc,EAAE,GAAG,MAAM,CAEzE;AAED;;GAEG;AACH,wBAAgB,YAAY,CAAC,QAAQ,EAAE,SAAS,cAAc,EAAE,GAAG,MAAM,CAExE;AAED;;;;;;;;GAQG;AACH,wBAAgB,aAAa,CAC3B,IAAI,EAAE,SAAS,OAAO,EAAE,EACxB,IAAI,EAAE,SAAS,CAAC,SAAS,cAAc,EAAE,CAAC,EAAE,EAC5C,EAAE,EAAE,MAAM,GACT,aAAa,CAUf"}
|
|
@@ -0,0 +1,47 @@
|
|
|
1
|
+
import { toCharacters } from "../script/characters.js";
|
|
2
|
+
/**
|
|
3
|
+
* How far past a number a run's decode is allowed to see, in characters.
|
|
4
|
+
*
|
|
5
|
+
* The same four {@link import("./edge-context.js").wordStartingAt} looks ahead
|
|
6
|
+
* by, and for the same reason: four characters is as far as a rule can reach,
|
|
7
|
+
* so carrying more would be lattice nobody reads. A prefix can only take words
|
|
8
|
+
* away and never invent one, so cutting the run here costs the decode nothing
|
|
9
|
+
* it would otherwise have had.
|
|
10
|
+
*/
|
|
11
|
+
const READS_AHEAD = 4;
|
|
12
|
+
/**
|
|
13
|
+
* The 汉字 a number in front of a Han run stands for, for that run's decode.
|
|
14
|
+
*
|
|
15
|
+
* Only the last segment of the run before, because only that one touches the
|
|
16
|
+
* Han: the D of 3D银行 comes between them, and a decode of 银行 that saw 三
|
|
17
|
+
* beside it would be reading a text nobody wrote.
|
|
18
|
+
*/
|
|
19
|
+
export function numeralBefore(segments) {
|
|
20
|
+
return segments.at(-1)?.hanzi ?? "";
|
|
21
|
+
}
|
|
22
|
+
/**
|
|
23
|
+
* The same, for the number after a Han run: only the first segment of it.
|
|
24
|
+
*/
|
|
25
|
+
export function numeralAfter(segments) {
|
|
26
|
+
return segments[0]?.hanzi ?? "";
|
|
27
|
+
}
|
|
28
|
+
/**
|
|
29
|
+
* The 汉字 either side of a Han run, once the digits around it have been read.
|
|
30
|
+
*
|
|
31
|
+
* 那条河长300公里 is four runs, and the 长 ends the first of them, so a rule
|
|
32
|
+
* asking what the 长 is measured in used to see nothing at all: the numeral is
|
|
33
|
+
* one run away and the 公里 two. This hands the decode what the sentence would
|
|
34
|
+
* have said spelled out, which is the number in front, and the number behind
|
|
35
|
+
* with the Han that carries on from it.
|
|
36
|
+
*/
|
|
37
|
+
export function contextAround(runs, read, at) {
|
|
38
|
+
const said = numeralAfter(read[at + 1] ?? []);
|
|
39
|
+
const following = runs[at + 2];
|
|
40
|
+
return {
|
|
41
|
+
before: numeralBefore(read[at - 1] ?? []),
|
|
42
|
+
after: said === "" || following?.isHan !== true
|
|
43
|
+
? said
|
|
44
|
+
: said + toCharacters(following.text).slice(0, READS_AHEAD).join(""),
|
|
45
|
+
};
|
|
46
|
+
}
|
|
47
|
+
//# sourceMappingURL=run-context.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"run-context.js","sourceRoot":"","sources":["../../src/decode/run-context.ts"],"names":[],"mappings":"AAQA,OAAO,EAAE,YAAY,EAAE,MAAM,yBAAyB,CAAC;AAIvD;;;;;;;;GAQG;AACH,MAAM,WAAW,GAAG,CAAC,CAAC;AAEtB;;;;;;GAMG;AACH,MAAM,UAAU,aAAa,CAAC,QAAmC;IAC/D,OAAO,QAAQ,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,KAAK,IAAI,EAAE,CAAC;AACtC,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,YAAY,CAAC,QAAmC;IAC9D,OAAO,QAAQ,CAAC,CAAC,CAAC,EAAE,KAAK,IAAI,EAAE,CAAC;AAClC,CAAC;AAED;;;;;;;;GAQG;AACH,MAAM,UAAU,aAAa,CAC3B,IAAwB,EACxB,IAA4C,EAC5C,EAAU;IAEV,MAAM,IAAI,GAAG,YAAY,CAAC,IAAI,CAAC,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;IAC9C,MAAM,SAAS,GAAG,IAAI,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC;IAC/B,OAAO;QACL,MAAM,EAAE,aAAa,CAAC,IAAI,CAAC,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;QACzC,KAAK,EACH,IAAI,KAAK,EAAE,IAAI,SAAS,EAAE,KAAK,KAAK,IAAI;YACtC,CAAC,CAAC,IAAI;YACN,CAAC,CAAC,IAAI,GAAG,YAAY,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,WAAW,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC;KACzE,CAAC;AACJ,CAAC"}
|
|
@@ -29,13 +29,14 @@ export declare function wordFrom(dictionary: Dictionary, lattice: Lattice, group
|
|
|
29
29
|
*/
|
|
30
30
|
export declare function ruledLattice(dictionary: Dictionary, run: string, rules: readonly EdgeRule[], hints: ResolvedHints | undefined): Lattice;
|
|
31
31
|
/**
|
|
32
|
-
* Cut a decode of the whole lattice down to the run
|
|
32
|
+
* Cut a decode of the whole lattice down to the run inside it.
|
|
33
33
|
*
|
|
34
|
-
* Everything
|
|
35
|
-
*
|
|
36
|
-
* work.
|
|
34
|
+
* Everything outside `at` and `to` was context rather than run — 汉字 a caller
|
|
35
|
+
* put either side so that the decode could see them — and is dropped once it
|
|
36
|
+
* has done its work. A unit is kept only where it lies wholly inside the run,
|
|
37
|
+
* which is what `isJoinedAt` guarantees for the ones that cannot be cut.
|
|
37
38
|
*/
|
|
38
|
-
export declare function runGroups<Unit extends ReadingUnit>({ lattice, at }: RunLattice, units: readonly Unit[]): readonly (readonly Unit[])[];
|
|
39
|
+
export declare function runGroups<Unit extends ReadingUnit>({ lattice, at, to }: RunLattice, units: readonly Unit[]): readonly (readonly Unit[])[];
|
|
39
40
|
/**
|
|
40
41
|
* Decode a Han run with the lattice: the recommended path.
|
|
41
42
|
*
|
|
@@ -43,10 +44,12 @@ export declare function runGroups<Unit extends ReadingUnit>({ lattice, at }: Run
|
|
|
43
44
|
* lock the positions that cannot vary, score the rest — where the greedy
|
|
44
45
|
* baseline commits to the longest match at each position and never revisits it.
|
|
45
46
|
*
|
|
46
|
-
* `before`
|
|
47
|
-
* not report: the 汉字 a number written in digits would have been
|
|
48
|
-
*
|
|
49
|
-
*
|
|
50
|
-
*
|
|
47
|
+
* `before` and `after` are 汉字 standing either side of the run that the decode
|
|
48
|
+
* should see and not report: the 汉字 a number written in digits would have been
|
|
49
|
+
* written with, and whatever carries on from it. A run is decoded on its own
|
|
50
|
+
* without them, and a run reached that way has no idea what surrounded it —
|
|
51
|
+
* 个人 alone is the word `gèrén`, while the 两 of 2个人 makes 两个 the word and
|
|
52
|
+
* leaves 人 to itself, and 那条河长 ends where 那条河长三百公里 goes on to say
|
|
53
|
+
* how long the river is.
|
|
51
54
|
*/
|
|
52
55
|
//# sourceMappingURL=run-lattice.d.ts.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"run-lattice.d.ts","sourceRoot":"","sources":["../../src/decode/run-lattice.ts"],"names":[],"mappings":"AAAA,OAAO,EAAgB,KAAK,OAAO,EAAE,MAAM,cAAc,CAAC;AAC1D,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,YAAY,CAAC;AAChD,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,oBAAoB,CAAC;AAErD,OAAO,EAAE,UAAU,EAAE,MAAM,oBAAoB,CAAC;AAShD,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,6BAA6B,CAAC;AAC9D,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,cAAc,CAAC;AAChD,OAAO,EAAkB,KAAK,QAAQ,EAAE,MAAM,YAAY,CAAC;AAC3D,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,WAAW,CAAC;AAE7C;;;;;;;GAOG;AACH,wBAAgB,UAAU,CAAC,IAAI,SAAS,WAAW,EACjD,KAAK,EAAE,SAAS,IAAI,EAAE,EACtB,UAAU,EAAE,SAAS,MAAM,EAAE,GAC5B,SAAS,CAAC,SAAS,IAAI,EAAE,CAAC,EAAE,CAkB9B;AAED;;GAEG;AACH,wBAAgB,QAAQ,CACtB,UAAU,EAAE,UAAU,EACtB,OAAO,EAAE,OAAO,EAChB,KAAK,EAAE,SAAS,WAAW,EAAE,GAC5B,WAAW,CAiBb;AAED;;;;;;;GAOG;AACH,wBAAgB,YAAY,CAC1B,UAAU,EAAE,UAAU,EACtB,GAAG,EAAE,MAAM,EACX,KAAK,EAAE,SAAS,QAAQ,EAAE,EAC1B,KAAK,EAAE,aAAa,GAAG,SAAS,GAC/B,OAAO,CAMT;AAED
|
|
1
|
+
{"version":3,"file":"run-lattice.d.ts","sourceRoot":"","sources":["../../src/decode/run-lattice.ts"],"names":[],"mappings":"AAAA,OAAO,EAAgB,KAAK,OAAO,EAAE,MAAM,cAAc,CAAC;AAC1D,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,YAAY,CAAC;AAChD,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,oBAAoB,CAAC;AAErD,OAAO,EAAE,UAAU,EAAE,MAAM,oBAAoB,CAAC;AAShD,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,6BAA6B,CAAC;AAC9D,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,cAAc,CAAC;AAChD,OAAO,EAAkB,KAAK,QAAQ,EAAE,MAAM,YAAY,CAAC;AAC3D,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,WAAW,CAAC;AAE7C;;;;;;;GAOG;AACH,wBAAgB,UAAU,CAAC,IAAI,SAAS,WAAW,EACjD,KAAK,EAAE,SAAS,IAAI,EAAE,EACtB,UAAU,EAAE,SAAS,MAAM,EAAE,GAC5B,SAAS,CAAC,SAAS,IAAI,EAAE,CAAC,EAAE,CAkB9B;AAED;;GAEG;AACH,wBAAgB,QAAQ,CACtB,UAAU,EAAE,UAAU,EACtB,OAAO,EAAE,OAAO,EAChB,KAAK,EAAE,SAAS,WAAW,EAAE,GAC5B,WAAW,CAiBb;AAED;;;;;;;GAOG;AACH,wBAAgB,YAAY,CAC1B,UAAU,EAAE,UAAU,EACtB,GAAG,EAAE,MAAM,EACX,KAAK,EAAE,SAAS,QAAQ,EAAE,EAC1B,KAAK,EAAE,aAAa,GAAG,SAAS,GAC/B,OAAO,CAMT;AAED;;;;;;;GAOG;AACH,wBAAgB,SAAS,CAAC,IAAI,SAAS,WAAW,EAChD,EAAE,OAAO,EAAE,EAAE,EAAE,EAAE,EAAE,EAAE,UAAU,EAC/B,KAAK,EAAE,SAAS,IAAI,EAAE,GACrB,SAAS,CAAC,SAAS,IAAI,EAAE,CAAC,EAAE,CAK9B;AAED;;;;;;;;;;;;;;GAcG"}
|
|
@@ -7,7 +7,7 @@ export { runLattice } from "./lattice-slice.js";
|
|
|
7
7
|
* is where a lattice is cut down to a run and the words are read back out of
|
|
8
8
|
* it. `decode.ts` is what asks for one.
|
|
9
9
|
*/
|
|
10
|
-
import { decodeSpacing } from "./decode.js";
|
|
10
|
+
import { decodeSpacing } from "./lattice-decode.js";
|
|
11
11
|
import { applyEdgeRules } from "./rules.js";
|
|
12
12
|
/**
|
|
13
13
|
* Gather the reading units into the words the boundaries make of them.
|
|
@@ -66,14 +66,15 @@ export function ruledLattice(dictionary, run, rules, hints) {
|
|
|
66
66
|
return applyEdgeRules(buildLattice(dictionary, run, hints), dictionary, rules);
|
|
67
67
|
}
|
|
68
68
|
/**
|
|
69
|
-
* Cut a decode of the whole lattice down to the run
|
|
69
|
+
* Cut a decode of the whole lattice down to the run inside it.
|
|
70
70
|
*
|
|
71
|
-
* Everything
|
|
72
|
-
*
|
|
73
|
-
* work.
|
|
71
|
+
* Everything outside `at` and `to` was context rather than run — 汉字 a caller
|
|
72
|
+
* put either side so that the decode could see them — and is dropped once it
|
|
73
|
+
* has done its work. A unit is kept only where it lies wholly inside the run,
|
|
74
|
+
* which is what `isJoinedAt` guarantees for the ones that cannot be cut.
|
|
74
75
|
*/
|
|
75
|
-
export function runGroups({ lattice, at }, units) {
|
|
76
|
-
return groupUnits(units.filter((unit) => unit.from >= at), decodeSpacing(lattice).filter((from) => from >= at));
|
|
76
|
+
export function runGroups({ lattice, at, to }, units) {
|
|
77
|
+
return groupUnits(units.filter((unit) => unit.from >= at && unit.to <= to), decodeSpacing(lattice).filter((from) => from >= at && from < to));
|
|
77
78
|
}
|
|
78
79
|
/**
|
|
79
80
|
* Decode a Han run with the lattice: the recommended path.
|
|
@@ -82,10 +83,12 @@ export function runGroups({ lattice, at }, units) {
|
|
|
82
83
|
* lock the positions that cannot vary, score the rest — where the greedy
|
|
83
84
|
* baseline commits to the longest match at each position and never revisits it.
|
|
84
85
|
*
|
|
85
|
-
* `before`
|
|
86
|
-
* not report: the 汉字 a number written in digits would have been
|
|
87
|
-
*
|
|
88
|
-
*
|
|
89
|
-
*
|
|
86
|
+
* `before` and `after` are 汉字 standing either side of the run that the decode
|
|
87
|
+
* should see and not report: the 汉字 a number written in digits would have been
|
|
88
|
+
* written with, and whatever carries on from it. A run is decoded on its own
|
|
89
|
+
* without them, and a run reached that way has no idea what surrounded it —
|
|
90
|
+
* 个人 alone is the word `gèrén`, while the 两 of 2个人 makes 两个 the word and
|
|
91
|
+
* leaves 人 to itself, and 那条河长 ends where 那条河长三百公里 goes on to say
|
|
92
|
+
* how long the river is.
|
|
90
93
|
*/
|
|
91
94
|
//# sourceMappingURL=run-lattice.js.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"run-lattice.js","sourceRoot":"","sources":["../../src/decode/run-lattice.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,YAAY,EAAgB,MAAM,cAAc,CAAC;AAI1D,OAAO,EAAE,UAAU,EAAE,MAAM,oBAAoB,CAAC;AAChD;;;;;;GAMG;AACH,OAAO,EAAE,aAAa,EAAE,MAAM,
|
|
1
|
+
{"version":3,"file":"run-lattice.js","sourceRoot":"","sources":["../../src/decode/run-lattice.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,YAAY,EAAgB,MAAM,cAAc,CAAC;AAI1D,OAAO,EAAE,UAAU,EAAE,MAAM,oBAAoB,CAAC;AAChD;;;;;;GAMG;AACH,OAAO,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AAGpD,OAAO,EAAE,cAAc,EAAiB,MAAM,YAAY,CAAC;AAG3D;;;;;;;GAOG;AACH,MAAM,UAAU,UAAU,CACxB,KAAsB,EACtB,UAA6B;IAE7B,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;IACvD,MAAM,IAAI,GAAG,IAAI,GAAG,CAAC,UAAU,CAAC,MAAM,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;IAChE,MAAM,MAAM,GAAa,EAAE,CAAC;IAC5B,IAAI,IAAI,GAAW,EAAE,CAAC;IAEtB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAC3C,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YAClB,IAAI,GAAG,EAAE,CAAC;QACZ,CAAC;QACD,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAClB,CAAC;IACD,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACpB,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IACpB,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,QAAQ,CACtB,UAAsB,EACtB,OAAgB,EAChB,KAA6B;IAE7B,MAAM,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;IACvB,MAAM,IAAI,GAAG,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;IAC1B,wEAAwE;IACxE,IAAI,KAAK,KAAK,SAAS,IAAI,IAAI,KAAK,SAAS,EAAE,CAAC;QAC9C,MAAM,IAAI,KAAK,CAAC,6CAA6C,CAAC,CAAC;IACjE,CAAC;IACD,MAAM,IAAI,GAAG,OAAO,CAAC,UAAU,CAAC,KAAK,CAAC,KAAK,CAAC,IAAI,EAAE,IAAI,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IACpE,MAAM,KAAK,GAAG,UAAU,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;IAEtC,OAAO;QACL,IAAI;QACJ,OAAO,EAAE,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,GAAG,IAAI,CAAC,OAAO,CAAC,CAAC;QACnD,YAAY,EAAE,KAAK,EAAE,YAAY,IAAI,KAAK;QAC1C,YAAY,EAAE,KAAK,EAAE,YAAY,IAAI,EAAE;QACvC,OAAO,EAAE,KAAK,KAAK,SAAS;KAC7B,CAAC;AACJ,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,YAAY,CAC1B,UAAsB,EACtB,GAAW,EACX,KAA0B,EAC1B,KAAgC;IAEhC,OAAO,cAAc,CACnB,YAAY,CAAC,UAAU,EAAE,GAAG,EAAE,KAAK,CAAC,EACpC,UAAU,EACV,KAAK,CACN,CAAC;AACJ,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,SAAS,CACvB,EAAE,OAAO,EAAE,EAAE,EAAE,EAAE,EAAc,EAC/B,KAAsB;IAEtB,OAAO,UAAU,CACf,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,IAAI,EAAE,IAAI,IAAI,CAAC,EAAE,IAAI,EAAE,CAAC,EACxD,aAAa,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,IAAI,EAAE,IAAI,IAAI,GAAG,EAAE,CAAC,CACjE,CAAC;AACJ,CAAC;AAED;;;;;;;;;;;;;;GAcG"}
|