@chaffjs/lang-en 0.3.0 → 0.5.0
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/dist/lexicons.d.ts.map +1 -1
- package/dist/lexicons.js +6 -1
- package/dist/lexicons.js.map +1 -1
- package/dist/pos.d.ts.map +1 -1
- package/dist/pos.js +9 -2
- package/dist/pos.js.map +1 -1
- package/lexicons/contraction.yaml +34 -0
- package/package.json +1 -1
- package/src/lexicons.ts +6 -1
- package/src/pos.ts +9 -2
package/dist/lexicons.d.ts.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"lexicons.d.ts","sourceRoot":"","sources":["../src/lexicons.ts"],"names":[],"mappings":"AAIA,OAAO,KAAK,EAAE,OAAO,EAAgB,MAAM,gBAAgB,CAAC;
|
|
1
|
+
{"version":3,"file":"lexicons.d.ts","sourceRoot":"","sources":["../src/lexicons.ts"],"names":[],"mappings":"AAIA,OAAO,KAAK,EAAE,OAAO,EAAgB,MAAM,gBAAgB,CAAC;AAiB5D;;;GAGG;AACH,eAAO,MAAM,YAAY,GAAI,MAAK,MAAY,KAAG,MAAM,CAAC,MAAM,EAAE,OAAO,CAQ7D,CAAC"}
|
package/dist/lexicons.js
CHANGED
|
@@ -8,7 +8,12 @@ const toEntry = (raw) => {
|
|
|
8
8
|
if (!isRecord(raw) || typeof raw["pattern"] !== "string")
|
|
9
9
|
return undefined;
|
|
10
10
|
const weight = raw["weight"];
|
|
11
|
-
|
|
11
|
+
const instead = raw["instead_of"];
|
|
12
|
+
return {
|
|
13
|
+
pattern: raw["pattern"],
|
|
14
|
+
weight: typeof weight === "number" ? weight : undefined,
|
|
15
|
+
instead_of: typeof instead === "string" ? instead : undefined,
|
|
16
|
+
};
|
|
12
17
|
};
|
|
13
18
|
/**
|
|
14
19
|
* 語彙表はアダプタが持つ。detector は共通で、これだけが言語別。spec §11。
|
package/dist/lexicons.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"lexicons.js","sourceRoot":"","sources":["../src/lexicons.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,YAAY,EAAE,WAAW,EAAE,MAAM,SAAS,CAAC;AACpD,OAAO,EAAE,aAAa,EAAE,MAAM,UAAU,CAAC;AACzC,OAAO,EAAE,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;AAC1C,OAAO,EAAE,KAAK,EAAE,MAAM,MAAM,CAAC;AAG7B,MAAM,GAAG,GAAG,IAAI,CAAC,OAAO,CAAC,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,EAAE,UAAU,CAAC,CAAC;AAE5E,MAAM,QAAQ,GAAG,CAAC,KAAc,EAAoC,EAAE,CAAC,OAAO,KAAK,KAAK,QAAQ,IAAI,KAAK,KAAK,IAAI,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC;AAE5I,MAAM,OAAO,GAAG,CAAC,GAAY,EAA4B,EAAE;IACzD,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,OAAO,GAAG,CAAC,SAAS,CAAC,KAAK,QAAQ;QAAE,OAAO,SAAS,CAAC;IAC3E,MAAM,MAAM,GAAG,GAAG,CAAC,QAAQ,CAAC,CAAC;IAC7B,OAAO,
|
|
1
|
+
{"version":3,"file":"lexicons.js","sourceRoot":"","sources":["../src/lexicons.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,YAAY,EAAE,WAAW,EAAE,MAAM,SAAS,CAAC;AACpD,OAAO,EAAE,aAAa,EAAE,MAAM,UAAU,CAAC;AACzC,OAAO,EAAE,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;AAC1C,OAAO,EAAE,KAAK,EAAE,MAAM,MAAM,CAAC;AAG7B,MAAM,GAAG,GAAG,IAAI,CAAC,OAAO,CAAC,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,EAAE,UAAU,CAAC,CAAC;AAE5E,MAAM,QAAQ,GAAG,CAAC,KAAc,EAAoC,EAAE,CAAC,OAAO,KAAK,KAAK,QAAQ,IAAI,KAAK,KAAK,IAAI,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC;AAE5I,MAAM,OAAO,GAAG,CAAC,GAAY,EAA4B,EAAE;IACzD,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,OAAO,GAAG,CAAC,SAAS,CAAC,KAAK,QAAQ;QAAE,OAAO,SAAS,CAAC;IAC3E,MAAM,MAAM,GAAG,GAAG,CAAC,QAAQ,CAAC,CAAC;IAC7B,MAAM,OAAO,GAAG,GAAG,CAAC,YAAY,CAAC,CAAC;IAClC,OAAO;QACL,OAAO,EAAE,GAAG,CAAC,SAAS,CAAC;QACvB,MAAM,EAAE,OAAO,MAAM,KAAK,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,SAAS;QACvD,UAAU,EAAE,OAAO,OAAO,KAAK,QAAQ,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,SAAS;KAC9D,CAAC;AACJ,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,MAAc,GAAG,EAA2B,EAAE,CACzE,WAAW,CAAC,GAAG,CAAC;KACb,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC;KACxC,MAAM,CAA0B,CAAC,GAAG,EAAE,IAAI,EAAE,EAAE;IAC7C,MAAM,GAAG,GAAY,KAAK,CAAC,YAAY,CAAC,IAAI,CAAC,GAAG,EAAE,IAAI,CAAC,EAAE,MAAM,CAAC,CAAC,CAAC;IAClE,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,OAAO,GAAG,CAAC,IAAI,CAAC,KAAK,QAAQ,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;QAAE,OAAO,GAAG,CAAC;IAClG,MAAM,OAAO,GAAG,GAAG,CAAC,SAAS,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,KAAK,SAAS,CAAC,CAAC;IACnF,OAAO,EAAE,GAAG,GAAG,EAAE,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,EAAE,OAAO,EAAE,CAAC;AAC1C,CAAC,EAAE,EAAE,CAAC,CAAC"}
|
package/dist/pos.d.ts.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"pos.d.ts","sourceRoot":"","sources":["../src/pos.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,KAAK,EAAE,MAAM,gBAAgB,CAAC;AA4E5C,eAAO,MAAM,IAAI,GAAI,KAAK,MAAM,KAAG,MAAiE,CAAC;
|
|
1
|
+
{"version":3,"file":"pos.d.ts","sourceRoot":"","sources":["../src/pos.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,KAAK,EAAE,MAAM,gBAAgB,CAAC;AA4E5C,eAAO,MAAM,IAAI,GAAI,KAAK,MAAM,KAAG,MAAiE,CAAC;AA4BrG,eAAO,MAAM,OAAO,QAAO,IAE1B,CAAC;AAEF,eAAO,MAAM,OAAO,QAAO,OAAoC,CAAC;AAwBhE,eAAO,MAAM,QAAQ,GAAI,MAAM,MAAM,KAAG,KAAK,EAAE,GAAG,SAIjD,CAAC"}
|
package/dist/pos.js
CHANGED
|
@@ -67,14 +67,21 @@ export const upos = (tag) => (PUNCTUATION.has(tag) ? "PUNCT" : (BY_TAG[tag] ?? "
|
|
|
67
67
|
* 受動は be + 過去分詞。過去分詞だけでは完了形(has reviewed)と見分けられないので、
|
|
68
68
|
* 直前の be を見る。間に副詞が挟まる(was quickly approved)ぶんだけ遡る。
|
|
69
69
|
*/
|
|
70
|
-
|
|
70
|
+
/**
|
|
71
|
+
* be と過去分詞の間に立てるもの。副詞と、副詞をつなぐ接続詞(was fully and carefully reviewed)。
|
|
72
|
+
*
|
|
73
|
+
* これでも拾えない形がある。`was fully and finally approved` では wink が `approved` を
|
|
74
|
+
* VBN ではなく **VBD** と付けるため、そもそも過去分詞として見えない。解析器の限界で、
|
|
75
|
+
* VBD も受動と見なすと `The team was here and approved it` まで受動になる。直さない。
|
|
76
|
+
*/
|
|
77
|
+
const SKIPPABLE = new Set(["RB", "RBR", "RBS", "CC", ","]);
|
|
71
78
|
const BE = new Set(["be", "am", "is", "are", "was", "were", "been", "being"]);
|
|
72
79
|
const isBe = (entry) => BE.has(entry.lemma ?? entry.value.toLowerCase()) || BE.has(entry.value.toLowerCase());
|
|
73
80
|
const isPassive = (tagged, at) => {
|
|
74
81
|
if (tagged[at]?.pos !== "VBN")
|
|
75
82
|
return false;
|
|
76
83
|
const before = tagged.slice(0, at).reverse();
|
|
77
|
-
const head = before.find((entry) => !
|
|
84
|
+
const head = before.find((entry) => !SKIPPABLE.has(entry.pos));
|
|
78
85
|
return head !== undefined && isBe(head);
|
|
79
86
|
};
|
|
80
87
|
const state = { ready: undefined };
|
package/dist/pos.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"pos.js","sourceRoot":"","sources":["../src/pos.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,aAAa,EAAE,MAAM,aAAa,CAAC;AAG5C,MAAM,OAAO,GAAG,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;AAO/C,MAAM,QAAQ,GAAG,CAAC,KAAc,EAAoC,EAAE,CAAC,OAAO,KAAK,KAAK,QAAQ,IAAI,KAAK,KAAK,IAAI,CAAC;AAEnH,MAAM,UAAU,GAAG,CAAC,KAAc,EAAqD,EAAE,CAAC,OAAO,KAAK,KAAK,UAAU,CAAC;AAEtH,MAAM,QAAQ,GAAG,CAAC,KAAc,EAAmB,EAAE,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,OAAO,KAAK,CAAC,OAAO,CAAC,KAAK,QAAQ,IAAI,OAAO,KAAK,CAAC,KAAK,CAAC,KAAK,QAAQ,CAAC;AAEhJ,MAAM,OAAO,GAAG,CAAC,KAAc,EAAsB,EAAE,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;AAE5F,kEAAkE;AAClE,MAAM,UAAU,GAAG,CAAC,KAA8B,EAAE,IAAY,EAAE,IAAwB,EAAW,EAAE;IACrG,MAAM,MAAM,GAAY,KAAK,CAAC,IAAI,CAAC,CAAC;IACpC,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC;QAAE,MAAM,IAAI,KAAK,CAAC,qBAAqB,IAAI,WAAW,CAAC,CAAC;IAC/E,OAAO,OAAO,CAAC,KAAK,CAAC,MAAM,EAAE,KAAK,EAAE,IAAI,CAAC,CAAC;AAC5C,CAAC,CAAC;AAEF,MAAM,KAAK,GAAG,GAAW,EAAE;IACzB,MAAM,OAAO,GAAY,OAAO,CAAC,iBAAiB,CAAC,CAAC;IACpD,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC;QAAE,MAAM,IAAI,KAAK,CAAC,oCAAoC,CAAC,CAAC;IAChF,MAAM,MAAM,GAAY,OAAO,EAAE,CAAC;IAClC,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC;QAAE,MAAM,IAAI,KAAK,CAAC,oCAAoC,CAAC,CAAC;IAC7E,OAAO,MAAM,CAAC;AAChB,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,MAAM,GAAqC;IAC/C,EAAE,EAAE,OAAO;IACX,EAAE,EAAE,KAAK;IACT,EAAE,EAAE,KAAK;IACT,EAAE,EAAE,MAAM;IACV,EAAE,EAAE,GAAG;IACP,EAAE,EAAE,KAAK;IACT,EAAE,EAAE,KAAK;IACT,GAAG,EAAE,KAAK;IACV,GAAG,EAAE,KAAK;IACV,EAAE,EAAE,KAAK;IACT,EAAE,EAAE,MAAM;IACV,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,OAAO;IACZ,IAAI,EAAE,OAAO;IACb,GAAG,EAAE,KAAK;IACV,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,IAAI,EAAE,MAAM;IACZ,EAAE,EAAE,KAAK;IACT,GAAG,EAAE,KAAK;IACV,GAAG,EAAE,KAAK;IACV,EAAE,EAAE,KAAK;IACT,GAAG,EAAE,KAAK;IACV,EAAE,EAAE,MAAM;IACV,EAAE,EAAE,MAAM;IACV,EAAE,EAAE,MAAM;IACV,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,KAAK;IACV,EAAE,EAAE,MAAM;IACV,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,KAAK;CACX,CAAC;AAEF,MAAM,WAAW,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAE7E,MAAM,CAAC,MAAM,IAAI,GAAG,CAAC,GAAW,EAAU,EAAE,CAAC,CAAC,WAAW,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,GAAG,CAAC,CAAC,CAAC;AAErG;;;GAGG;AACH,MAAM,
|
|
1
|
+
{"version":3,"file":"pos.js","sourceRoot":"","sources":["../src/pos.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,aAAa,EAAE,MAAM,aAAa,CAAC;AAG5C,MAAM,OAAO,GAAG,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;AAO/C,MAAM,QAAQ,GAAG,CAAC,KAAc,EAAoC,EAAE,CAAC,OAAO,KAAK,KAAK,QAAQ,IAAI,KAAK,KAAK,IAAI,CAAC;AAEnH,MAAM,UAAU,GAAG,CAAC,KAAc,EAAqD,EAAE,CAAC,OAAO,KAAK,KAAK,UAAU,CAAC;AAEtH,MAAM,QAAQ,GAAG,CAAC,KAAc,EAAmB,EAAE,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,OAAO,KAAK,CAAC,OAAO,CAAC,KAAK,QAAQ,IAAI,OAAO,KAAK,CAAC,KAAK,CAAC,KAAK,QAAQ,CAAC;AAEhJ,MAAM,OAAO,GAAG,CAAC,KAAc,EAAsB,EAAE,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;AAE5F,kEAAkE;AAClE,MAAM,UAAU,GAAG,CAAC,KAA8B,EAAE,IAAY,EAAE,IAAwB,EAAW,EAAE;IACrG,MAAM,MAAM,GAAY,KAAK,CAAC,IAAI,CAAC,CAAC;IACpC,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC;QAAE,MAAM,IAAI,KAAK,CAAC,qBAAqB,IAAI,WAAW,CAAC,CAAC;IAC/E,OAAO,OAAO,CAAC,KAAK,CAAC,MAAM,EAAE,KAAK,EAAE,IAAI,CAAC,CAAC;AAC5C,CAAC,CAAC;AAEF,MAAM,KAAK,GAAG,GAAW,EAAE;IACzB,MAAM,OAAO,GAAY,OAAO,CAAC,iBAAiB,CAAC,CAAC;IACpD,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC;QAAE,MAAM,IAAI,KAAK,CAAC,oCAAoC,CAAC,CAAC;IAChF,MAAM,MAAM,GAAY,OAAO,EAAE,CAAC;IAClC,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC;QAAE,MAAM,IAAI,KAAK,CAAC,oCAAoC,CAAC,CAAC;IAC7E,OAAO,MAAM,CAAC;AAChB,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,MAAM,GAAqC;IAC/C,EAAE,EAAE,OAAO;IACX,EAAE,EAAE,KAAK;IACT,EAAE,EAAE,KAAK;IACT,EAAE,EAAE,MAAM;IACV,EAAE,EAAE,GAAG;IACP,EAAE,EAAE,KAAK;IACT,EAAE,EAAE,KAAK;IACT,GAAG,EAAE,KAAK;IACV,GAAG,EAAE,KAAK;IACV,EAAE,EAAE,KAAK;IACT,EAAE,EAAE,MAAM;IACV,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,OAAO;IACZ,IAAI,EAAE,OAAO;IACb,GAAG,EAAE,KAAK;IACV,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,IAAI,EAAE,MAAM;IACZ,EAAE,EAAE,KAAK;IACT,GAAG,EAAE,KAAK;IACV,GAAG,EAAE,KAAK;IACV,EAAE,EAAE,KAAK;IACT,GAAG,EAAE,KAAK;IACV,EAAE,EAAE,MAAM;IACV,EAAE,EAAE,MAAM;IACV,EAAE,EAAE,MAAM;IACV,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,KAAK;IACV,EAAE,EAAE,MAAM;IACV,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,KAAK;CACX,CAAC;AAEF,MAAM,WAAW,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAE7E,MAAM,CAAC,MAAM,IAAI,GAAG,CAAC,GAAW,EAAU,EAAE,CAAC,CAAC,WAAW,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,GAAG,CAAC,CAAC,CAAC;AAErG;;;GAGG;AACH;;;;;;GAMG;AACH,MAAM,SAAS,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,EAAE,GAAG,CAAC,CAAC,CAAC;AAE3D,MAAM,EAAE,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,OAAO,CAAC,CAAC,CAAC;AAE9E,MAAM,IAAI,GAAG,CAAC,KAAa,EAAW,EAAE,CAAC,EAAE,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,IAAI,KAAK,CAAC,KAAK,CAAC,WAAW,EAAE,CAAC,IAAI,EAAE,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,WAAW,EAAE,CAAC,CAAC;AAE/H,MAAM,SAAS,GAAG,CAAC,MAAyB,EAAE,EAAU,EAAW,EAAE;IACnE,IAAI,MAAM,CAAC,EAAE,CAAC,EAAE,GAAG,KAAK,KAAK;QAAE,OAAO,KAAK,CAAC;IAC5C,MAAM,MAAM,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,OAAO,EAAE,CAAC;IAC7C,MAAM,IAAI,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,SAAS,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC;IAC/D,OAAO,IAAI,KAAK,SAAS,IAAI,IAAI,CAAC,IAAI,CAAC,CAAC;AAC1C,CAAC,CAAC;AAEF,MAAM,KAAK,GAAkC,EAAE,KAAK,EAAE,SAAS,EAAE,CAAC;AAElE,MAAM,CAAC,MAAM,OAAO,GAAG,GAAS,EAAE;IAChC,KAAK,CAAC,KAAK,KAAK,KAAK,EAAE,CAAC;AAC1B,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,OAAO,GAAG,GAAY,EAAE,CAAC,KAAK,CAAC,KAAK,KAAK,SAAS,CAAC;AAEhE;;;GAGG;AACH,MAAM,MAAM,GAAG,CAAC,IAAY,EAAE,MAAyB,EAAW,EAAE,CAClE,MAAM,CAAC,MAAM,CACX,CAAC,GAAG,EAAE,KAAK,EAAE,EAAE,EAAE,EAAE;IACjB,MAAM,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC;IACpD,IAAI,KAAK,KAAK,CAAC,CAAC;QAAE,OAAO,GAAG,CAAC;IAC7B,MAAM,GAAG,GAAG,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,MAAM,CAAC;IACvC,MAAM,KAAK,GAAG;QACZ,IAAI,EAAE,EAAE,KAAK,EAAE,GAAG,EAAE;QACpB,OAAO,EAAE,KAAK,CAAC,KAAK;QACpB,GAAG,EAAE,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC;QACpB,GAAG,CAAC,KAAK,CAAC,KAAK,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,CAAC;QAC5D,GAAG,CAAC,SAAS,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;KAClE,CAAC;IACF,OAAO,EAAE,MAAM,EAAE,CAAC,GAAG,GAAG,CAAC,MAAM,EAAE,KAAK,CAAC,EAAE,MAAM,EAAE,GAAG,EAAE,CAAC;AACzD,CAAC,EACD,EAAE,MAAM,EAAE,EAAE,EAAE,MAAM,EAAE,CAAC,EAAE,CAC1B,CAAC,MAAM,CAAC;AAEX,MAAM,CAAC,MAAM,QAAQ,GAAG,CAAC,IAAY,EAAuB,EAAE;IAC5D,MAAM,MAAM,GAAG,KAAK,CAAC,KAAK,CAAC;IAC3B,IAAI,MAAM,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IAC3C,OAAO,MAAM,CAAC,IAAI,EAAE,OAAO,CAAC,UAAU,CAAC,MAAM,EAAE,aAAa,EAAE,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAC;AAC3F,CAAC,CAAC"}
|
|
@@ -0,0 +1,34 @@
|
|
|
1
|
+
# Contractions and the long forms that say the same thing.
|
|
2
|
+
#
|
|
3
|
+
# Neither is wrong. The rule only asks whether one document picks one and keeps it.
|
|
4
|
+
id: contraction
|
|
5
|
+
language: en
|
|
6
|
+
entries:
|
|
7
|
+
- pattern: "don't"
|
|
8
|
+
instead_of: do not
|
|
9
|
+
- pattern: "doesn't"
|
|
10
|
+
instead_of: does not
|
|
11
|
+
- pattern: "didn't"
|
|
12
|
+
instead_of: did not
|
|
13
|
+
- pattern: "can't"
|
|
14
|
+
instead_of: cannot
|
|
15
|
+
- pattern: "won't"
|
|
16
|
+
instead_of: will not
|
|
17
|
+
- pattern: "isn't"
|
|
18
|
+
instead_of: is not
|
|
19
|
+
- pattern: "aren't"
|
|
20
|
+
instead_of: are not
|
|
21
|
+
- pattern: "wasn't"
|
|
22
|
+
instead_of: was not
|
|
23
|
+
- pattern: "it's"
|
|
24
|
+
instead_of: it is
|
|
25
|
+
- pattern: "we're"
|
|
26
|
+
instead_of: we are
|
|
27
|
+
- pattern: "they're"
|
|
28
|
+
instead_of: they are
|
|
29
|
+
- pattern: "you're"
|
|
30
|
+
instead_of: you are
|
|
31
|
+
- pattern: "we'll"
|
|
32
|
+
instead_of: we will
|
|
33
|
+
- pattern: "we've"
|
|
34
|
+
instead_of: we have
|
package/package.json
CHANGED
package/src/lexicons.ts
CHANGED
|
@@ -11,7 +11,12 @@ const isRecord = (value: unknown): value is Record<string, unknown> => typeof va
|
|
|
11
11
|
const toEntry = (raw: unknown): LexiconEntry | undefined => {
|
|
12
12
|
if (!isRecord(raw) || typeof raw["pattern"] !== "string") return undefined;
|
|
13
13
|
const weight = raw["weight"];
|
|
14
|
-
|
|
14
|
+
const instead = raw["instead_of"];
|
|
15
|
+
return {
|
|
16
|
+
pattern: raw["pattern"],
|
|
17
|
+
weight: typeof weight === "number" ? weight : undefined,
|
|
18
|
+
instead_of: typeof instead === "string" ? instead : undefined,
|
|
19
|
+
};
|
|
15
20
|
};
|
|
16
21
|
|
|
17
22
|
/**
|
package/src/pos.ts
CHANGED
|
@@ -81,7 +81,14 @@ export const upos = (tag: string): string => (PUNCTUATION.has(tag) ? "PUNCT" : (
|
|
|
81
81
|
* 受動は be + 過去分詞。過去分詞だけでは完了形(has reviewed)と見分けられないので、
|
|
82
82
|
* 直前の be を見る。間に副詞が挟まる(was quickly approved)ぶんだけ遡る。
|
|
83
83
|
*/
|
|
84
|
-
|
|
84
|
+
/**
|
|
85
|
+
* be と過去分詞の間に立てるもの。副詞と、副詞をつなぐ接続詞(was fully and carefully reviewed)。
|
|
86
|
+
*
|
|
87
|
+
* これでも拾えない形がある。`was fully and finally approved` では wink が `approved` を
|
|
88
|
+
* VBN ではなく **VBD** と付けるため、そもそも過去分詞として見えない。解析器の限界で、
|
|
89
|
+
* VBD も受動と見なすと `The team was here and approved it` まで受動になる。直さない。
|
|
90
|
+
*/
|
|
91
|
+
const SKIPPABLE = new Set(["RB", "RBR", "RBS", "CC", ","]);
|
|
85
92
|
|
|
86
93
|
const BE = new Set(["be", "am", "is", "are", "was", "were", "been", "being"]);
|
|
87
94
|
|
|
@@ -90,7 +97,7 @@ const isBe = (entry: Tagged): boolean => BE.has(entry.lemma ?? entry.value.toLow
|
|
|
90
97
|
const isPassive = (tagged: readonly Tagged[], at: number): boolean => {
|
|
91
98
|
if (tagged[at]?.pos !== "VBN") return false;
|
|
92
99
|
const before = tagged.slice(0, at).reverse();
|
|
93
|
-
const head = before.find((entry) => !
|
|
100
|
+
const head = before.find((entry) => !SKIPPABLE.has(entry.pos));
|
|
94
101
|
return head !== undefined && isBe(head);
|
|
95
102
|
};
|
|
96
103
|
|