From 5b0b8c82a8477cfe49a538f267805488daa7f5bd Mon Sep 17 00:00:00 2001
From: lonkaars <loek@pipeframe.xyz>
Date: Mon, 3 Jul 2023 16:38:56 +0200
Subject: more correct sentence parsing

---
 import/jmdict/jmdict.ts | 36 +++++++++++++++++++++++++++---------
 1 file changed, 27 insertions(+), 9 deletions(-)

(limited to 'import')

diff --git a/import/jmdict/jmdict.ts b/import/jmdict/jmdict.ts
index 1f391e5..6109c9b 100644
--- a/import/jmdict/jmdict.ts
+++ b/import/jmdict/jmdict.ts
@@ -1,4 +1,5 @@
 import type { JMdict } from "npm:@scriptin/jmdict-simplified-types";
+import { Tag } from "../../language/tags.ts";
 
 // this script is very messy right now, and doesn't transfer all information
 // present in the dictionary.
@@ -14,15 +15,31 @@ const jmdict = JSON.parse(new TextDecoder().decode(input)) as JMdict;
 
 // TODO: more tags
 const tagLookup = {
-	["misc/uk"]: "aux:uk",
-	["class/adv"]: "class:adverb",
-	["class/vs"]: "class:verb:suru",
-	["class/v1"]: "class:verb:ru",
-	["class/v5"]: "class:verb:u",
-	["class/n"]: "class:noun",
-	["class/suf"]: "class:suffix",
-	["class/prt"]: "class:part",
-	["class/exp"]: "class:expr",
+	["misc/uk"]: Tag.Auxiliary.UsuallyKana,
+	["class/adv"]: Tag.Class.Adverb,
+	["class/vs"]: Tag.Class.Verb.Suru,
+	["class/v1"]: Tag.Class.Verb.Ru,
+	["class/v5"]: Tag.Class.Verb.U,
+	["class/v5k"]: Tag.Class.Verb.U,
+	["class/v5uru"]: Tag.Class.Verb.U,
+	["class/v5r-i"]: Tag.Class.Verb.U,
+	["class/v5u-s"]: Tag.Class.Verb.U,
+	["class/v5aru"]: Tag.Class.Verb.U,
+	["class/v5b"]: Tag.Class.Verb.U,
+	["class/v5g"]: Tag.Class.Verb.U,
+	["class/v5n"]: Tag.Class.Verb.U,
+	["class/v5m"]: Tag.Class.Verb.U,
+	["class/v5r"]: Tag.Class.Verb.U,
+	["class/v5t"]: Tag.Class.Verb.U,
+	["class/v5s"]: Tag.Class.Verb.U,
+	["class/v5u"]: Tag.Class.Verb.U,
+	["class/vk"]: Tag.Class.Verb.Ru, // TODO: this is possibly risky? (should be kuru, but kuru is a ru verb)
+	["class/n"]: Tag.Class.Noun,
+	["class/suf"]: Tag.Class.Suffix,
+	["class/prt"]: Tag.Class.Particle,
+	["class/exp"]: Tag.Class.Expression,
+	["class/adj-i"]: Tag.Class.Adjective.I,
+	["class/adj-na"]: Tag.Class.Adjective.Na,
 } as { [map: string]: string };
 
 console.log(`.param set :dict 'jmdict_${LANG}'`);
@@ -55,6 +72,7 @@ for (let i = 0; i < jmdict.words.length; i++) {
 		...term.sense.map(s => s.partOfSpeech).reduce((acc, current) => [...acc, ...current], []).map(i => `class/${i}`),
 	])];
 	var tags = tags.filter(i => i in tagLookup).map(i => tagLookup[i]);
+	// if (writing == "来る") console.log(term);
 	for (let j = 0; j < definitions.length; j++) {
 		var out = `\t('${writing}', '${reading}', '${tags.join(" ")}', ${j+1}, '${definitions[j].replaceAll("'", "''")}')${(last && j == definitions.length-1) ? ';' : ','}`;
 		if (j == 0 && other_writings.length > 0) out += ` -- TODO: alts: ${other_writings.join(", ")}`;
-- 
cgit v1.2.3