fix(rule): fix index of match word

textlint-ja · Nov 27, 2015 · 110b303 · 110b303
1 parent 6abeaac
commit 110b303
Show file tree

Hide file tree

Showing 3 changed files with 44 additions and 48 deletions.
diff --git a/package.json b/package.json
@@ -36,11 +36,14 @@
     "espower-babel": "^3.3.0",
     "mocha": "^2.3.0",
     "power-assert": "^1.0.0",
-    "textlint": "^3.2.0",
-    "textlint-tester": "^0.2.0"
+    "textlint": "^5.0.3",
+    "textlint-tester": "^0.4.1"
   },
   "dependencies": {
+    "kuromojin": "^1.0.2",
     "object-assign": "^4.0.1",
+    "sentence-splitter": "^1.2.0",
+    "structured-source": "^3.0.2",
     "textlint-rule-helper": "^1.1.3"
   }
 }
diff --git a/src/max-ten.js b/src/max-ten.js
@@ -2,69 +2,62 @@
 "use strict";
 import {RuleHelper} from "textlint-rule-helper"
 import ObjectAssign from "object-assign"
+import {getTokenizer} from "kuromojin";
+import splitSentences from "sentence-splitter";
+import Source from "structured-source";
 const defaultOptions = {max: 3};
-function countTen(text) {
-    return text.split("、").length - 1;
-}
 /**
  * @param {RuleContext} context
  * @param {object} options
  */
 export default function (context, options = {}) {
     options = ObjectAssign({}, defaultOptions, options);
     const maxLen = options.max;
-    const punctuation = /[。]/;
     let helper = new RuleHelper(context);
     let {Syntax, RuleError, report, getSource} = context;
-    let currentParagraphTexts = [];
     return {
-        [Syntax.Paragraph](){
-            currentParagraphTexts = []
-        },
-        [Syntax.Str](node){
-            // ignore text from external factor
-            if (helper.isChildNode(node, [Syntax.Link, Syntax.Image, Syntax.BlockQuote])) {
+        [Syntax.Paragraph](node){
+            if (helper.isChildNode(node, [Syntax.BlockQuote])) {
                 return;
             }
-            currentParagraphTexts.push(node);
-        },
-        [Syntax.Paragraph + ":exit"](){
-            let currentTenCount = 0;
+            let sentences = splitSentences(getSource(node), {
+                charRegExp: /[。\?\!？！]/,
+                newLineCharacters: "\n\n"
+            });
             /*
             <p>
             <str><code><img><str>
             <str>
             </p>
              */
-            currentParagraphTexts.forEach(strNode => {
-                let paddingLine = 0;
-                let paddingColumn = 0;
-                let text = getSource(strNode);
-                let characters = text.split("");
-                characters.forEach(char => {
-                    if (char === "、") {
-                        currentTenCount++;
-                    }
-                    if (char === "。") {
-                        // reset
-                        currentTenCount = 0;
-                    }
-                    // report
-                    if (currentTenCount >= maxLen) {
-                        var ruleError = new context.RuleError(`一つの文で"、"を${maxLen}つ以上使用しています`, {
-                            line: paddingLine,
-                            column: paddingColumn
-                        });
-                        report(strNode, ruleError);
-                        currentTenCount = 0;
-                    }
-                    // calc padding{line,column}
-                    if (char === "\n") {
-                        paddingLine++;
-                        paddingColumn = 0;
-                    } else {
-                        paddingColumn++;
-                    }
+            return getTokenizer().then(tokenizer => {
+                sentences.forEach(sentence => {
+                    let text = sentence.value;
+                    let source = new Source(text);
+                    let currentTenCount = 0;
+                    let tokens = tokenizer.tokenizeForSentence(text);
+                    let lastToken = null;
+                    tokens.forEach(token => {
+                        let surface = token.surface_form;
+                        if (surface === "、") {
+                            currentTenCount++;
+                            lastToken = token;
+                        }
+                        if (surface === "。") {
+                            // reset
+                            currentTenCount = 0;
+                        }
+                        // report
+                        if (currentTenCount >= maxLen) {
+                            let position = source.indexToPosition(lastToken.word_position - 1);
+                            let ruleError = new context.RuleError(`一つの文で"、"を${maxLen}つ以上使用しています`, {
+                                line: position.line-1,
+                                column:position.column
+                            });
+                            report(node, ruleError);
+                            currentTenCount = 0;
+                        }
+                    });
                 });
             });
         }

diff --git a/test/max-ten-test.js b/test/max-ten-test.js
@@ -1,6 +1,6 @@
 import rule from "../src/max-ten"
 function textIncludeTen(count) {
-    return (new Array(count + 1)).join("テスト、") + "です";
+    return (new Array(count + 1)).join("テスト文章において、") + "です";
 }
 var TextLintTester = require("textlint-tester");
 var tester = new TextLintTester();
@@ -24,7 +24,7 @@ tester.run("max-ten", rule, {
     ],
     invalid: [
         {
-            text: `a、b、       c
+            text: `これは、これは、これは
 、d`
             ,
             errors: [