add rule: no middle Tanween

aborazmeh · Jul 9, 2024 · afe4180 · afe4180
1 parent 9a1a294
commit afe4180
Show file tree

Hide file tree

Showing 3 changed files with 48 additions and 1 deletion.
diff --git a/README.md b/README.md
@@ -47,6 +47,10 @@ Shadda can't be combined with Sukun
 
     ضيّْق
 
+### No Middle Tanween
+
+    لا يمكن للتنوين أن يأتي في وسًٍط وسٍط وسٌط الكلمة
+
 ## Usage
 
 Via `.textlintrc.json`(Recommended)
@@ -90,7 +94,6 @@ Test textlint rule by [textlint-tester](https://github.com/textlint/textlint-tes
 
 - Normalize diacritics forms before some regexes like for duplicated diacritics
 - No diacritics or Tanween on Alef without Hamza except Tanween Al-Nasb
-- No middle Tanween
 - No consecutive five Harakat 
 - No consecutive three Sokoon
 - Twneen on Alef or the letter before

diff --git a/src/index.ts b/src/index.ts
@@ -6,10 +6,12 @@ export interface Options {
     no_shadda_with_madda?: boolean;
     no_shadda_with_sukun?: boolean;
     no_duplicated_diacritics?: boolean;
+    no_middle_tanween?: boolean
 }
 
 const regex = {
     diacritics: "[\u064B-\u0653]",
+    tanween: "[\u064B-\u064D]",
     shadda: "[\u0651\u0AFB\uFC5E-\uFC63\uFCF2-\uFCF4\uFE7C\uFE7D\u11237]",
     madda: "\u0653",
     alefMadda: "[\u0622\uFE81\uFE82\uFEF5\uFEF6]",
@@ -100,6 +102,21 @@ function noDuplicatedDiacritics(node: TxtStrNode, text: string, context: Readonl
     }
 }
 
+function noMiddleTanween(node: TxtStrNode, text: string, context: Readonly<TextlintRuleContext>) {
+    const { report, locator, RuleError } = context;
+
+    // FIXME dont't report if the next character is Alef and it's the last letter of the word
+    const matches = text.matchAll(new RegExp(`(${regex.tanween})(?=[^\\s]*\\p{Letter})`, "ug"));
+    for (const match of matches) {
+        const index = match.index ?? 0;
+        const matchRange = [index, index + match[0].length] as const;
+        const ruleError = new RuleError("Found middle Tanween.", {
+            padding: locator.range(matchRange)
+        });
+        report(node, ruleError);
+    }
+}
+
 const report: TextlintRuleModule<Options> = (context, options = {}) => {
     const { getSource, Syntax } = context;
     return {
@@ -108,6 +125,7 @@ const report: TextlintRuleModule<Options> = (context, options = {}) => {
             const shaddaWithMaddaOpt = options.no_shadda_with_madda ?? true;
             const shaddaWithSukunOpt = options.no_shadda_with_sukun ?? true;
             const duplicatedDiacriticsOpt = options.no_duplicated_diacritics ?? true;
+            const noMiddleTanweenOpt = options.no_middle_tanween ?? true;
 
             const text = getSource(node); // Get text
             noLooseDiacritics(node, text, context, removeLooseDiacritics);
@@ -123,6 +141,10 @@ const report: TextlintRuleModule<Options> = (context, options = {}) => {
             if (duplicatedDiacriticsOpt) {
                 noDuplicatedDiacritics(node, text, context);
             }
+
+            if (noMiddleTanweenOpt) {
+                noMiddleTanween(node, text, context);
+            }
         }
     };
 };

diff --git a/test/index-test.ts b/test/index-test.ts
@@ -9,6 +9,7 @@ tester.run("rule", rule, {
         "الآن",
         "ضيّق",
         "يونَُِس: قال أبو عبيدة، «يقال:يونس بضم النون وكسرها». والمشهور في القراءة يونُس برفع النون من غير همز.",
+        "من علامات الاسم: التنوين آخر الاسم المنصرف؛ السماء صافيةٌ، رأيت عامراً، مررت بدكانٍ",
         {
             text: "ضيّْق",
             options: {
@@ -113,6 +114,27 @@ tester.run("rule", rule, {
                     range: [9, 12]
                 }
             ]
+        },
+        {
+            text: "لا يمكن للتنوين أن يأتي في وسًٍط وسٍط وسٌط الكلمة",
+            errors: [
+                {
+                    message: "Found middle Tanween.",
+                    range: [29, 30]
+                },
+                {
+                    message: "Found middle Tanween.",
+                    range: [30, 31]
+                },
+                {
+                    message: "Found middle Tanween.",
+                    range: [35, 36]
+                },
+                {
+                    message: "Found middle Tanween.",
+                    range: [40, 41]
+                }
+            ]
         }
     ]
 });