feat(options): support exclusionPatterns options

Author: azu

README.md

@@ -25,6 +25,10 @@ Add "sentence-length" to your `.textlintrc`.
     - default: 100
     - The total number of characters allowed on each sentences.
     - Sentence.length > 100 and throw Error
+- `exclusionPatterns`: `string[]`
+    - A strings that match the patterns is uncount of the sentence.
+    - Set an array of RegExp-like string.
+    - See https://github.com/textlint/regexp-string-matcher
 
 ```
 {
@@ -36,6 +40,22 @@ Add "sentence-length" to your `.textlintrc`.
 }
 ```
 
+Uncount `(...)` from `A sentence(...).`
+
+```
+{
+    "rules": {
+        "sentence-length": {
+            "max": 100,
+            exclusionPatterns: [
+                "/\\(.*\\)$\\./"
+            ]
+
+        }
+    }
+}
+```
+
 ## Exception
 
 - Except BlockQuote

package.json

@@ -32,6 +32,7 @@
     "prettier": "prettier --write **/*.js"
   },
   "dependencies": {
+    "@textlint/regexp-string-matcher": "^1.1.0",
     "sentence-splitter": "^3.0.11",
     "textlint-rule-helper": "^2.1.1",
     "textlint-util-to-string": "^3.0.0"

src/sentence-length.js

@@ -3,12 +3,29 @@
 import { splitAST, Syntax as SentenceSyntax } from "sentence-splitter";
 import { StringSource } from "textlint-util-to-string";
 import { RuleHelper } from "textlint-rule-helper";
+import { createRegExp } from "@textlint/regexp-string-matcher";
+
+
+function removeRangeFromString(string, regExpStrings) {
+    const patterns = regExpStrings.map(pattern => {
+        return createRegExp(pattern);
+    });
+    let result = string;
+    patterns.forEach(pattern => {
+        result = result.replace(pattern, "");
+    });
+    return result;
+}
 
 const defaultOptions = {
-    max: 100
+    max: 100,
+    // The strings that match following patterns is uncount of the sentence
+    // See https://github.com/textlint/regexp-string-matcher
+    exclusionPatterns: []
 };
 module.exports = function(context, options = {}) {
     const maxLength = options.max || defaultOptions.max;
+    const exclusionPatterns = options.exclusionPatterns || defaultOptions.exclusionPatterns;
     const helper = new RuleHelper(context);
     const { Syntax, RuleError, report } = context;
     // toPlainText
@@ -26,14 +43,20 @@ module.exports = function(context, options = {}) {
             const paragraph = splitAST(node);
             paragraph.children.filter(sentence => sentence.type === SentenceSyntax.Sentence).forEach(sentence => {
                 const source = new StringSource(sentence);
-                const sentenceText = source.toString();
+                const actualText = source.toString();
+                const sentenceText = removeRangeFromString(actualText, exclusionPatterns);
                 // larger than > 100
+                const actualTextLength = actualText.length;
                 const sentenceLength = sentenceText.length;
                 if (sentenceLength > maxLength) {
                     const startLine = sentence.loc.start.line;
                     report(
                         sentence,
-                        new RuleError(`Line ${startLine} sentence length(${sentenceLength}) exceeds the maximum sentence length of ${maxLength}.
+                        new RuleError(`Line ${startLine} sentence length(${
+                            sentenceLength !== actualTextLength
+                                ? `${sentenceLength}, original:${actualTextLength}`
+                                : sentenceLength
+                            }) exceeds the maximum sentence length of ${maxLength}.
 Over ${sentenceLength - maxLength} characters.`)
                     );
                 }

test/sentence-length-test.js

@@ -12,7 +12,7 @@ tester.run("textlint-rule-sentence-length", rule, {
         "This is a article",
         "Test`code`です。",
         {
-            text: '"This" is code.',
+            text: "\"This\" is code.",
             options: {
                 max: 15
             }
@@ -35,22 +35,32 @@ tester.run("textlint-rule-sentence-length", rule, {
         },
         {
             // == 12345
-            text: '[123](http://example.com "123456")45',
+            text: "[123](http://example.com \"123456\")45",
             options: {
                 max: 5
             }
         },
+        {
+            // ignore /()$/。
+            text: "1234(56789)",
+            options: {
+                max: 5,
+                exclusionPatterns: [
+                    "/\\(.*\\)$/"
+                ]
+            }
+        },
         {
             // html node
             // == 12345
-            text: '<s>123</s><b>45</b>',
+            text: "<s>123</s><b>45</b>",
             options: {
                 max: 5
             }
         },
         {
             // List
-            text: '- [abc](http://example.com "abc")de',
+            text: "- [abc](http://example.com \"abc\")de",
             options: {
                 max: 5
             }
@@ -136,7 +146,7 @@ Over 1 characters.`
             errors: [
                 {
                     message: `Line 1 sentence length(18) exceeds the maximum sentence length of 5.
-Over 13 characters.`,
+Over 13 characters.`
                 }
             ]
         },
@@ -190,6 +200,20 @@ Over 2 characters.`,
                     column: 1
                 }
             ]
+        },
+        {
+            // ignore /()$/。
+            text: "123456789(56789)",
+            options: {
+                max: 5,
+                exclusionPatterns: [
+                    "/\\(.*\\)$/"
+                ]
+            },
+            errors: [{
+                message: "Line 1 sentence length(9, original:16) exceeds the maximum sentence length of 5.\n" +
+                    "Over 4 characters."
+            }]
         }
     ]
 });

yarn.lock

@@ -798,6 +798,18 @@
   resolved "https://registry.yarnpkg.com/@textlint/module-interop/-/module-interop-1.0.1.tgz#765ca9ccb9b66657d65061d3546b3fe51e48c983"
   integrity sha512-gqx1Te+lMnXX6xyGTUdzGhm8RT7IfiSRMtWoH1FeTMg2InArRT+lTksCFc/x5dtaPN4vwOFZUvU8oTzYQzXbyg==
 
+"@textlint/regexp-string-matcher@^1.1.0":
+  version "1.1.0"
+  resolved "https://registry.yarnpkg.com/@textlint/regexp-string-matcher/-/regexp-string-matcher-1.1.0.tgz#e19975029ce228a214d50c6a7e9dbcbef29ad8cd"
+  integrity sha512-uTPnE1Dw1j+9clXPn61ZUdtg+WyhbgeXHwCTfBev7quHjeCP9PS8NdRkR6wEgmjuLg+xZlI4r/e1r6Bd0xyusQ==
+  dependencies:
+    escape-string-regexp "^1.0.5"
+    execall "^1.0.0"
+    lodash.sortby "^4.7.0"
+    lodash.uniq "^4.5.0"
+    lodash.uniqwith "^4.5.0"
+    to-regex "^3.0.2"
+
 "@textlint/text-to-ast@^3.1.5":
   version "3.1.5"
   resolved "https://registry.yarnpkg.com/@textlint/text-to-ast/-/text-to-ast-3.1.5.tgz#927d971712ac24d6056a0a178cc154c177427e4c"
@@ -1224,6 +1236,14 @@ cli-truncate@^0.2.1:
     slice-ansi "0.0.4"
     string-width "^1.0.1"
 
+clone-regexp@^1.0.0:
+  version "1.0.1"
+  resolved "https://registry.yarnpkg.com/clone-regexp/-/clone-regexp-1.0.1.tgz#051805cd33173375d82118fc0918606da39fd60f"
+  integrity sha512-Fcij9IwRW27XedRIJnSOEupS7RVcXtObJXbcUOX93UCLqqOdRpkvzKywOOSizmEK/Is3S/RHX9dLdfo6R1Q1mw==
+  dependencies:
+    is-regexp "^1.0.0"
+    is-supported-regexp-flag "^1.0.0"
+
 [email protected]:
   version "3.1.0"
   resolved "https://registry.yarnpkg.com/co/-/co-3.1.0.tgz#4ea54ea5a08938153185e15210c68d9092bc1b78"
@@ -1587,6 +1607,13 @@ execa@^2.0.3:
     signal-exit "^3.0.2"
     strip-final-newline "^2.0.0"
 
+execall@^1.0.0:
+  version "1.0.0"
+  resolved "https://registry.yarnpkg.com/execall/-/execall-1.0.0.tgz#73d0904e395b3cab0658b08d09ec25307f29bb73"
+  integrity sha1-c9CQTjlbPKsGWLCNCewlMH8pu3M=
+  dependencies:
+    clone-regexp "^1.0.0"
+
 expand-brackets@^2.1.4:
   version "2.1.4"
   resolved "https://registry.yarnpkg.com/expand-brackets/-/expand-brackets-2.1.4.tgz#b77735e315ce30f6b6eff0f83b04151a22449622"
@@ -2281,6 +2308,11 @@ is-stream@^2.0.0:
   resolved "https://registry.yarnpkg.com/is-stream/-/is-stream-2.0.0.tgz#bde9c32680d6fae04129d6ac9d921ce7815f78e3"
   integrity sha512-XCoy+WlUr7d1+Z8GgSuXmpuUFC9fOhRXglJMx+dwLKTkL44Cjd4W1Z5P+BQZpr+cR93aGP4S/s7Ftw6Nd/kiEw==
 
+is-supported-regexp-flag@^1.0.0:
+  version "1.0.1"
+  resolved "https://registry.yarnpkg.com/is-supported-regexp-flag/-/is-supported-regexp-flag-1.0.1.tgz#21ee16518d2c1dd3edd3e9a0d57e50207ac364ca"
+  integrity sha512-3vcJecUUrpgCqc/ca0aWeNu64UGgxcvO60K/Fkr1N6RSvfGCTU60UKN68JDmKokgba0rFFJs12EnzOQa14ubKQ==
+
 is-symbol@^1.0.1:
   version "1.0.1"
   resolved "https://registry.yarnpkg.com/is-symbol/-/is-symbol-1.0.1.tgz#3cc59f00025194b6ab2e38dbae6689256b660572"
@@ -2502,6 +2534,21 @@ locate-path@^3.0.0:
     p-locate "^3.0.0"
     path-exists "^3.0.0"
 
+lodash.sortby@^4.7.0:
+  version "4.7.0"
+  resolved "https://registry.yarnpkg.com/lodash.sortby/-/lodash.sortby-4.7.0.tgz#edd14c824e2cc9c1e0b0a1b42bb5210516a42438"
+  integrity sha1-7dFMgk4sycHgsKG0K7UhBRakJDg=
+
+lodash.uniq@^4.5.0:
+  version "4.5.0"
+  resolved "https://registry.yarnpkg.com/lodash.uniq/-/lodash.uniq-4.5.0.tgz#d0225373aeb652adc1bc82e4945339a842754773"
+  integrity sha1-0CJTc662Uq3BvILklFM5qEJ1R3M=
+
+lodash.uniqwith@^4.5.0:
+  version "4.5.0"
+  resolved "https://registry.yarnpkg.com/lodash.uniqwith/-/lodash.uniqwith-4.5.0.tgz#7a0cbf65f43b5928625a9d4d0dc54b18cadc7ef3"
+  integrity sha1-egy/ZfQ7WShiWp1NDcVLGMrcfvM=
+
 lodash@^4.0.0:
   version "4.17.4"
   resolved "https://registry.yarnpkg.com/lodash/-/lodash-4.17.4.tgz#78203a4d1c328ae1d86dca6460e369b57f4055ae"