graphemesplit - npm Package Compare versions

Comparing version 2.0.1 to 2.0.3

generate.js

@@ -1,9 +0,9 @@
-const fs = require('fs')
-const https = require('https')
-const stream = require('stream')
+const fs = require("fs");
+const https = require("https");
+const stream = require("stream");
 
-const linesStream = require('@orisano/lines-stream')
-const UnicodeTrieBuilder = require('unicode-trie/builder')
+const linesStream = require("@orisano/lines-stream");
+const UnicodeTrieBuilder = require("unicode-trie/builder");
 
-const types = require('./types')
+const types = require("./types");
 
@@ -16,63 +16,73 @@ function parseLine() {
     transform(line, encoding, callback) {
-      const body = line.split('#')[0]
+      const body = line.split("#")[0];
       if (body.trim().length === 0) {
-        callback()
-        return
+        callback();
+        return;
       }
-      const [rawRange, type] = body.split(';').map(x => x.trim())
-      const range = rawRange.split('..').map(x => parseInt(x, 16))
+      const [rawRange, type] = body.split(";").map(x => x.trim());
+      const range = rawRange.split("..").map(x => parseInt(x, 16));
       if (range.length > 1) {
-        this.push({start: range[0], end: range[1], type})
+        this.push({ start: range[0], end: range[1], type });
       } else {
-        this.push({start: range[0], end: range[0], type})
+        this.push({ start: range[0], end: range[0], type });
       }
-      callback()
-    },
-  })
+      callback();
+    }
+  });
 }
 
-https.get('https://www.unicode.org/Public/UCD/latest/ucd/auxiliary/GraphemeBreakProperty.txt', res => {
-  const {statusCode} = res
-  if (statusCode !== 200) {
-    console.error(`failed to request: ${statusCode}`)
-    res.resume()
-    return
+https.get(
+  "https://www.unicode.org/Public/UCD/latest/ucd/auxiliary/GraphemeBreakProperty.txt",
+  res => {
+    const { statusCode } = res;
+    if (statusCode !== 200) {
+      console.error(`failed to request: ${statusCode}`);
+      res.resume();
+      return;
+    }
+
+    const trie = new UnicodeTrieBuilder(types.Other);
+    res
+      .setEncoding("utf8")
+      .pipe(linesStream())
+      .pipe(parseLine())
+      .on("data", ({ start, end, type }) => {
+        trie.setRange(start, end, types[type]);
+      })
+      .on("end", () => {
+        fs.writeFileSync(
+          "./typeTrie.json",
+          JSON.stringify({
+            data: trie.toBuffer().toString("base64")
+          })
+        );
+      });
   }
+);
 
-  const trie = new UnicodeTrieBuilder(types.Other)
-  res
-  .setEncoding('utf8')
-  .pipe(linesStream())
-  .pipe(parseLine())
-  .on('data', ({start, end, type}) => {
-    trie.setRange(start, end, types[type])
-  })
-  .on('end', () => {
-    fs.writeFileSync('./typeTrie.json', JSON.stringify({
-      data: trie.toBuffer().toString('base64'),
-    }))
-  })
-})
-
-https.get('https://unicode.org/Public/emoji/11.0/emoji-data.txt', res => {
-  const {statusCode} = res
+https.get("https://unicode.org/Public/emoji/11.0/emoji-data.txt", res => {
+  const { statusCode } = res;
   if (statusCode !== 200) {
-    console.error(`failed to request: ${statusCode}`)
-    res.resume()
-    return
+    console.error(`failed to request: ${statusCode}`);
+    res.resume();
+    return;
   }
 
-  const trie = new UnicodeTrieBuilder()
+  const trie = new UnicodeTrieBuilder();
   res
-  .setEncoding('utf8')
-  .pipe(linesStream())
-  .pipe(parseLine())
-  .on('data', ({start, end, type}) => {
-    if (type === 'Extended_Pictographic') trie.setRange(start, end, types.Extended_Pictographic)
-  })
-  .on('end', () => {
-    fs.writeFileSync('./extPict.json', JSON.stringify({
-      data: trie.toBuffer().toString('base64'),
-    }))
-  })
-})
+    .setEncoding("utf8")
+    .pipe(linesStream())
+    .pipe(parseLine())
+    .on("data", ({ start, end, type }) => {
+      if (type === "Extended_Pictographic")
+        trie.setRange(start, end, types.Extended_Pictographic);
+    })
+    .on("end", () => {
+      fs.writeFileSync(
+        "./extPict.json",
+        JSON.stringify({
+          data: trie.toBuffer().toString("base64")
+        })
+      );
+    });
+});

index.js

@@ -1,12 +0,12 @@
-const types = require('./types')
-const typeTrieData = require('./typeTrie').data
-const extPictData = require('./extPict').data
+const types = require("./types");
+const typeTrieData = require("./typeTrie").data;
+const extPictData = require("./extPict").data;
 
-const UnicodeTrie = require('unicode-trie')
+const UnicodeTrie = require("unicode-trie");
 
-const typeTrie = new UnicodeTrie(Buffer.from(typeTrieData, 'base64'))
-const extPict = new UnicodeTrie(Buffer.from(extPictData, 'base64'))
+const typeTrie = new UnicodeTrie(Buffer.from(typeTrieData, "base64"));
+const extPict = new UnicodeTrie(Buffer.from(extPictData, "base64"));
 
 function is(type, bit) {
-  return (type & bit) !== 0
+  return (type & bit) !== 0;
 }
@@ -17,9 +17,9 @@
   ExtendOrZWJ: 1,
-  NotBoundary: 2,
-}
+  NotBoundary: 2
+};
 
 function nextGraphemeClusterSize(ts, start) {
-  const L = ts.length
+  const L = ts.length;
 
-  let ri = 0
+  let ri = 0;
   let gb11State = GB11State.Initial;
@@ -29,8 +29,8 @@
   for (let i = start; i + 1 < L; i++) {
-    const curr = ts[i + 0]
-    const next = ts[i + 1]
+    const curr = ts[i + 0];
+    const next = ts[i + 1];
 
     // for GB12, GB13
     if (!is(curr, types.Regional_Indicator)) {
-      ri = 0
+      ri = 0;
     }
@@ -43,16 +43,19 @@
         if (is(curr, types.Extended_Pictographic)) {
-          gb11State = GB11State.ExtendOrZWJ
+          gb11State = GB11State.ExtendOrZWJ;
         } else {
-          gb11State = GB11State.Initial
+          gb11State = GB11State.Initial;
         }
-        break
+        break;
       case GB11State.ExtendOrZWJ:
         if (is(curr, types.Extend)) {
-          gb11State = GB11State.ExtendOrZWJ
-        } else if (is(curr, types.ZWJ) && is(next, types.Extended_Pictographic)) {
-          gb11State = GB11State.NotBoundary
+          gb11State = GB11State.ExtendOrZWJ;
+        } else if (
+          is(curr, types.ZWJ) &&
+          is(next, types.Extended_Pictographic)
+        ) {
+          gb11State = GB11State.NotBoundary;
         } else {
-          gb11State = GB11State.Initial
+          gb11State = GB11State.Initial;
         }
-        break
+        break;
     }
@@ -62,65 +65,81 @@
     if (is(curr, types.CR) && is(next, types.LF)) {
-      continue
+      continue;
     }
     // GB4: (Control | CR | LF) ÷
     if (is(curr, types.Control | types.CR | types.LF)) {
-      return (i + 1) - start
+      return i + 1 - start;
     }
     // GB5: ÷ (Control | CR | LF)
     if (is(next, types.Control | types.CR | types.LF)) {
-      return (i + 1) - start
+      return i + 1 - start;
     }
     // GB6: L x (L | V | LV | LVT)
-    if (is(curr, types.L) && is(next, types.L | types.V | types.LV | types.LVT)) {
-      continue
+    if (
+      is(curr, types.L) &&
+      is(next, types.L | types.V | types.LV | types.LVT)
+    ) {
+      continue;
     }
     // GB7: (LV | V) x (V | T)
     if (is(curr, types.LV | types.V) && is(next, types.V | types.T)) {
-      continue
+      continue;
     }
     // GB8: (LVT | T) x T
     if (is(curr, types.LVT | types.T) && is(next, types.T)) {
-      continue
+      continue;
     }
     // GB9: x (Extend | ZWJ)
     if (is(next, types.Extend | types.ZWJ)) {
-      continue
+      continue;
     }
     // GB9a: x SpacingMark
     if (is(next, types.SpacingMark)) {
-      continue
+      continue;
     }
     // GB9b: Prepend x
     if (is(curr, types.Prepend)) {
-      continue
+      continue;
     }
     // GB11: \p{Extended_Pictographic} Extend* ZWJ x \p{Extended_Pictographic}
     if (gb11State === GB11State.NotBoundary) {
-      continue
+      continue;
     }
     // GB12: sot (RI RI)* RI x RI
     // GB13: [^RI] (RI RI)* RI x RI
-    if (is(curr, types.Regional_Indicator) && is(next, types.Regional_Indicator) && ri % 2 === 0) {
-      ri++
-      continue
+    if (
+      is(curr, types.Regional_Indicator) &&
+      is(next, types.Regional_Indicator) &&
+      ri % 2 === 0
+    ) {
+      ri++;
+      continue;
     }
     // GB999: Any ÷ Any
-    return (i + 1) - start
+    return i + 1 - start;
   }
   // GB2: Any ÷ eot
-  return L - start
+  return L - start;
 }
 
 module.exports = function split(str) {
-  const graphemeClusters = []
+  const graphemeClusters = [];
 
-  const codePoints = [...str].map(x => x.codePointAt(0))
-  const ts = codePoints.map(c => typeTrie.get(c) | extPict.get(c))
-  for (let offset = 0; offset < codePoints.length;) {
-    const size = nextGraphemeClusterSize(ts, offset)
-    const graphemeCluster = codePoints.slice(offset, offset + size)
-    graphemeClusters.push(String.fromCodePoint(...graphemeCluster))
-    offset += size
+  const map = [0];
+  const ts = [];
+  for (let i = 0; i < str.length; ) {
+    const code = str.codePointAt(i);
+    ts.push(typeTrie.get(code) | extPict.get(code));
+    i += code > 65535 ? 2 : 1;
+    map.push(i);
   }
-  return graphemeClusters
-}
+
+  for (let offset = 0; offset < ts.length; ) {
+    const size = nextGraphemeClusterSize(ts, offset);
+    const start = map[offset];
+    const end = map[offset + size];
+    graphemeClusters.push(str.slice(start, end));
+    offset += size;
+  }
+
+  return graphemeClusters;
+};

package.json

 {
   "name": "graphemesplit",
-  "version": "2.0.1",
+  "version": "2.0.3",
   "main": "index.js",
@@ -8,4 +8,8 @@ "author": "Nao YONASHIRO",
   "repository": "github:orisano/graphemesplit",
+  "scripts": {
+      "fmt": "prettier --write **/*.js"
+  },
   "devDependencies": {
-    "@orisano/lines-stream": "^0.1.2"
+    "@orisano/lines-stream": "^0.1.2",
+    "prettier": "^1.16.4"
   },
@@ -12,0 +16,0 @@ "dependencies": {

test.js

@@ -1,44 +0,54 @@
-const assert = require('assert')
-const https = require('https')
-const stream = require('stream')
+const assert = require("assert");
+const https = require("https");
+const stream = require("stream");
 
-const linesStream = require('@orisano/lines-stream')
+const linesStream = require("@orisano/lines-stream");
 
-const graphemesplit = require('./')
+const graphemesplit = require("./");
 
-https.get('https://www.unicode.org/Public/UNIDATA/auxiliary/GraphemeBreakTest.txt', res => {
-  const {statusCode} = res
-  if (statusCode !== 200) {
-    console.error(`failed to http request: ${statusCode}`)
-    res.resume()
-    return
-  }
-  res
-  .pipe(linesStream())
-  .pipe(new stream.Transform({
-    decodeStrings: false,
-    readableObjectMode: true,
+https.get(
+  "https://www.unicode.org/Public/UNIDATA/auxiliary/GraphemeBreakTest.txt",
+  res => {
+    const { statusCode } = res;
+    if (statusCode !== 200) {
+      console.error(`failed to http request: ${statusCode}`);
+      res.resume();
+      return;
+    }
+    res
+      .pipe(linesStream())
+      .pipe(
+        new stream.Transform({
+          decodeStrings: false,
+          readableObjectMode: true,
 
-    transform(line, encoding, callback) {
-      callback()
+          transform(line, encoding, callback) {
+            callback();
 
-      if (line.trim().length === 0) return
-      const [body, description] = line.split('#')
-      const test = body.trim()
-      if (test.length === 0) return
-      const graphemeClusters = test
-      .split('÷')
-      .filter(x => x.length > 0)
-      .map(x => {
-        const codePoints = x.split('×')
-        .map(y => parseInt(y.trim(), 16))
-        return String.fromCodePoint(...codePoints)
-      })
-      this.push({expected: graphemeClusters, description})
-    }
-  }))
-  .on('data', ({expected, description}) => {
-    const got = graphemesplit(expected.join(''))
-    assert.deepStrictEqual(got, expected, `unexpected grapheme clusters. expected: ${expected}, but got: ${got} # ${description}`)
-  })
-})
+            if (line.trim().length === 0) return;
+            const [body, description] = line.split("#");
+            const test = body.trim();
+            if (test.length === 0) return;
+            const graphemeClusters = test
+              .split("÷")
+              .filter(x => x.length > 0)
+              .map(x => {
+                const codePoints = x
+                  .split("×")
+                  .map(y => parseInt(y.trim(), 16));
+                return String.fromCodePoint(...codePoints);
+              });
+            this.push({ expected: graphemeClusters, description });
+          }
+        })
+      )
+      .on("data", ({ expected, description }) => {
+        const got = graphemesplit(expected.join(""));
+        assert.deepStrictEqual(
+          got,
+          expected,
+          `unexpected grapheme clusters. expected: ${expected}, but got: ${got} # ${description}`
+        );
+      });
+  }
+);

types.js

@@ -16,3 +16,3 @@ module.exports = {
   LVT: 1 << 12,
-  Extended_Pictographic: 1 << 13,
-}
+  Extended_Pictographic: 1 << 13
+};

New alerts

License Policy Violation

License

This package is not allowed per your license policy. Review the package's license to ensure compliance.

Found 1 instance in 1 package

Major refactor

Supply chain risk

Package has recently undergone a major refactor. It may be unstable or indicate significant internal changes. Use caution when updating to versions that include significant changes.

Found 1 instance in 1 package

Network access

Supply chain risk

This module accesses the network.

Found 1 instance in 1 package

Debug access

Supply chain risk

Uses debug, reflection and dynamic code execution features.

Found 1 instance in 1 package

Filesystem access

Supply chain risk

Accesses the file system, and could potentially read sensitive data.

Found 1 instance in 1 package

Fixed alerts

License Policy Violation

License

This package is not allowed per your license policy. Review the package's license to ensure compliance.

Found 1 instance in 1 package

Network access

Supply chain risk

This module accesses the network.

Found 1 instance in 1 package

Filesystem access

Supply chain risk

Accesses the file system, and could potentially read sensitive data.

Found 1 instance in 1 package

Improved metrics

Lines of code

294

increased by26.18%

Worsened metrics

Total package byte prevSize

14216

decreased by-49.61%

Dev dependency count

2

increased by100%

Number of package files

10

decreased by-41.18%

Number of low supply chain risk alerts

3

increased by50%

Number of medium supply chain risk alerts

3

increased by50%

No dependency changes