wip: improve node api

packages/check-html-links/CHANGELOG.md

@@ -1,11 +1,5 @@
 # check-html-links
 
-## 0.1.2
-
-### Patch Changes
-
-- f343c50: When reading bigger files, especially bigger files with all content on one line it could mean a read chunk is in the middle of a character. This can lead to strange symbols in the resulting string. The `hightWaterMark` is now increased from the node default of 16KB to 256KB. Additionally, the `hightWaterMark` is now synced for reading and parsing.
-
 ## 0.1.1
 
 ### Patch Changes

packages/check-html-links/index.js

@@ -1,2 +1,3 @@
 export { validateFolder } from './src/validateFolder.js';
 export { formatErrors } from './src/formatErrors.js';
+export { checkHtmlLinks } from './src/checkHtmlLinks.js';

packages/check-html-links/package.json

@@ -1,6 +1,6 @@
 {
   "name": "check-html-links",
-  "version": "0.1.2",
+  "version": "0.1.1",
   "publishConfig": {
     "access": "public"
   },

packages/check-html-links/src/checkHtmlLinks.js

@@ -0,0 +1,50 @@
+#!/usr/bin/env node
+
+import chalk from 'chalk';
+import { validateFiles } from './validateFolder.js';
+import { formatErrors } from './formatErrors.js';
+import { listFiles } from './listFiles.js';
+
+export async function checkHtmlLinks({ rootDir = process.cwd() }) {
+  const performanceStart = process.hrtime();
+
+  console.log('👀 Checking if all internal links work...');
+  const files = await listFiles('**/*.html', rootDir);
+
+  const filesOutput =
+    files.length == 0
+      ? '🧐 No files to check. Did you select the correct folder?'
+      : `🔥 Found a total of ${chalk.green.bold(files.length)} files to check!`;
+  console.log(filesOutput);
+
+  const { errors, numberLinks } = await validateFiles(files, rootDir);
+
+  console.log(`🔗 Found a total of ${chalk.green.bold(numberLinks)} links to validate!\n`);
+
+  const performance = process.hrtime(performanceStart);
+  if (errors.length > 0) {
+    let referenceCount = 0;
+    for (const error of errors) {
+      referenceCount += error.usage.length;
+    }
+    const output = [
+      `❌ Found ${chalk.red.bold(
+        errors.length.toString(),
+      )} missing reference targets (used by ${referenceCount} links) while checking ${
+        files.length
+      } files:`,
+      ...formatErrors(errors)
+        .split('\n')
+        .map(line => `  ${line}`),
+      `Checking links duration: ${performance[0]}s ${performance[1] / 1000000}ms`,
+    ];
+    console.error(output.join('\n'));
+    process.exit(1);
+  } else {
+    console.log(
+      `✅ All internal links are valid. (executed in ${performance[0]}s ${
+        performance[1] / 1000000
+      }ms)`,
+    );
+  }
+}

packages/check-html-links/src/cli.js

@@ -1,55 +1,17 @@
 #!/usr/bin/env node
 
 import path from 'path';
-import chalk from 'chalk';
-import { validateFiles } from './validateFolder.js';
-import { formatErrors } from './formatErrors.js';
-import { listFiles } from './listFiles.js';
+import { checkHtmlLinks } from 'check-html-links';
 
-async function main() {
+async function cli() {
   const userRootDir = process.argv[2];
   const rootDir = userRootDir ? path.resolve(userRootDir) : process.cwd();
-  const performanceStart = process.hrtime();
 
-  console.log('👀 Checking if all internal links work...');
-  const files = await listFiles('**/*.html', rootDir);
-
-  const filesOutput =
-    files.length == 0
-      ? '🧐 No files to check. Did you select the correct folder?'
-      : `🔥 Found a total of ${chalk.green.bold(files.length)} files to check!`;
-  console.log(filesOutput);
-
-  const { errors, numberLinks } = await validateFiles(files, rootDir);
-
-  console.log(`🔗 Found a total of ${chalk.green.bold(numberLinks)} links to validate!\n`);
-
-  const performance = process.hrtime(performanceStart);
-  if (errors.length > 0) {
-    let referenceCount = 0;
-    for (const error of errors) {
-      referenceCount += error.usage.length;
-    }
-    const output = [
-      `❌ Found ${chalk.red.bold(
-        errors.length.toString(),
-      )} missing reference targets (used by ${referenceCount} links) while checking ${
-        files.length
-      } files:`,
-      ...formatErrors(errors)
-        .split('\n')
-        .map(line => `  ${line}`),
-      `Checking links duration: ${performance[0]}s ${performance[1] / 1000000}ms`,
-    ];
-    console.error(output.join('\n'));
-    process.exit(1);
-  } else {
-    console.log(
-      `✅ All internal links are valid. (executed in %ds %dms)`,
-      performance[0],
-      performance[1] / 1000000,
-    );
+  try {
+    await checkHtmlLinks({ rootDir });
+  } catch (error) {
+    console.log('Check Html Links CLI failed to execute', error);
   }
 }
 
-main();
+cli();

packages/check-html-links/src/validateFolder.js

@@ -15,12 +15,10 @@ import path from 'path';
 const require = createRequire(import.meta.url);
 const { SaxEventType, SAXParser } = saxWasm;
 
-const streamOptions = { highWaterMark: 256 * 1024 };
-
 const saxPath = require.resolve('sax-wasm/lib/sax-wasm.wasm');
 const saxWasmBuffer = fs.readFileSync(saxPath);
-const parserReferences = new SAXParser(SaxEventType.Attribute, streamOptions);
-const parserIds = new SAXParser(SaxEventType.Attribute, streamOptions);
+const parserReferences = new SAXParser(SaxEventType.Attribute);
+const parserIds = new SAXParser(SaxEventType.Attribute /*, { highWaterMark: 256 * 1024 } */);
 
 /** @type {Error[]} */
 let checkLocalFiles = [];
@@ -78,7 +76,7 @@ function extractReferences(htmlFilePath) {
   };
 
   return new Promise(resolve => {
-    const readable = fs.createReadStream(htmlFilePath, streamOptions);
+    const readable = fs.createReadStream(htmlFilePath);
     readable.on('data', chunk => {
       // @ts-expect-error
       parserReferences.write(chunk);
@@ -114,7 +112,7 @@ function idExists(filePath, id) {
   };
 
   return new Promise(resolve => {
-    const readable = fs.createReadStream(filePath, streamOptions);
+    const readable = fs.createReadStream(filePath);
     readable.on('data', chunk => {
       // @ts-expect-error
       parserIds.write(chunk);