feat: open file after scan on Windows and MacOS, --open-file

README.md

@@ -57,6 +57,7 @@ Options:
   --out-dir <dir>              Output directory (default: ".")
   --csv <path>                 Skip scan, only convert csv to xlsx
   --no-color                   No console colors
+  --open-file                  Open file after scan (default: yes on Windows and MacOS)
   --no-console-validate        Don't output validate messages in console
   -V, --version                output the version number
   -h, --help                   display help for command
@@ -79,7 +80,7 @@ Options:
 - Title is right-aligned to reveal the common part
 - Validation of some columns (status, request time, description length)
 
-### Fields list (20.04.2020):
+### Fields list (20.07.2020):
 - url
 - mixed_content_url
 - canonical
@@ -99,6 +100,7 @@ Options:
 - h2_count
 - h3_count
 - h4_count
+- canonical_count
 - images
 - images_without_alt
 - images_alt_empty
@@ -215,7 +217,6 @@ site-audit-seo -d 1 -u https://example -f '{ "title": "$(`title`).text()" }'
 
 ## TODO:
 - Unique links
-- Scraper freezes when scrap doc in 2 threads
 - [Offline w3c validation](https://www.npmjs.com/package/html-validator)
 - [Words count](https://github.com/IonicaBizau/count-words)
 - [Sentences count](https://github.com/NaturalNode/natural)

src/index.js

@@ -5,6 +5,8 @@ const { program } = require('commander');
 const packageJson = require('../package.json');
 const scrapSite = require('./scrap-site');
 const saveAsXlsx = require('./save-as-xlsx');
+const { exec } = require('child_process');
+const os = require('os');
 
 const list = val => {
   return val ? val.split(',') : [];
@@ -28,19 +30,25 @@ program
   .option('--out-dir <dir>', `Output directory`, '.')
   .option('--csv <path>', `Skip scan, only convert csv to xlsx`)
   .option('--no-color', `No console colors`)
+  .option('--open-file', `Open file after scan (default: yes on Windows and MacOS)`)
   .option('--no-console-validate', `Don't output validate messages in console`)
   .name("site-audit-seo")
   .version(packageJson.version)
   .usage("-u https://example.com")
   .parse(process.argv);
 
 async function start() {
+  if(program.openFile === undefined) {
+    program.openFile = ['darwin', 'win32'].includes(os.platform());
+  }
+
   if(program.csv) {
     const csvPath = program.csv
     const xlsxPath = csvPath.replace(/\.csv$/, '.xlsx');
     try {
       saveAsXlsx(csvPath, xlsxPath);
-      console.log(`${xlsxPath} saved`)
+      console.log(`${xlsxPath} saved`);
+      if(program.openFile) exec(`"${xlsxPath}"`);
     } catch(e) {
       console.error(e)
     }
@@ -79,6 +87,7 @@ async function start() {
       docsExtensions: program.docsExtensions,     // расширения, которые будут добавлены в таблицу
       outDir: program.outDir,                     // папка, куда сохраняются csv
       color: program.color,                       // раскрашивать консоль
+      openFile: program.openFile,                 // открыть файл после сканирования
       fields: program.fields,                     // дополнительные поля
       removeCsv: program.removeCsv,               // удалять csv после генерации xlsx
       consoleValidate: program.consoleValidate,   // выводить данные валидации в консоль

src/scrap-site.js

@@ -5,6 +5,7 @@ const HCCrawler = require('@popstas/headless-chrome-crawler');
 const CSVExporter = require('@popstas/headless-chrome-crawler/exporter/csv');
 const url = require('url');
 const {validateResults, getValidationSum} = require('./validate');
+const { exec } = require('child_process');
 
 const DEBUG = true; // выключить, если не нужны console.log на каждый запрос (не будет видно прогресс)
 
@@ -357,6 +358,7 @@ module.exports = async (baseUrl, options = {}) => {
 
     console.log(`\n${color.yellow}Saved to ${xlsxPath}${color.reset}`);
     console.log(`Finish: ${t} sec (${perPage} per page)`);
+    if(options.openFile) exec(`"${xlsxPath}"`);
   };
 
   let isSuccess = true;