How to fix npm ERR! missing script: start

Problem:

You want to run a NodeJS app using

npm start

but you only see this error message:

npm ERR! missing script: start

Solution:

You need to tell npm what to do when you run npm start explicitly by editing package.json.

First, identify the main file of your application. Most often it is called index.js, server.js or app.js. When you open package.json in an editor, you can also often find a line like

"main": "index.js",

In this example, index.js is the main file of the application.

Now we can edit package.json to add a start script.

In package.json, find the "scripts" section. If you have a default package.json, it will look like this:

"scripts": {
  "test": "echo \"Error: no test specified\" && exit 1"
},

Now add a comma to the end of the "test" line and add this line after it:

"start": "node index.js"

and replace index.js by the main file of your application (e.g. app.js, server.js, index.js etc).

The "scripts" section should now look like this:

"scripts": {
  "test": "echo \"Error: no test specified\" && exit 1",
  "start": "node index.js"
},

Now save and close package.json and run

npm start

to start your application.

How to read NodeJS child-process.exec stdout/stderr using async/await Promises

You want to run a command like file my.pdf using NodeJS child-process.exec and get its stdout after it’s finished.

Solution:

TL;DR: (await exec('file my.pdf')).stdout

We’re using child-process-promise here in order to simplify our implementation. Install it using npm i --save child-process-promise !

const { exec } = require('child-process-promise');

async function run () {
    const ret = await exec(`file my.pdf`);
    return ret.stdout;
}

run().then(console.log).catch(console.error);

You can also use .stderr instead of .stdout to get the stderr output as a string

How to fix NodeJS MongoDB ‘Cannot read property ‘high_’ of null’

When encountering an error message like

TypeError: Cannot read property 'high_' of null
    at Long.equals (/home/uli/dev/NMUN/node_modules/bson/lib/bson/long.js:236:31)
    at nextFunction (/home/uli/dev/NMUN/node_modules/mongodb-core/lib/cursor.js:473:16)
    at Cursor.next (/home/uli/dev/NMUN/node_modules/mongodb-core/lib/cursor.js:763:3)
    at Cursor._next (/home/uli/dev/NMUN/node_modules/mongodb/lib/cursor.js:211:36)
    at nextObject (/home/uli/dev/NMUN/node_modules/mongodb/lib/operations/cursor_ops.js:192:10)
    at hasNext (/home/uli/dev/NMUN/node_modules/mongodb/lib/operations/cursor_ops.js:135:3)
    (...)

you likely have code like this:

const cursor = db.getCollection('mycollection').find({})
while (cursor.hasNext()) {
    const doc = cursor.next();
    // ... handle doc ...
}

The solution is quite simple: Since find(), cursor.hasNext() and cursor.next() all return Promises, you can’t use their results directly.

This example shows you how to do it properly using async/await:

const cursor = await db.getCollection('mycollection').find({})
while (await cursor.hasNext()) {
    const doc = await cursor.next();
    // ... handle doc ...
}

In order to do this remember that the function containing this code will need to be an async function. See the Mozilla documentation or google for Javascript async tutorial in order to learn about the details!

Puppeteer: Get text content / inner HTML of an element

Problem:

You want to use puppeteer to automate testing a webpage. You need to get either the text or the inner HTML of some element, e.g. of

<div id="mydiv">
</div>

on the page.

Solution:

# Get inner text
const innerText = await page.evaluate(() => document.querySelector('#mydiv').innerText);

# Get inner HTML
const innerHTML = await page.evaluate(() => document.querySelector('#mydiv').innerHTML);

Note that .innerText includes the text of sub-elements. You can use the complete DOM API inside page.evaluate(...). You can use any CSS selector as an argument for document.querySelector(...).

How to fix WebPack error describe: optionsSchema.definitions.output.properties.path.description

Problem:

You are trying to build your webpack project, but you see an error message like this:

/home/uli/project/node_modules/webpack-cli/bin/config-yargs.js:89
                                describe: optionsSchema.definitions.output.properties.path.description,
                                                                           ^

TypeError: Cannot read property 'properties' of undefined
    at module.exports (/home/uli/project/node_modules/webpack-cli/bin/config-yargs.js:89:48)
    at /home/uli/project/node_modules/webpack-cli/bin/webpack.js:60:27
    at Object.<anonymous> (/home/uli/project/node_modules/webpack-cli/bin/webpack.js:515:3)
    at Module._compile (internal/modules/cjs/loader.js:723:30)
    at Object.Module._extensions..js (internal/modules/cjs/loader.js:734:10)
    at Module.load (internal/modules/cjs/loader.js:620:32)
    at tryModuleLoad (internal/modules/cjs/loader.js:560:12)
    at Function.Module._load (internal/modules/cjs/loader.js:552:3)
    at Module.require (internal/modules/cjs/loader.js:659:17)
    at require (internal/modules/cjs/helpers.js:22:18)

Solution:

This is a known bug in webpack 4.20.0 – you can circumvent this issue by using webpack 4.19.0.

Look for a line like

"webpack": "^4.7.0",

in your package.json. The caret (^) allows npm to use any 4.x.x version – including the broken 4.20.0.

Replace the aforementioned line by

"webpack": "4.19.0",

to use only webpack 4.19.0.

After that, run npm install and retry building your application.

Fixing MomentJS interpreting dates as local time

Problem:

You live in a non-UTC timezone. Dates parsed using MomentJS like

moment("2017-01-01")

are interpreted as local time as opposed to UTC (as would be appropiate based on the Z in the ISO8601 timestamp) and therefore trying to format them yields an offset timestamp:

moment("2017-01-01").toDate().toISOString()
// "2016-12-31T23:00:00.000Z" <-- Offset by 1 hour (MEZ - UTC)

This causes you trouble as often the original date is not preserved: In the example above, the correct date would be 2017-01-01 but it is 2016-12-31 instead.

Solution 1: Force moment to parse the date as UTC

moment.utc("2017-01-01").toDate().toISOString()
// "2017-01-01T00:00:00.000Z" <-- Correct

Solution 2: Manually subtract the timezone offset

let m = moment("2017-01-01");
// Subtract the difference between timezone and UTC
m = m.subtract(m.toDate().getTimezoneOffset(), 'minutes');
// m.toDate().toISOString() === "2017-01-01T01:00:00.000Z" - CORRECT

Use this if Solution 1 does not work or you can’t modify the parsing code: This solution works even if the date has already been parsed.

DO NOT use

new Date().getTimezoneOffset() // NEVER USE THIS! MIGHT USE THE WRONG OFFSET!

because it uses the timezone offset of the current date instead of the date parsed by MomentJS. Obviously this is wrong in countries with summertime, since dates may have different offsets depending on the date. Moreover, in rare cases the client’s computer timezone may have changed due to travel etc in the meantime and might therefore might not represent the correct offset to use.

Fixing Angular Error: Unexpected value ‘undefined’ declared by the module

Symptom:

For a module of yours, you get an error message like this on load:

Error: Unexpected value 'undefined' declared by the module 'MyModule'
    at syntaxError (compiler.js:1021)
    at compiler.js:10623
    at Array.forEach (<anonymous>)
    at CompileMetadataResolver.push../node_modules/@angular/compiler/fesm5/compiler.js.CompileMetadataResolver.getNgModuleMetadata (compiler.js:10621)
    at JitCompiler.push../node_modules/@angular/compiler/fesm5/compiler.js.JitCompiler._loadModules (compiler.js:23876)
    at JitCompiler.push../node_modules/@angular/compiler/fesm5/compiler.js.JitCompiler._compileModuleAndComponents (compiler.js:23857)
    at JitCompiler.push../node_modules/@angular/compiler/fesm5/compiler.js.JitCompiler.compileModuleAsync (compiler.js:23817)
    at CompilerImpl.push../node_modules/@angular/platform-browser-dynamic/fesm5/platform-browser-dynamic.js.CompilerImpl.compileModuleAsync (platform-browser-dynamic.js:143)
    at core.js:4999
[...]

Solution:

First, try to restart ng serve. In some cases this will outright fix the issue.

The error message is caused by some element of your @NgModule declarations: [ /* ... */ ] to be undefined.

For example, if you have a @NgModule declaration like this:

@NgModule({
  imports: [
    RecallCommonModule,
    CommonModule,
    MyRoutingModule
  ],
  declarations: [,
    MyDetailComponent,
    MySearchComponent
  ],
  providers: []
})

the issue is in the stray comma in the declarations line: When compiled, it results in [undefined, MyDetailComponent, MySearchComponent]. Removing the stray comma fixes the issue.

If you don’t have stray commata in your declarations, comment out every element in your declarations array and see which causes the error message to disappear.

Fixing NodeJS Intl.DateTimeFormat not formatting properly for locales

Symptom:

NodeJS starting from version v9.x supports the ES6 Intl.DateTimeFormat

When you use it with the ‘en-US’ locale, it works properly and prints "August 13, 2018":

const df = new Intl.DateTimeFormat('en-US', {day: 'numeric', month: 'long', year: 'numeric', timeZone: 'UTC'});
console.log(df.format(new Date("2018-08-13T04:00:00.000Z")));

However, using it with a different locale fails:

const df = new Intl.DateTimeFormat('de', {day: 'numeric', month: 'long', year: 'numeric', timeZone: 'UTC'});
console.log(df.format(new Date("2018-08-13T04:00:00.000Z")));

While you would expect this to print "13. August 2018" , it will print "2018 M08 13"

Reason:

By default, NodeJS is only built with small-icu support, thereby only installing the en-US locale in order to reduce the installation filesize.

Solution 1 (preferred):

You can use the intl polyfill module to completely replace NodeJS’s implementation of intl:

Installation:
npm i --save intl
Usage:
// Replace Intl by polyfill
Intl = require("intl")

const df = new Intl.DateTimeFormat('de', {day: 'numeric', month: 'long', year: 'numeric', timeZone: 'UTC'});
console.log(df.format(new Date("2018-08-13T04:00:00.000Z")));

This will print 13. August 2018 as expected.

Solution 2 (alternate):

You can use the full-icu package to continue using the NodeJS ICU implementation (i.e. no polyfill), but just install the ICU data.

While this reduces the total installation filesize, installation is slow and the exact method depends on the NodeJS version and requires more work than just using the intl polyfill.

In order to install, use

npm i --save full-icu

This will take some time to compile the data and then will print instructions like this:

 √ icudt62l.dat (link)
Node will use this ICU datafile if the environment variable NODE_ICU_DATA is set to “node_modules/full-icu”
or with node --icu-data-dir=node_modules/full-icu YOURAPP.js
 For package.json:
{"scripts":{"start":"node --icu-data-dir=node_modules/full-icu YOURAPP.js"}}

By the way, if you have full data, running this in node:
> new Intl.DateTimeFormat('es',{month:'long'}).format(new Date(9E8));
... will show “enero”. If it shows “January” you don't have full data.
News: Please see https://github.com/icu-project/full-icu-npm/issues/6

In order to actually use full-icu, you need to use the --icu-data-dir=node_modules/full-icu argument every time you run node. In order to run node interactively, use

node --icu-data-dir=node_modules/full-icu

If you use scripts in your application (e.g. the start script, i.e. what gets executed if you run npm start), you need to adjust the configuration in package.json:

Instead of

// [...]
"scripts": {
    "start": "node index.js"
}
// [...]

use

// [...]
"scripts": {
    "start": "node --icu-data-dir=node_modules/full-icu index.js"
}
// [...]

Depending on your application, you might need to use a different script name than index.js – common names include server.js and start.js

How to fix Angular6 “_getAriaLabel is not a function” with production build

Problem:

You are building an Angular6 application and in development mode everything works fine. However, if you build in production mode:

ng build --prod --aot

you see an error like this in the client:

main.4d1baabffbba5677af03.js:1 ERROR TypeError: i.ɵnov(...)._getAriaLabel is not a function
    at Object.updateRenderer (main.4d1baabffbba5677af03.js:1)
    at Object.updateRenderer (main.4d1baabffbba5677af03.js:1)
[...]

Solution:

The issue appears to be caused by incorrectly updated NodeJS modules. You can fix it by simply deleting your node_modules folder:

rm -rf node_modules

Furthermore, it’s recommended to update @angular/cli as the bug does not seem to be present in newer versions of @angular/cli:

sudo npm i -g @angular/cli

Source & discussion on GitHub

How to set cell value to string using js-xlsx

This snippet reads a XLSX file using js-xlsx, sets the C2 cell to abc123 and writes the result to another file:

const XLSX = require('xlsx');

const table = XLSX.readFile('mytable.xlsx');
// Use first sheet
const sheet = table.Sheets[table.SheetNames[0]];

// Option 1: If you have numeric row and column indexes
sheet[XLSX.utils.encode_cell({r: 1 /* 2 */, c: 2 /* C */})] = {t: 's' /* type: string */, v: 'abc123' /* value */};
// Option 2: If you have a cell coordinate like 'C2' or 'D15'
sheet['C2'] = {t: 's' /* type: string */, v: 'abc123' /* value */};

XLSX.writeFile(table, 'result.xlsx');

 

How to iterate over XLSX rows using js-xlsx

This snippet allows you to easily iterate over rows in any XLSX files using the js-xlsx library (in this example we don’t iterate over all columns but rather only get the B column as an example):

const table = XLSX.readFile('mytable.xlsx');
const sheet = table.Sheets[table.SheetNames[0]];

var range = XLSX.utils.decode_range(sheet['!ref']);
for (let rowNum = range.s.r; rowNum <= range.e.r; rowNum++) {
    // Example: Get second cell in each row, i.e. Column "B"
    const secondCell = sheet[XLSX.utils.encode_cell({r: rowNum, c: 1})];
    // NOTE: secondCell is undefined if it does not exist (i.e. if its empty)
    console.log(secondCell); // secondCell.v contains the value, i.e. string or number
}

 

How to fix Angular4/5/6 Unexpected token ‘px’

If you encounter an error message like this:

Parser Error: Unexpected token 'px' at column 3 in [70px] in ng:///AppModule/MyComponent.html@5:34 ("="let string of strings">

look at the line the error is referring to. It will look similar to this:

<mat-expansion-panel-header [collapsedHeight]="70px">

You have two options of fixing this:

Option 1: Recommended if the value (70px in this case) is always constant.

Remove the brackets from the attribute: [collapsedHeight] to collapsedHeight. The brackets mean that the value shall be interpreted as Javascript and removing them means interpreting the value as attribute. You code should look like this:

<mat-expansion-panel-header collapsedHeight="70px">

Option 2: Force angular to interpret the value (70px in this case) as a string:

Add single quotes before and after the value makes it valid Javascript:

<mat-expansion-panel-header [collapsedHeight]="'70px'">

I recommend to use this option only if you expect the value to be a non-constant javascript expression in the future.

PDFJS: Read PDF from memory Buffer in NodeJS

Note: This post uses async/await and therefore requires NodeJS 8+.

This is how to read a PDF file from a file, e.g. mypdf.pdf:

pdfjs.getDocument('mypdf.pdf');

Full example:

const pdfjs = require('pdfjs-dist');

async function readPDF() {
    const pdf = await pdfjs.getDocument('mypdf.pdf');
    // ...
}

Here’s how you can read the PDF from a memory buffer:

pdfjs.getDocument({data: buffer});

Full example

const fs = require('mz/fs')
const pdfjs = require('pdfjs-dist');

async function readPDF() {
    // Read file into buffer
    const buffer = await fs.readFile('mypdf.pdf')
    // Parse PDF from buffer
    const pdf = await pdfjs.getDocument({data: buffer});
    // ...
}

Using mz/fs is not required, it’s just used as an utility library to be able to use await with files.

 

Convert pt (postscript/PDF unit) to inch or mm in Javascript

Here are some simple utility functions to convert the preprint unit pt (defined as 1/72 inch) into inches or mm.

function convertPtToInch(pt) { return pt / 72; }
function convertInchToMM(inch) { return inch * 25.4; }
function convertPtToMM(pt) {
  return convertInchToMM(convertPtToInch(pt)); }

// Example usage
console.log(convertPtToMM(595)) // Prints 209.90277777777777

Note that while this conversion is exact, there is some tolerance required when comparing these units:
An ISO A4 paper is defined as 210x297 mm – or 595x842 pt.

However, converting 595×842 pt into mm results in 209.902777 mm and 297.038888 mm respectively. Watch out for those tolerances if you try to compare paper sizes. I recommend a tolerance of at least 0.25 mm.

Extract PDF page sizes using PDFJS & NodeJS

Although most PDFs have some pages with only one page size (e.g. DIN A4 or Letter in portrait orientation), PDFs sometimes also have pages that have another size or orientation (which is treated just like another size) that other pages in the same document.

This post provides an easy-to-reuse example on how to use PDFJS in NodeJS (though it will be just as easy to do in the browser) to extract the PDF

It is based on this previous post on how to read all pages from a PDF document using PDFJS, so be sure to check that out first.

First install the required dependencies:

npm install bereich pdfjs-dist

then you can use this source code to read the page sizes of mypdf.pdf:

const pdfjs = require('pdfjs-dist');
const bereich = require('bereich');

class PageSize {
  constructor(width, height) {
    this.width = width;
    this.height = height
  }
}

function getPageSize (page) {
    const [x, y, w, h] = page.pageInfo.view;
    const width = w - x;
    const height = h - y;
    const rotate = page.pageInfo.rotate;
    // Consider rotation
    return (rotate === 90 || rotate === 270)
        ? new PageSize(height, width) : new PageSize(width, height);
}

async function readPDFPageSizes() {
  const pdf = await pdfjs.getDocument('mypdf.pdf');
  const numPages = pdf.numPages;

  const pageNumbers = Array.from(bereich(1, numPages));
  // Start reading all pages 1...numPages
  const promises = pageNumbers.map(pageNo => pdf.getPage(pageNo));
  // Wait until all pages have been read
  const pages = await Promise.all(promises);
  // You can do something with pages here.
  return pages.map(getPageSize);
}

readPDFPageSizes()
    .then(pageSizes => {console.log(pageSizes)})
    .catch(err => {console.error(`Error while reading PDF: ${err}`)})

Running this with a document having a single A4 page will result in

[ PageSize { width: 595, height: 842 } ]

Note that the width & height unit is pt (Points). One pt is defined as 1/72 inches. A DIN A4 page (portrait) is 595x842pt, therefore you see those values here.
See this TechOverflow post for code to convert pt to mm and inches.

PDFJS: Read all pages using async/await in NodeJS

PDFJS has an official example that – among other things, reads all pages from a PDF document.
However, their promise-based method is rather complex to understand and to write. Luckily, there is an easier way using async/await (which is supported starting from NodeJS 8.x).

I’m using the bereich library (bereich is german for range) in order to generate an array of page numbers (1..numPages).
Install the required libraries using

npm install pdfjs-dist bereich

Here’s the source code example:

const pdfjs = require('pdfjs-dist');
const bereich = require('bereich');

async function readPDFPages() {
  const pdf = await pdfjs.getDocument('mypdf.pdf');
  const numPages = pdf.numPages;

  const pageNumbers = Array.from(bereich(1, numPages));
  // Start reading all pages 1...numPages
  const promises = pageNumbers.map(pageNo => pdf.getPage(pageNo));
  // Wait until all pages have been read
  const pages = await Promise.all(promises);
  // You can do something with pages here.
  return pages;
}

readPDFPages().then(pages => {
    console.log(pages)
}).catch(err => {
    console.error(`Error while reading PDF: ${err}`)
})