node package manager
Share your code. npm Orgs help your team discover, share, and reuse code. Create a free org »

@studio/log

Studio Log

👻 Log ndjson to an output stream, format the output with emoji ✨

Features

  • Fancy formats with emoji for log file reading pleasure.
  • API designed to produce expressive source code.
  • Uses topics instead of log levels for more fine grained filtering.
  • Uses object streams to avoid serialize -> parse -> serialize when used in a command line application.
  • Disabled by default. If no output stream is specified, no logs are written.

Usage

Log output is disabled by default to ensure logs don't get in the way when writing unit tests. Therefore you want to set this up as the first thing in your main:

// Sending raw JSON logs to stdout, e.g. in a server application:
require('@studio/log').out(process.stdout);
 
// Sending fancy formatted logs to stdout, e.g. in a command line tool:
require('@studio/log')
  .transform(require('@studio/log/format/fancy')())
  .out(process.stdout);

Next, create a logger instance in a module and start writing logs:

const logger = require('@studio/log');
 
const log = logger('app');
 
exports.startService = function (port) {
  log.launch('my service', { port: 433 });
};
 

In the server example above, this output is produced:

{"ts":1486630378584,"ns":"app","topic":"launch","msg":"my service","data":{"port":433}}

Send your logs to the emojilog command for pretty printing:

$ cat logs.ndjson | emojilog
09:52:58 🚀  app my service port=433

Install

$ npm install @studio/log --save

Install this module globally to get the emojilog command line tool to format logs:

$ npm install @studio/log -g
# ... 
$ node app.js | emojilog

Topics

Instead of log levels, this logger uses a set of topics to categorize, format and filter logs. Unlike log levels, topics are not ordered by severity.

These topics are available:

  • ✅ = ok
  • ⚠️ = warn
  • 🐛 = issue
  • 🚨 = error
  • 🙈 = ignore
  • 🔺 = input
  • 🔻 = output
  • 📤 = send
  • 📥 = receive
  • 📡 = fetch
  • 🏁 = finish
  • 🚀 = launch
  • ⛔️ = terminate
  • ✨ = spawn
  • 📣 = broadcast
  • 💾 = disk
  • ⏱ = timing
  • 💰 = money
  • 🔢 = numbers
  • 👻 = wtf

Log format

  • ns: The logger instance namespace.
  • ts: The timestamp as returned by Date.now().
  • topic: The topic name.
  • msg: The message.
  • data: The data.
  • stack: The stack of error object.
  • cause: The cause stack of error.cause object, if available.

API

Creating a logger

  • log = logger(ns[, data]): Creates a new logger with the given namespace. The namespace is added to each log entry as the ns property. If data is provided, it is added to each log entry. Multiple calls with the same ns property return the same logger instance while data is replaced.
  • log.child(ns[, data]): Creates a child logger of a log instance. The namespaces are joined with a blank and data is merged. Multiple calls with the same ns property return the same logger instance while data is replaced.

Log instance functions

  • log.{topic}([message][, data][, error]): Create a new log entry with these behaviors:
    • The topic is added as the "topic".
    • If message is present, it's added as the "msd".
    • If data is present, it's added as the "data".
    • If error is present, the stack property of the error is added as the "stack". If no stack is present, the toString representation of the error is used.
    • If error.code is present, it is added to the "data" without modifying the original object.
    • If error.cause is present, the stack property of the cause is added as the "cause". If no stack is present, the toString representation of the cause is used.
    • If error.cause.code is present, a cause object is added to the "data" with { code: cause.code } and without modifying the original object.
  • log.filter([ns, ]stream): Configure a filter stream for this logger or a child logger namespace. See "Filter Streams".
  • log.mute(): Mute this logger namespace.

Global functions

  • logger.filter([namespace, ]stream): Configures a filter stream for the given namespace, or a global filter stream. See "Filter Streams".
  • logger.mute(namespace[, topic]): Mute the given namespace or only the topic in the namespace, if given.
  • logger.muteAll(topic): Mute the given topic in all namespaces.
  • logger.out(stream): Configure the output stream to write logs to. If not specified, no logs are written.
  • logger.transform(stream): Configure a transform stream to format logs. The given stream must be in readableObjectMode. See "Format Transforms". further down. Defaults to @studio/ndjson/stringify.
  • logger.hasStream(): Whether an output stream was set.
  • logger.reset(): Resets everything to the defaults.

CLI Options

  • --format or -f set the formatter to use. Defaults to "fancy".
  • --no-ts hide timestamps
  • --no-topic hide topics
  • --no-ns hide namespaces
  • --no-data hide data
  • --no-stack hide stacks
  • --stack message only show the error message
  • --stack peek show the message and the first line of the trace (default)
  • --stack full show the message and the full trace
  • --stack same as --stack full
  • --map use given source maps file to map stack traces

Filter Streams

Filter streams can be used to alter the data before passing it to the transform stream. Filter streams must be in objectMode. For example, Studio Log X is a filter stream implementation.

Format Transforms

Install a transform stream if you want to use Studio log in a command line application. The bundled transform streams have to be required separately:

const formatter = require('@studio/log/format/basic');
 
logger.transform(formatter({ ts: false }));

The following transform streams are available:

  • basic: Basic formatting with ISO dates and no colors
  • fancy: Colored output with localized dates. This is the default formatter when using the emojilog CLI.

Some advanced formatting is applied by naming conventions on top level properties of the data object. See demo.js for some examples.

  • ts or prefix ts_ formats a timestamp.
  • ms or prefix ms_ formats a millisecond value.
  • bytes or prefix bytes_ formats a byte value.

These options can be passed to the bundled format transforms:

  • ts: false hide timestamps
  • topic: false hide topics
  • ns: false hide namespaces
  • data: false hide data
  • stack: style with these stack styles:
    • false: hide the error entirely
    • message only show the error message
    • peek show the message and the first line of the trace (default)
    • full show the message and the full trace

The stack option is also used to format the "cause", if present.

Custom Format Transforms

Format transforms are node transform streams in writableObjectMode. Here is an example transform implementation, similar to the default transform:

const { Transform } = require('stream');
 
const ndjson = new Transform({
  writableObjectMode: true,
 
  transform(entry, enc, callback) {
    const str = JSON.stringify(entry);
    callback(null, `${str}\n`);
  }
});

Related modules

  • ☯️ Studio ndjson can be used to parse the ndjson produced by Studio log. The ParseTransform is internally used by the emojilog command and the StringifyTransform is used as the default transform to serialize to ndjson.
  • Studio Log X x-out confidential data in log entries.
  • 📦 Studio Changes is used to create the changelog for this module.

License

MIT

Made with ❤️ on 🌍