Serializers

A set of handy-if-janky wrappers for parsing and stringifying various serialization formats.

Caveats

• This package hides most details of each format, and allow arbitrary stuff to be thrown at a generic serializer interface.
• None of these implementations are meant to be used with large quantities of data; if you're parsing a hundred megs of JSON or a million line CSV, for the love of god, don't use this. It's a convenience wrapper, not an optimization.

The Rogues' Gallery

Each of the serializers adheres to a basic interface: A constructor accepts serializer-specific options, a stringify() function does what it says on the tin, and a parse() function does the opposite. Some serializers also support a validate() function if they're capable of checking validity before actually serializing.

• base64 is silly simple, but served as a nice demo while I was testing FS-Jetpack integration.
• csv and tsv, via csv-parse and csv-stringify. They're configured to assume header columns exist, parse the records to an array of objects, and be forgiving of missing/extra columns from line to line.
• ejson, a JSON variant used by MongoDB that can roundtrip dates, regexes, and buffers, via the ejson project.
• frontmatter in markdown files, in either JSON or YAML format, via the grey-matter project.
• ini, the old reliable, via the ini project.
• json, json5, and nsjson, via the json5 project. A convenience reviver that parses JSON-serialized dates is also provided. Just new NdJson({ reviver: JsonDateReviver }) and Bob's your unkle.
• php. Yeah, various PHP-based CMSs like Wordpress and Drupal have a tendency to jam internal configuration data into DB tables using PHP's serialize/deserialize functions. The code to decihper these little bundles of joy descends from php-js; I updated it with additional type-checking typing and error handling.
• plist, the XML doctype Apple uses to store tons of MacOS config files, via the plist project.
• toml, GitHub founder Tom Preston-Warner's personal config file language, via the @iarna/toml project.
• yaml, the venerable config storage format that shows up everywhere, via the yaml project.

Using them with FS Jetpack

Most of these are exposed as parse/stringify pairs that can be passed into my fork of fs-jetpack, where they'll transparently do their thing when reading and writing files with a specified extension. Again, this is absolutely not the most efficient way to handle large piles of data, but for small to middlin' ones it's extremely handy.

import jetpack from '@eatonfyi/fs-jetpack';
import { Yaml } from '@eatonfyi/serializers;

jetpack.setSerializer('.yaml', new Yaml());

const config = jetpack.read('./my-config.yaml', 'auto');

console.log(config); // Actual parsed data, not the raw string. Yay.