Schema-Fixer

What is Schema-Fixer?

Schema-Fixer is a lightweight utility designed to sanitize data while preserving its types.

What is it not?

Schema-Fixer is not a transformer or a validator—it doesn't validate correctness or enforce structure. It simply "fixes" data to align with a given schema.

Why use Schema-Fixer?

Handling null, undefined, or unexpected data formats can lead to application crashes or unintended behavior. Schema-Fixer ensures that your data conforms to the expected schema, allowing your application to run smoothly without errors caused by malformed input.

Key Features

• Reliable and Fail-Safe Schema-Fixer can process any data against a given schema without failing.

• Type Inference Automatically infers data types based on the provided schema.

• Graceful Handling of null and undefined Replaces null or undefined values with predefined default values, ensuring data integrity

• Prevents Unwanted Properties Excludes extraneous or unexpected properties, mitigating risks like code injection.

Install

# install from your project directory
npm install @gchumillas/schema-fixer

// and import it from your file.js or file.ts
import sf from '@gchumillas/schema-fixer'

console.log(sf.fix(100, 'string')) // returns '100'

Example

In the following example we obtain the data from an external API and ensure that it conforms to the expected format:

import sf from '@gchumillas/schema-fixer'

// unwanted `null` and `undefined` values are transformed to '', 0, [], etc.
// no need to indicate the return type, since it is inferred from the schema
function getAuthor = async (authorId: string) => {
  const res = await axios.get(`/api/v1/authors/${authorId}`)

  // 'fixes' the data against a given 'schema'
  return sf.fix(res.data, {
    name: sf.text(),
    middleName: sf.text(),
    lastName: sf.text(),
    age: sf.float(),
    isMarried: sf.bool(),
    // an array of strings
    children: sf.list({ of: sf.text() }),
    // an object
    address: {
      street: sf.text(),
      city: sf.text(),
      state: sf.text()
    },
    // array of objects
    books: sf.list({ of: {
      title: sf.text(),
      year: sf.float(),
      // combine multiple 'fixers'
      id: [sf.trim(), sf.upper()]
    }})
  })
}

Sugar syntax

For brevity, you can also use the "sugar syntax", which replaces "aliases" with the corresponding fixers:

import sf from '@gchumillas/schema-fixer'

const fixedData = sf.fix(data, {
  name: 'string',       // sf.text()
  age: 'number',        // sf.float()
  isMarried: 'boolean', // sf.bool()
  children: 'string[]', // sf.list({ of: sf.text() })
  years: 'number[]',    // sf.list({ of: sf.float() })
  items: 'boolean[]'    // sf.list({ of: sf.bool() })
})

API

fix(data, schema)        // fixes 'any data' againts a given schema
createFixer(def, fixer)  // creates a custom fixer

// built-in fixers
text({ def = '', required = true, coerce = true })     // fixes a string
float({ def = 0, required = true, coerce = true })     // fixes a number
bool({ def = false, required = true, coerce = true })  // fixes a boolean
list({ of: Schema, def = [], required = true })        // fixes an array

// additional built-in fixers
trim({ def = '', required = true })                    // trims a string
lower({ def = '', required = true })                   // lowercases a string
upper({ def = '', required = true })                   // uppercases a string
floor({ def = 0, required = true })                    // math floor

• A 'fixer' is the function that fixes the incoming data.

• A Schema can be a 'fixer', a list of 'fixers' or a record of 'fixers'. For example:

fix(1, bool()) // true
fix('Hello!', [text(), upper()]) // 'HELLO!'
fix({ name: 'John' }, { name: text(), age: float() }) // { name: 'John, age: 0 }

• The def parameter indicates the value to return when the 'fixer' cannot fix the incoming data or the data is null or undefined. For example:

fix(null, text()) // ''
fix(undefined, text({ def: 'aaa' })) // 'aaa'
fix({}, float({ def: 100 })) // 100, since {} cannot be fixed

• The coerce parameter indicates that you want to "force" the conversion (default is true). For example:

fix(100, text()) // '100'
fix(100, text({ coerce: false })) // '', since 100 is not a string

• The required parameter indicates that an incoming value is required. For example:

fix(undefined, float()) // 0, as we expect a number
fix(null, float({ required: false })) // undefined
fix(undefined, text({ required: false })) // undefined

[!NOTE] You'll probably never have to use required, def or coerce. But they're there!

Take a look at the TEST FILE for more examples.

Combine fixers

You can apply different "fixers" to the same incoming data by combining them. For example:

fix('Hello!', [text(), upper()]) // 'HELLO!'

Create your own fixer

This is a simple fixer:

import sf from '@gchumillas/schema-fixer'

// fixes a color
const colorFixer = sf.createFixer('#000000', (value) => {
  const color = `${value}`.trim().toUpperCase()

  if (color.match(/^#([0-9,A-F]{6})$/)) {
    // nice!
    return color
  } else {
    const matches = color.match(/^#([0-9,A-F]{3})$/)

    if (matches) {
      return `#${matches[1].split('').map(d => d.repeat(2)).join('')}`
    }
  }

  // a default value is provided
  throw new TypeError('not a color')
})

sf.fix('#f6f', colorFixer())    // '#FF66FF'
sf.fix('#f6ef6f', colorFixer()) // '#F6EF6F'
sf.fix('red', colorFixer())     // '#000000', as 'red' is not a valid color

Contributing

Do you want to contribute? Fantastic! Take a look at open issues or help me find and fix bugs.