Convert sets of semantic key/vale conditions to regex capture groups
Last updated 4 years ago by timkinnane .
MIT · Repository · Bugs · Original npm · Tarball · package.json
$ cnpm install conditioner-regex 
SYNC missed versions from official npm registry.

Conditioner - Regex in a bottle

NPM version

NB: This is a beta release, features are stable but documentation and coverage is incomplete.

What it do?

Conditioner provides semantic conditions for string comparison and capture.

It isn't intended to replace the full power of regex, just simplify it for some common use cases.

It was developed specifically for messanging and chat-bot scripts, to compare conversational streams against listeners and capture user inputs with easy to reference keys.

Options / Defaults

matchWords: true

Only whole words are matched, so words like "whoever" won't trigger false positives if "who" is what we're looking for.

ignoreCase: true

Conditions are case insensitive by default

ignorePunctuation: true

Punctuation is stripped from inputs by default, so that messages like "Hello World!" and "hello, world" can be treated equally.

escapeValues: false

By default, conditions won't escape special characters, so regex can be used within conditions to enhance capturing.

e.g. contains: 'coffee(s)?' will match "A coffee" and "1000 coffees".

Set to true if you actually need to match against special characters.

Condition Types

  • is matches the entire string
  • starts matches the start
  • ends matches the end
  • contains matches a segment
  • excludes negative match a segment
  • after matches following a segment
  • before matches preceding segment
  • range matches a number range


NB: It's written in coffee-script so that's what I use in demos, however the dist is compiled to js

Install: npm install conditioner-regex --save

Require: Conditioner = require 'conditioner-regex'

Initialize: new Conditioner( [conditions[, options]] )

Or (optionally) with some conditions and options...

c = new Conditioner
  starts: 'who'
  ends: '?'
  escapeValues: true
  ignorePunctuation: false

Would match any string that starts with "who" and ends in a question mark.

Conditions can be passed as an array or an object.

.add( condition[, key] )

Adds further conditions and can be chained on the constructor. You can also add regex objects or strings containing regex patterns as conditions.

re = new RegExp /test/, 'i'
c.add re, 'regexp'
c.add /test/i, 'regexpString'
c.add is: 'test', 'regexpCondition'

The above creates three identical conditions, match results can be accessed by their keys.


Object keeping passed conditions, each converted to regex.

.compare( string )

Compares the string against all loaded conditions. e.g. Looking for "Who... ?" 'Who goes there?'
# returns true 'Who? Tis I!'
# returns false

.capture( string )

Captures and returns the parts of the string that match each condition. e.g.

c = new Conditioner
  after: "name is"
  range: "1-99"

c.capture "My name is Isaac, I'm 72"
# returns { 0: 'Isaac', 1: '72' }

.compared and .captured

Keeps the results of each condition for the last comparison/capture.


Keeps the full returned results for a standard regex match


Clears the compare or capture results, leaving existing conditions.


Clears results and conditions.

Full Example

# determine replies to an array of coffee (or possibly not) orders
replyToOrders = (orders) ->

  # order conditions for validity
  validity = new Conditioner()
    .add starts: 'order|get'
    .add contains: 'coffee(s)?', 'coffeePlural'
    .add excludes: 'not'

  # order details
  deets = new Conditioner()
    .add contains: 'me', 'forSelf'
    .add range: '1-999', 'qty'
    .add after: 'for', 'for'
    .add ends: 'please', 'polite' (order) ->
    detail = deets.capture order # capture details
    valid = order # test validity

    # get parts
    coffeePlural = validity.matches.coffeePlural?[0] # coffee, coffees or undefined
    qty = detail.qty ? '1'
    who = if detail.forSelf then "you" else detail.for ? "I dunno?"
    polite = if detail.polite then yes else no

    # compose
      when valid and polite
        "#{ qty } #{ coffeePlural } for #{ who }. Have a nice day :)"
      when valid
        "#{ qty } #{ coffeePlural } for #{ who }"
      when not valid and polite
        "Sorry, no."
        "No coffee for you."

# coffee orders, input received
orders = [
  'Order me a coffee please'
  'Order 2 coffees for Otis'
  'Get me 100 coffees'
  'Order Borat 10 coffees please... NOT!'
  'Order me a horse, please'
  'I love lamp'

console.log replyToOrders orders

[ '1 coffee for you. Have a nice day :)',
  '2 coffees for Otis',
  '100 coffees for you',
  'No coffee for you.',
  'Sorry, no.',
  'No coffee for you.' ]

Download and open /docs for docco generated code comments for further detail.

Current Tags

  • 0.1.2                                ...           latest (4 years ago)

3 Versions

  • 0.1.2                                ...           4 years ago
  • 0.1.1                                ...           4 years ago
  • 0.1.0                                ...           4 years ago
Maintainers (1)
Today 0
This Week 0
This Month 0
Last Day 0
Last Week 0
Last Month 1
Dependencies (1)
Dev Dependencies (20)
Dependents (0)

Copyright 2014 - 2016 © |