Support plurals when adding contributions e.g. docs (and doc)

[jakebolam]

Is your feature request related to a problem? Please describe.
@AllContributorsBot please add jakebolam for docs, infra and code results in the additions of infra and code (not doc as well)

Describe the solution you'd like
The bot should be able to understand plurals. And add doc as well.

Describe alternatives you've considered

Additional context

[Berkmann18]

I've looked at parse-comment and here are my thoughts:

1. How about validContributionTypes includes the plural of each item on the list

Possible solution

validContributionTypes.forEach(type => validContributionTypes.push(`${type}s`))

Pros

1. It's One more line that will handle the valid types but in plural.
2. Doesn't require more dependencies or numerous code to achieve this.

Cons

1. It's not including the particular plural cases for: eventOrganizing (e.g.: eventsOrganizing), fundingFinding (e.g.: fundingsFinding), userTesting (e.g.: usersTesting).
2. It's an additional O(n)/Θ(n) (where n = validContributionTypes.length) which will make the process slower as validContributionTypes grows.

2. Using RegExp

Possible solution

const TYPE_RE = /([a-z]+)([A-Z][a-z]+)/;
validContributionTypes
  .map(type => {
    let part = TYPE_RE.exec(type);
    if (part === null) return [type, `${type}s`]
    let [, start, end] = part;
    return [`${start}${end}`, `${start}s${end}`, `${start}${end}s`, `${start}s${end}s`];
  })
  .reduce((a, b) => a.concat(b), [])

Pros

1. Native approach.
2. Handles all singular/plural variations.

Cons

1. Can be nasty and hard to maintain (especially as validContributionTypes changes).

3. Using what compromise offers

Possible solution

const contributions = doc
+       .forEach(type => contributions.push(type.toPlural()))
        .match('#Contribution')
        .data()
        .map(data => {
            // This removes whitespace, commas etc
            return data.normal
        })

Pros

1. It utilises more the NLP library that this bot uses.
2. It's less maintenance for this package when it comes to things like this.
3. It can account for a wide array of plural forms that would be valid in English

Cons

1. It might be tricky to get right (especially for people not familiar with compromise).

[jakebolam]

@Berkmann18 thanks for looking into this. Really like the break down of options, so great!

I like your 3rd solution (continuing to use compromise). I think there is a learning curve to compromise but I'm hoping that it will allow the bot to expand into better parsing/understanding of user intentions.

I have the vision of the bot being able to recognise terms such as
@AllContributorsBot please add Berkmann18 for documentation, ideas
e.g.

• doc could be: docs, documentation, documenting etc

Or even completly different wording.
@AllContributorsBot Berkmann18 has been great on the docs, could you add him as a contributor

Add maybe even additional languages

[allcontributors]

@jakebolam

I'm not sure how to thanks

Basic usage: @@AllContributorsBot please add jakebolam for code, doc

[jakebolam]

^ Need to figure out better ways of it responding to messages, not intended for it too (ticketed some stuff here #14)

[jakebolam]

@Berkmann18 did you want to put a PR for this? As you did all the work :)

[Berkmann18]

@jakebolam I started playing around with the parse-comment module and compromise, trying to find a way to get the plural forms of the #Contributions included.

[Berkmann18]

It turns out compromise doesn't handle that sort of thing without having a form of RegExp or array change and I was recommended to have the array with the plural forms in it before the .match(#Contribution) happens.

[Berkmann18]

@jakebolam Don't forget to close this 😉.

[jakebolam]

Closing. Fixed in #53