Improve singular method

[slawekjaranowski]

No description provided.

[gnodet]

Here's a proposal:

    private static final Map<String, String> PLURAL_EXCEPTIONS = new HashMap<>();

    static {
        // Irregular plurals
        PLURAL_EXCEPTIONS.put("men", "man");
        PLURAL_EXCEPTIONS.put("women", "woman");
        PLURAL_EXCEPTIONS.put("children", "child");
        PLURAL_EXCEPTIONS.put("mice", "mouse");
        PLURAL_EXCEPTIONS.put("people", "person");
        PLURAL_EXCEPTIONS.put("teeth", "tooth");
        PLURAL_EXCEPTIONS.put("feet", "foot");
        PLURAL_EXCEPTIONS.put("geese", "goose");

        // Invariant plurals
        PLURAL_EXCEPTIONS.put("series", "series");
        PLURAL_EXCEPTIONS.put("species", "species");
        PLURAL_EXCEPTIONS.put("sheep", "sheep");
        PLURAL_EXCEPTIONS.put("fish", "fish");
        PLURAL_EXCEPTIONS.put("deer", "deer");
        PLURAL_EXCEPTIONS.put("aircraft", "aircraft");

        // Special "oes" exceptions
        PLURAL_EXCEPTIONS.put("heroes", "hero");
        PLURAL_EXCEPTIONS.put("potatoes", "potato");
        PLURAL_EXCEPTIONS.put("tomatoes", "tomato");
        PLURAL_EXCEPTIONS.put("echoes", "echo");
        PLURAL_EXCEPTIONS.put("vetoes", "veto");
        PLURAL_EXCEPTIONS.put("torpedoes", "torpedo");
        PLURAL_EXCEPTIONS.put("cargoes", "cargo");
        PLURAL_EXCEPTIONS.put("haloes", "halo");
        PLURAL_EXCEPTIONS.put("mosquitoes", "mosquito");
        PLURAL_EXCEPTIONS.put("buffaloes", "buffalo");
    }

    public static String singular(String plural) {
        if (plural == null || plural.isEmpty()) return plural;

        String lower = plural.toLowerCase();

        if (PLURAL_EXCEPTIONS.containsKey(lower)) {
            return PLURAL_EXCEPTIONS.get(lower);
        }

        // Suffix-based rules
        if (lower.endsWith("ies") && plural.length() > 3) {
            return plural.substring(0, plural.length() - 3) + "y";
        }
        if (lower.endsWith("ves")) {
            return plural.substring(0, plural.length() - 3) + "f";
        }
        if (lower.endsWith("zzes")) {
            return plural.substring(0, plural.length() - 2);
        }
        if (lower.endsWith("sses")) {
            return plural.substring(0, plural.length() - 2);
        }
        if (lower.endsWith("ches") || lower.endsWith("shes")) {
            return plural.substring(0, plural.length() - 2);
        }
        if (lower.endsWith("xes")) {
            return plural.substring(0, plural.length() - 2);
        }
        if (lower.endsWith("oes")) {
            return plural.substring(0, plural.length() - 1);
        }
        if (lower.endsWith("s") && plural.length() > 1) {
            return plural.substring(0, plural.length() - 1);
        }

        return plural;
    }

With a more complete test pairs:

    // Known exceptions
    "men", "man",
    "women", "woman",
    "children", "child",
    "mice", "mouse",
    "people", "person",
    "teeth", "tooth",
    "feet", "foot",
    "geese", "goose",

    "series", "series",
    "species", "species",
    "sheep", "sheep",
    "fish", "fish",
    "deer", "deer",
    "aircraft", "aircraft",

    "heroes", "hero",
    "potatoes", "potato",
    "tomatoes", "tomato",
    "echoes", "echo",
    "vetoes", "veto",
    "torpedoes", "torpedo",
    "cargoes", "cargo",
    "haloes", "halo",
    "mosquitoes", "mosquito",
    "buffaloes", "buffalo",

    // Regular plural forms with suffixes
    "voes", "voe",
    "hoes", "hoe",
    "canoes", "canoe",
    "toes", "toe",
    "foes", "foe",
    "oboes", "oboe",
    "noes", "no",
    "boxes", "box",
    "wishes", "wish",
    "dishes", "dish",
    "brushes", "brush",
    "classes", "class",
    "buzzes", "buzz",
    "cars", "car",
    "dogs", "dog",
    "cats", "cat",
    "horses", "horse",

    // Some test cases with different rules
    "wolves", "wolf",
    "knives", "knife",
    "leaves", "leaf",
    "wives", "wife",
    "lives", "life",
    "babies", "baby",
    "parties", "party",
    "cities", "city",
    "buses", "bus",
    "boxes", "box",
    "churches", "church",
    "matches", "match",
    "watches", "watch",
    "riches", "rich",
    "dresses", "dress",
    "crosses", "cross",

    // More edge cases
    "heroes", "hero",
    "vetoes", "veto",
    "torpedoes", "torpedo",
    "tomatoes", "tomato",
    "potatoes", "potato",
    "echoes", "echo",
    "mosquitoes", "mosquito",
    "buffaloes", "buffalo",
    "volcanoes", "volcano",
    "goes", "go"

[gnodet]

Why is that one an exception?
The plural just adds an 's'

[slawekjaranowski]

end with es

[gnodet]

We can add a few more, here's a more extensive list:

"women", "woman", "men", "man", "children", "child", "mice", "mouse", "people", "person", "series", "series", "species", "species", "roses", "rose", "fezzes", "fez", "kisses", "kiss", "buses", "bus", "glasses", "glass", "heroes", "hero", "potatoes", "potato", "tomatoes", "tomato", "echoes", "echo", "torpedoes", "torpedo", "vetoes", "veto", "cargoes", "cargo", "haloes", "halo", "mosquitoes", "mosquito", "babies", "baby", "wolves", "wolf", "knives", "knife", "leaves", "leaf", "wives", "wife", "lives", "life", "boxes", "box", "wishes", "wish", "dishes", "dish", "churches", "church", "brushes", "brush", "classes", "class", "buzzes", "buzz", "cars", "car", "dogs", "dog", "voes", "voe", "does", "doe", "hoes", "hoe", "canoes", "canoe"

[gnodet]

The oes rule is a bit more complicated.
It usually looses the final s, but there are exceptions such as heroes, potatoes, tomatoes, echoes, torpedoes, vetoes, cargoes, haloes, and mosquitoes.
I think the rule should be loose the final s, with exceptions such as:

    "heroes", "hero",
    "potatoes", "potato",
    "tomatoes", "tomato",
    "echoes", "echo",
    "vetoes", "veto",
    "torpedoes", "torpedo",
    "cargoes", "cargo",
    "haloes", "halo",
    "mosquitoes", "mosquito",
    "buffaloes", "buffalo"

[slawekjaranowski]

rule:

        if (lower.endsWith("ves")) {
            return name.substring(0, name.length() - 3) + "f";
        }

not works for:

"knives, knife"
"wives, wife"
"lives, life",

[gnodet]

rule:

        if (lower.endsWith("ves")) {
            return name.substring(0, name.length() - 3) + "f";
        }

not works for:

"knives, knife"
"wives, wife"
"lives, life",

Those are actually incorrect plurals. the correct ones are ending with ves, knives, wives and lives.

[slawekjaranowski]

rule:

        if (lower.endsWith("ves")) {
            return name.substring(0, name.length() - 3) + "f";
        }

not works for:

"knives, knife"
"wives, wife"
"lives, life",

Those are actually incorrect plurals. the correct ones are ending with ves, knives, wives and lives.

I see knives the same - in my example and in your 😄

[slawekjaranowski]

@gnodet - based on your proposition I have a next fix ...

I'm afraid that it will be difficult to support all cases, so I added a parameter to Mojo, when we can add special exclusion in project.

[slawekjaranowski]

I have found it: https://github.com/dry-rb/dry-inflector/blob/main/lib/dry/inflector/inflections/defaults.rb

[hboutemy]

defining what "improving" means would be useful: as such, it's just a vague personal judgement

but I read the content, and it's related to being able to add exceptions and have a default list of classical ones

question: is there an official list somewhere?
i did a search, I could find https://www.thoughtco.com/irregular-plural-nouns-in-english-1692634 and many others
it would be nice to keep a pointer to the reference list used as a default

[slawekjaranowski]

As I started work on it .... I've hoped it will be only simple improvement ....

But the problem turned out to be more complicated I have added a list of some irregular form @gnodet comments - I don't know a source.

Finally I added a parameter to allow providing more special cases, as it is difficult to discover all irregular noun

Of course we can adjust change title to show what exactly was done.