Support recognizing abbreviated options and commands

[remkop]

http://bailando.berkeley.edu/papers/psb03.pdf

[remkop]

Micronaut does this:

// Apply command name expansion (rA for run-app, tA for test-app etc.)
cmd = commandsByName.values().find() { Command c ->
    ScriptNameResolver.resolvesTo(commandName, c.name)
}

where

class ScriptNameResolver {
    /**
     * Matches a camelCase scriptName to a potential scriptFileName in canonical form.<p>
     * The following scriptNames match FooBar: FB, FoB, FBa
     */
    static boolean resolvesTo(String scriptName, String scriptFileName) {
        def scriptFileNameTokens = NameUtils.getNameFromScript(scriptFileName).findAll(/[A-Z][a-z]+/)
        def scriptNameTokens = NameUtils.getNameFromScript(scriptName).findAll(/[A-Z][a-z]*/)

        if (scriptFileNameTokens.size() != scriptNameTokens.size()) return false
        for (int i = 0; i < scriptNameTokens.size(); i++) {
            String str = scriptNameTokens[i]
            if (!scriptFileNameTokens[i].startsWith(str)) return false
        }
        true
    }
}

and

//NameUtils
public static String getNameFromScript(String scriptName) {
    return getClassNameForLowerCaseHyphenSeparatedName(scriptName);
}
/**
 * Converts foo-bar into FooBar. Empty and null strings are returned as-is.
 *
 * @param name The lower case hyphen separated name
 * @return The class name equivalent.
 */
private static String getClassNameForLowerCaseHyphenSeparatedName(String name) {
    // Handle null and empty strings.
    if (isBlank(name)) {
        return name;
    }

    if (name.indexOf('-') == -1) {
        return name.substring(0, 1).toUpperCase() + name.substring(1);
    }

    StringBuilder buf = new StringBuilder();
    String[] tokens = name.split("-");
    for (String token : tokens) {
        if (token == null || token.length() == 0) {
            continue;
        }
        buf.append(token.substring(0, 1).toUpperCase())
            .append(token.substring(1));
    }
    return buf.toString();
}

[NewbieOrange]

Hi, any idea or suggestion about how this should be implemented? I saw your code snippet in #732, is that the preferred behaviour of this feature?

[remkop]

Yes, the code in #732 is a good start for how I think this should work.

Basically, options and commands can be abbreviated at word boundaries:

• --this-is-some-option can be abbreviated as --t-i-s-o, or --this-i-s or --t, or (without hyphens) as --tISO or --tIS, etc.
• --someCamelCaseOption can be appreviated as --sCCO, or --sC, or --s-c-c-o, etc.

Note that the code in #732 is just a proof of concept to explore this problem. If there is a better (shorter, clearer) way to do this that is fine.

The Map passed to the match function could perhaps also be a Set. I think the logic looks mainly at the keys and returns the (full) key that matches the specified abbreviated key. Once we have the full key, we can then get the CommandLine or OptionSpec from the map in CommandSpec.

Note that the code to loop over the characters in the string is incorrect: see https://stackoverflow.com/questions/1527856/how-can-i-iterate-through-the-unicode-codepoints-of-a-java-string (instead of i++ the variable should be incremented by Character.charCount(codepoint)).

[NewbieOrange]

Created a draft PR #1047.