dhowe / rita4j Goto Github PK

gpg: directory '/home/runner/.gnupg' created
gpg: keybox '/home/runner/.gnupg/pubring.kbx' created
gpg: no default secret key: No secret key
gpg: signing failed: No secret key
[INFO] ------------------------------------------------------------------------
[INFO] BUILD FAILURE
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 23.769 s
[INFO] Finished at: 2020-11-18T04:30:35Z
[INFO] ------------------------------------------------------------------------
Error: Failed to execute goal org.apache.maven.plugins:maven-gpg-plugin:1.6:sign (sign-artifacts) on project rita: Exit code: 2 -> [Help 1]
Error:
Error: To see the full stack trace of the errors, re-run Maven with the -e switch.
Error: Re-run Maven using the -X switch to enable full debug logging.
Error:
Error: For more information about the errors and possible solutions, please read the following articles:
Error: [Help 1] http://cwiki.apache.org/confluence/display/MAVEN/MojoExecutionException
Error: Process completed with exit code 1.
##[debug]Finishing: Publish to the Maven Central Repository

We need the superset of all tests for RiScript to be in both Java & JS.
Some will fail, this is OK, just submit a PR (in both repos) anyway.

if the code don't visit (e.g. print to console) this.root.children and the last word's children before calling Node.asTree(), the children will be missing somehow causing problems in calculating the probability

Should be a single .html page with (jquery is fine):

1. 1 div/text-area for input text
2. 1 div (no text-area needed) for output
3. 2 buttons (one for 'load', one for 'execute')
4. One knob called 'temperature' (0-1, default value is 0)
5. One knob called 'speed' (0-1, default value is .5)
6. One pulldown or knob called 'N' (1-5, default value is 3, integers only)

The 'input' text area should have a vertical slider if text is longer than window and should wrap at the width. The output div should auto-wrap at the width of the window. New lines will appear at top (don't worry about this for now)

• ‘load’ can handle local-files only for now (should bring up the finder/explorer)
• ‘execute’ should be disabled if there is no text in the input text-area (top)

https://github.com/dhowe/rita2js/blob/master/src/markov.js#L7

Asking because of test cases such as this one

This one?
https://docs.oracle.com/javase/8/docs/api/org/w3c/dom/Node.html

We need to verify the library can be used on Android as well. I'm not sure the best way to do this with maven, but we certainly don't want to open android studio

please compare to js version

is add to knownIssues by #53
transform will fail when assign and choice is use together, e.g
$a=a\n($a).toUpperCase() and $a=a\n($a | $a).toUpperCase()

this problem cause many fails in Grammar and Riscript

seems that with choice, transform is wrongly added before the choice is done
it might be some problem in the tree...
@dhowe maybe you will want to look into this

more info in the code comment in KnownIssues

in rita.js

kwic: function(word, numWords) {
..
}

int Rita

String[] kwic(String text, String word, Map<String, Object> opts)
or
String[] kwic(String text, int numWords, Map<String, Object> opts)

?

in rita-script but not in Java yet

let rm = new RiTa.RiMarkov(3);
expect(rm.n).eq(3);

Lexicon.java
Parser.java
Grammar.java
Markov.java

https://docs.github.com/en/actions/language-and-framework-guides/publishing-java-packages-with-maven

Migrate new website code from https://github.com/dhowe/RiTa to https://github.com/dhowe/RiTaWeb

See https://github.com/dhowe/RiTa/issues/561

draft already written -- need to make it live

you can import the file inside this zip to your eclipse formatter preferences formatter-prefs.xml.zip

can close when done

Use cases:

• RiTa.env()
• RiTa.conjugate() (with default named args?)

in rita-script
normally posTags will return String Array
RiTa.posTags("He outnumbers us") ===> ["prp", "vbz", "prp"]

and for { inline: true } will return String
RiTa.posTags("asdfaasd", { inline: true }) ===> "asdfaasd/nn"

are we doing the same in RiTa Java? or we should align both to return String Array?

should include java/js/node/android/browser, etc (adapt from the existing version)

in toPhoneArray test

should this be public or private or test in another way?

let result = RiTa._lexicon().toPhoneArray(RiTa._lexicon()._rawPhones("tornado"));     eql(result, ['t', 'ao', 'r', 'n', 'ey', 'd', 'ow'], 'got:' + result);

js testcases: https://github.com/dhowe/rita2js/blob/master/test/tagger-tests.js#L7
We don't need to support array input for RiTa.pos() for java version, right?

• Gradle or Maven (#14)
• logger (SLF4J) (#21)
• update Java (#23)

is third parameter optional?

isRhyme(String word1, String word2, boolean useLTS)
isAlliteration(String word1, String word2, boolean useLTS)

if yes, what is default (true or false)

also check if it occurs in js

• dealing with line breaks like \n
• dealing with rules starting with number, like $1foo
• throw on bad choice
• dealing with Rita function transform e.g. using RiTa.env() in the input -> on hold
• Custom transform using RiTa.addTransform()
• dealing with norep()
• rseq() -> the dont repeat part only
• seq()
• using assignment with symbol in context refer to #64 (grammar-tests)

In js

RiTa.evaluate('I ate the\n$dog.', { dog: 'lab' }, { trace: 0 })

or

let ctx = { bar: { color: 'blue' } };
RiTa.evaluate('$foo=$bar.color\n$foo', ctx, { trace: 0 })

What's the best way to define ctx in java? It could be any data type in js...

LexParser.java is empty at the moment.

in js

Util.shuffle(results, RiTa);

What is the second parameter in JAVA and the usage of it?

*in Lexicon.java

Setup continuous integration (CI) as well as auto-publishing releases on maven/github:

• setup CI to run tests on each commit (probably via jenkins)
• publish initial maven package
• publish initial github package
• create workflow for programmatically publishing to both (or skip maven central)

See docs here: https://docs.github.com/en/free-pro-team@latest/actions/guides/publishing-java-packages-with-maven

• Tokenizer
• Markov (completions(), probability(), probabilities())
• Grammar (on hold)

For each class above, do the following:

1. make sure tests are in sync with JS
2. implement functions to make tests pass
3. ask questions about the best way to do things in Java
4. make sure code compiles and is correctly formatted (using this formatter)

do we need in rita2?
opening a new re.java?

class RE {

  constructor(regex, offset, suffix) {
    this.raw = regex;
    this.regex = new RegExp(regex);
    this.offset = offset;
    this.suffix = suffix || '';
  }

  applies(word) {
    return this.regex.test(word);
  }

  fire(word) {
    //return this.truncate(Utils.trim(word)) + this.suffix;
    return this.truncate(word) + this.suffix;
  }

  analyze(word) {
    return this.suffix != '' && word.endsWith(this.suffix);
  }

  truncate(word) {
    return (this.offset === 0) ? word : word.substr(0, word.length - this.offset);
  }

  toString() {
    return '/' + this.raw + '/';
  }
}

Just a simple test or two for each should be fine:

public static final <T> List<T> randomOrdering(List<T> list);
public static final float[] randomOrdering(float[] arr);
public static final boolean[] randomOrdering(boolean[] arr);
public static final double[] randomOrdering(double[] arr);
public static final int[] randomOrdering(int[] arr);
public static final <T> T[] randomOrdering(final T[] arr);

in rita-script

expect(RiTa.alliterations).to.be.a('function');

for LetterToSound.java

When I'm working on RiScriptTests.java I was trying to find the best way to translate the following in js:
expect(() => RiTa.evaluate('{$a<} foo', { a: 2 }, { silent: 1 })).to.throw();

For some of the throw error tests I have used the following syntax that is available in JUnit 4.12:

@Test(expected = RiTaException.class)
	public void testBadConditionals() {
		Map<String, Object> ctx = opts();
		ctx.put("a", 2);
		RiTa.evaluate("{$a<} foo", ctx, opts("silent", true));
	}

But I'm not sure what's the best method I could use for multiple throw errors asserts within one test case.
Example: https://github.com/cqx931/rita2/blob/master/src/test/java/rita/test/RiScriptTests.java#L633

We are now using JUnit 4.12 and JUnit 5 Jupiter assertions API introduces the assertThrows method for asserting exceptions that seem to be very easy to use. Could we switch to JUnit 5?

Reference:
https://www.baeldung.com/junit-assert-exception

• riscript-tests.js 61 (priority)
• analyzer-tests.js 7
• conjugator-tests.js 4
• grammar-tests.js 15
• lexicon-tests.js 25
• markov-tests.js 25
• rita-tests.js 6
• tagger-tests.js 14
• tokenizer-tests.js 3

Total: 160

• tagger-tests.js (sc)
• tokenizer-tests.js (sc)
• conjugator-tests.js (sc)
• markov-tests.js (sc)
• analyzer-tests.js (dh)
• lex-loader-tests.js (dh)
• lexicon-tests.js (dh)
• riscript-tests.js (dh)
• grammar-tests.js (dh)
• rita-tests.js

I'm trying to find the equivalent for expect(1.5).to.be.closeTo(1, 0.5); in java.
From a quick research I found a good match implemented in Harmcrest: http://hamcrest.org/JavaHamcrest/javadoc/1.3/org/hamcrest/number/IsCloseTo.html

Could we add Harmcrest as another library for the JUnit test in java?

[ERROR] Tests run: 26, Failures: 0, Errors: 1, Skipped: 0, Time elapsed: 0.43 s <<< FAILURE! - in rita.test.MarkovTests
[ERROR] rita.test.MarkovTests.callGenerate Time elapsed: 0.099 s <<< ERROR!
rita.RiTaException: [Markov] Exceeded maxAttempts:99
at rita.test.MarkovTests.callGenerate(MarkovTests.java:231)

To generate docs (for now):

1. Right-click on docgen.DocGenerator.java and select 'run-as-application'
2. This should generate the docs in /web/reference

Note: the html here was converted from php so may need some cleanup. It also doesn't include RiGrammar or RiMarkov

We need to do better with the test naming. See example below.

Given the following JS-tests

describe('Symbol',` () => {
    it('Should eval linebreak-defined variables', () => {
       ... 
    });

    it('Should return input for undefined', () => {
       ... 
    });
})

Map to Java tests as follows:

@Test
public void evalLinebreakDefinedVariablesSYMBOL() {
  ... 
}

@Test
public void returnInputForUndefinedSYMBOL() {
  ...
}

I'll work on RiScript and Grammar, you do the others:

See https://github.com/dhowe/rita2/blob/master/src/test/java/rita/test/KnownIssues.java#L30 and https://github.com/dhowe/rita2/blob/master/src/test/java/rita/test/KnownIssues.java#L53

Need to use debugger to figure out what is going on here

I'm working on the test cases for RiTa.addTransform() and I'm not sure how can I translate this from js to java
(s) => s || 'B'
https://github.com/dhowe/rita2js/blob/master/test/riscript-tests.js#L580
Is this translatable in java? Shouldn't input value always match with the predefined type?

I can also use another function here, as the test case is testing addTransform() in context.
Function<String, String> func = (x) -> {return "A";};

We need to verify the library can be used on Android (with and without processing) -- I'm not sure the best way to do this with junit/maven... we may need a separate way to test

use SL4J, as discussed here: philburk/jsyn#83

Should still export 1.8 jar

One option: https://github.com/bsideup/jabel

many duplicate pluralize/singularize tests in here

add everything to handleSingularPluralPairs and remove other pluralize/singularize tests?