lang-gen

name	lang-gen
description	Generate a new random_name_generator language from a free-text theme (e.g. "german curse words") — assembles flagged pre/mid/sur syllable collections, registers the File constant, adds a smoke spec and README entry, and samples names to verify. Use when asked to create, generate, or add a new language, dialect, or syllable file for the name generator.

lang-gen

Turn a free-text theme into a fully integrated random_name_generator language file. This workflow is written to be model-agnostic: any assistant that can edit files and run commands can follow it.

The language-file DSL

Authoritative source: lib/random_name_generator/syllable.rb. Each line of a language .txt is one syllable:

• Leading - → prefix (pre) bucket — a name beginning.
• Leading + → suffix (sur) bucket — a name ending.
• No sign → middle (mid) bucket — an interior fragment.

Optional whitespace-separated flags after the syllable constrain adjacency:

• +v — the NEXT syllable must start with a vowel.
• +c — the NEXT syllable must start with a consonant.
• -v — THIS syllable may only follow one ending in a vowel.
• -c — THIS syllable may only follow one ending in a consonant.

A name needs all three buckets non-empty. Generator#compose picks a prefix, fills with middles, ends with a suffix, honoring the flags.

Workflow

Follow these steps in order. Do not skip verification.

1. Parse the theme and derive names

From the argument string derive and then show the user for confirmation:

• slug — concise kebab-case, 1–2 meaningful words, drop filler ("words", "generator", articles). german curse words → german-curse.
• constant — SCREAMING_SNAKE of the slug. german-curse → GERMAN_CURSE.
• destination — edgy themes (curse, insult, swear, NSFW, drunk, etc.) → lib/languages/experimental/<slug>.txt; tame themes → lib/languages/<slug>.txt.

Before writing anything, check for collisions: if the target .txt OR the constant already exists, STOP and ask the user how to proceed. Never overwrite.

2. Author the syllable file (hybrid: seed, then extend)

1. Recall a handful of real roots for the theme. (German curses: schei, arsch, sau, verdammt, dreck, mist, depp, blöd, …)
2. Slot fragments into buckets: prefixes (-) = natural beginnings, middles (bare) = connective interiors, suffixes (+) = natural endings / intensifiers.
3. Extend each bucket with same-flavor invented syllables until each holds ~20–40 entries (match lib/languages/goblin.txt scale).
4. Add +v/+c/-v/-c flags only where they prevent awkward clusters — e.g. a consonant-final prefix that would collide, or a lone vowel that reads badly after another vowel. Most lines carry no flag. Keep it judicious.

Write the file to the destination from step 1.

3. Register the constant

Edit lib/random_name_generator.rb. Add, next to the matching group of constants (edgy → under the # Experimental comment; tame → with the curated constants):

CONSTANT_NAME = File.new("#{dirname}/languages/<relative/path>.txt")

Do NOT touch flip_mode or flip_mode_cyrillic — those arrays are curated.

4. Add a smoke spec

Add an example to spec/random_name_generator_spec.rb following existing conventions (see the with languages context). It must assert the constant is a File, its path includes the expected filename, and that a Generator built from it composes a non-empty String:

it "when it has a <Theme> language" do
  expect(RandomNameGenerator::CONSTANT_NAME).to be_a(File)
  expect(RandomNameGenerator::CONSTANT_NAME.path).to include("languages/<relative/path>.txt")
  name = RandomNameGenerator::Generator.new(RandomNameGenerator::CONSTANT_NAME).compose
  expect(name).to be_a(String)
  expect(name).not_to be_empty
end

5. Add a README entry

Add the new language to README.md. Edgy/experimental languages go under a short "Experimental languages" note; tame ones join the curated list. Link to the file, matching how the curated four are listed.

6. Verify — parse-check then sample

• Parse-check: confirm every line is a valid syllable and all three buckets are non-empty. Fastest via the gem:

  bundle exec ruby -Ilib -rrandom_name_generator -e '
    g = RandomNameGenerator::Generator.new(RandomNameGenerator::CONSTANT_NAME)
    abort "empty pre"  if g.pre_syllables.empty?
    abort "empty mid"  if g.mid_syllables.empty?
    abort "empty sur"  if g.sur_syllables.empty?
    puts "buckets OK: #{g.pre_syllables.size}/#{g.mid_syllables.size}/#{g.sur_syllables.size}"
  '
  

• Sample: print ~10 composed names for the user to eyeball:

  bundle exec ruby -Ilib -rrandom_name_generator -e '
    g = RandomNameGenerator::Generator.new(RandomNameGenerator::CONSTANT_NAME)
    10.times { puts g.compose }
  '
  

• Run the spec: bundle exec rspec spec/random_name_generator_spec.rb

If names look broken or monotonous, iterate on the .txt (more variety, better flags) and re-sample before finishing.

7. Hand off the commit

Git is manual in this workflow. Do NOT run git yourself. Surface the exact command for the user:

git add lib/languages README.md lib/random_name_generator.rb spec/random_name_generator_spec.rb && git commit -m "Add <Theme> language via /lang-gen"

Guardrails

• Never overwrite an existing language file or constant without confirmation.
• Never run state-changing git commands — always suggest them.
• Keep bucket sizes and flag style consistent with existing lib/languages/*.txt.