Find similarities and differences in object APIs by tracking patterns.
See where your objects' methods are following common patterns and where they are not.
Install the gem and add to the application's Gemfile by executing:
$ bundle add klee
If bundler is not being used to manage dependencies, install the gem by executing:
$ gem install klee
class Something
# define methods that might match some of the patterns
end
patterns = Klee.patterns do
prefix("has_")
prefix("verify_")
infix("_in_")
suffix("_value?")
end
gestalt = Klee[Something.new, patterns: patterns]
gestalt.trace(threshold: 6) # threshold for levenshtein distance between unusual method names
puts gestalt.plot
gestalt.trace(threshold: 9) # clear the plot and trace again
puts gestalt.plot
puts gestalt["unusual"] # unusual method namesFind hidden concepts based upon word repetition
class Something
# define methods that have common word parts
end
concept = Klee.object_concepts(Something)
concept[4] #=> Set of words that appear at least 4 times
concept[2] #=> larger Set of words that appear at least 2 times
filtered = Klee.object_concepts(Something, modifiers: %i[fill_in_ hover_over_ _message])
filtered[4] #=> Set of concepts excluding any common modifiersDiscover hidden vocabulary across your entire codebase by analyzing class and method names.
codebase = Klee.scan("app/**/*.rb", "lib/**/*.rb")
# See all concepts ranked by frequency
codebase.concepts.rank
# => {
# "user" => { classes: Set["User", "UserSession"], methods: Set["current_user", "find_user"] },
# "account" => { classes: Set["Account", "AccountManager"], methods: Set["account_balance"] },
# ...
# }
# Drill into a specific concept
codebase.concepts[:account]
# => { classes: Set["Account", "AccountManager"], methods: Set["account_balance", "close_account"] }
# Filter noise with ignore list and threshold
codebase = Klee.scan("app/**/*.rb",
ignore: %i[new create get set find all],
threshold: 3)Discover which objects frequently work together across your codebase.
codebase = Klee.scan("app/**/*.rb", threshold: 3)
# Pairwise co-occurrence (which objects appear together in files)
codebase.collaborators.pairs
# => { ["account", "user"] => 15, ["session", "user"] => 12, ... }
# Method-level co-occurrence (tighter coupling)
codebase.collaborators.pairs(scope: :method)
# => { ["account", "user"] => 8, ... }
# What collaborates with a specific object
codebase.collaborators.for(:user)
# => { "account" => 15, "session" => 12, "order" => 5 }
# Derived clusters (connected components)
codebase.collaborators.clusters
# => [Set["user", "account", "session"], Set["order", "payment", "invoice"]]Expose klee's vocabulary analysis to Claude Code or other MCP clients.
# Install the mcp gem (optional dependency)
gem install mcp
# Run the server
klee-mcpConfigure Claude Code in ~/.claude.json:
{
"mcpServers": {
"klee": {
"command": "klee-mcp"
}
}
}Available tools:
discover_vocabulary- Extract domain language from a codebasefind_concept_clusters- Identify groups of related conceptsexplore_concept- Deep-dive into a specific conceptfind_collaborators- Find objects that work togethercheck_naming_consistency- Detect naming pattern deviationscodebase_summary- High-level vocabulary overview
After checking out the repo, run bin/setup to install dependencies. Then, run rake test to run the tests. You can also run bin/console for an interactive prompt that will allow you to experiment.
To install this gem onto your local machine, run bundle exec rake install. To release a new version, update the version number in version.rb, and then run bundle exec rake release, which will create a git tag for the version, push git commits and the created tag, and push the .gem file to rubygems.org.
Bug reports and pull requests are welcome on GitHub at https://github.com/saturnflyer/klee. This project is intended to be a safe, welcoming space for collaboration, and contributors are expected to adhere to the code of conduct.
The gem is available as open source under the terms of the MIT License.
Everyone interacting in the Klee project's codebases, issue trackers, chat rooms and mailing lists is expected to follow the code of conduct.