feat: add thread safety to Bayes and LSI classifiers

[cardmagic]

Summary

• Add Mutex_m inclusion to both Bayes and LSI classifiers
• Wrap all public methods accessing shared state in synchronize blocks
• Add marshal_dump/marshal_load for serialization (Mutex can't be marshaled)

Problem

Concurrent train() and classify() calls corrupt internal state:

classifier = Classifier::Bayes.new('Spam', 'Ham')

threads = 10.times.map do
  Thread.new do
    100.times do
      classifier.train_spam("buy now cheap")
      classifier.classify("special offer")
    end
  end
end

threads.each(&:join)  # Corrupted state

Solution

Use the Mutex_m mixin (already a gem dependency) with a pattern that avoids deadlock:

def public_method(...)
  synchronize { private_unlocked_method(...) }
end

private

def private_unlocked_method(...)
  # actual implementation
end

Test plan

• All existing tests pass with bundle exec rake test
• All tests pass with NATIVE_VECTOR=true bundle exec rake test
• Marshal serialization still works (test_serialize_safe)

Closes #66

[Copilot]

Pull request overview

This PR adds thread safety to the Bayes and LSI classifiers by incorporating the Mutex_m mixin and wrapping all public methods that access shared state in synchronize blocks. The implementation follows a pattern that prevents deadlock by creating unlocked private versions of methods that are called within synchronized blocks.

Key Changes:

• Added Mutex_m inclusion and initialization to both Bayes and LSI classes
• Wrapped all public methods accessing shared state in synchronize blocks
• Implemented custom marshal_dump/marshal_load methods to handle serialization (Mutex objects cannot be marshaled)

Reviewed changes

Copilot reviewed 2 out of 3 changed files in this pull request and generated no comments.

File	Description
lib/classifier/lsi.rb	Added thread safety via Mutex_m, refactored methods into public synchronized versions and private unlocked versions, implemented custom marshaling
lib/classifier/bayes.rb	Added thread safety via Mutex_m, wrapped state-accessing methods in synchronize blocks, implemented custom marshaling

Comments suppressed due to low confidence (1)

lib/classifier/bayes.rb:1

• Missing closing end for the synchronize block. The block opened on line 193 is not properly closed before the method ends on line 202.

# rbs_inline: enabled

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 2 out of 3 changed files in this pull request and generated 2 comments.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Calling mu_initialize directly may not properly initialize the mutex state. Consider using super() or the appropriate Mutex_m initialization method to ensure the mutex is correctly set up during deserialization.

Suggested change

	mu_initialize
	@mu = Mutex.new

[Copilot]

Calling mu_initialize directly may not properly initialize the mutex state. Consider using super() or the appropriate Mutex_m initialization method to ensure the mutex is correctly set up during deserialization.

Suggested change

	mu_initialize
	@mutex = Mutex.new

[cardmagic]

Regarding the Copilot suggestion to replace mu_initialize with @mu = Mutex.new:

The current implementation using mu_initialize is correct. This is the proper Mutex_m API for reinitializing the mutex state after deserialization. The suggested @mu = Mutex.new or @mutex = Mutex.new would not work because:

1. Mutex_m uses internal instance variables (not @mu or @mutex)
2. The synchronize method relies on state set up by mu_initialize

Verified that serialization/deserialization works correctly:

b = Classifier::Bayes.new('Spam', 'Ham')
b.train_spam('buy now')
loaded = Marshal.load(Marshal.dump(b))
loaded.train_spam('test')  # Works correctly