PyInj v1.2.0 - Performance & Reliability

PyInj v1.2.0 Release Notes

Release Date: 2025-09-14

Overview

PyInj v1.2.0 brings significant performance improvements, critical bug fixes, and enhanced monitoring capabilities. This release focuses on production reliability with O(1) circular dependency detection, proper memory management, and correct transient scope semantics.

🚀 Performance Improvements

O(1) Circular Dependency Detection

• Before: O(n²) complexity using tuple concatenation
• After: O(1) complexity using set-based tracking
• Impact: Detection time remains constant even with 1000+ dependency depth
• Benchmark: < 100ms for 1000-depth chains (previously would timeout)

Memory Leak Fixes

• Singleton Lock Cleanup: Locks are now properly removed after initialization
• Before: Unbounded growth with defaultdict(threading.Lock)
• After: Locks cleaned up immediately after successful singleton creation
• Impact: Long-running applications no longer accumulate memory

Transient Scope Correctness

• Fixed: Removed incorrect WeakValueDictionary caching
• Before: Transients were incorrectly cached, violating scope semantics
• After: Each resolution creates a new instance as expected
• Impact: Proper garbage collection and predictable behavior

🐛 Bug Fixes

1. Transient Scope Violation - Transients now correctly create new instances every time
2. Singleton Lock Memory Leak - Proper cleanup prevents memory accumulation
3. Dead Code Removal - Eliminated unused dependency tracking code

✨ New Features

Batch Operations

# Register multiple dependencies at once
container.batch_register([
    (TOKEN_A, provider_a),
    (TOKEN_B, provider_b),
    (TOKEN_C, provider_c),
])

# Resolve multiple dependencies efficiently
results = container.batch_resolve([TOKEN_A, TOKEN_B, TOKEN_C])

# Async batch resolution with parallel execution
results = await container.batch_resolve_async([TOKEN_A, TOKEN_B, TOKEN_C])

Performance Monitoring

# Get container statistics
stats = container.get_stats()
# Returns: {
#     'total_providers': 150,
#     'singletons': 45,
#     'cache_hits': 1200,
#     'cache_misses': 150,
#     'cache_hit_rate': 0.889,
#     'avg_resolution_time': 0.000004
# }

# Direct cache hit rate property
hit_rate = container.cache_hit_rate  # 0.0 to 1.0

Contextual Overrides

# Temporarily override dependencies in a context
with container.use_overrides({LOGGER: fake_logger, DB: test_db}):
    service = container.get(SERVICE)
    # service uses fake_logger and test_db within this context
# Original dependencies restored outside the context

📊 Testing Improvements

• 31 new tests added for comprehensive coverage
• 7 memory profiling tests using tracemalloc
• 10 cycle detection tests validating O(1) performance
• 4 performance benchmark tests
• Total: 156 tests, all passing

📚 Documentation

• Cleaned up code comments: Removed redundant inline comments, keeping only non-obvious explanations
• Updated API documentation: Added all v1.2.0 features with examples
• Performance documentation: Real benchmarks and complexity analysis
• CHANGELOG: Comprehensive change history

💥 Breaking Changes

None. This release maintains full backward compatibility with v1.1.x.

🔄 Migration Guide

No migration required. Simply update your dependency:

# Using pip
pip install --upgrade pyinj==1.2.0

# Using uv
uv pip install --upgrade pyinj==1.2.0

# Using poetry
poetry add pyinj@^1.2.0

🎯 Highlights for Production Use

1. Predictable Performance: O(1) operations ensure consistent response times
2. Memory Safe: No memory leaks in long-running applications
3. Correct Semantics: Transient scope behaves as documented
4. Observable: Built-in performance monitoring for production insights
5. Thread-Safe: Proper lock management without memory overhead

🙏 Acknowledgments

Thanks to all contributors and users who reported issues and provided feedback for this release.

📦 Installation

pip install pyinj==1.2.0

🔗 Links

• GitHub Repository
• PyPI Package
• Documentation
• Changelog

---

Full Changelog: v1.1.1...v1.2.0

PyInj 1.1.1

PyInj 1.1.1

Improvements and fixes since 1.1.0:

• CI gates: format→lint→types, then tests and docs
• PR head checkout; avoid duplicate runs on PR branches; concurrency cancel-in-progress
• Pytest: coverage (pytest-cov), junit xml, xdist; Codecov upload
• README: status badges
• Cleanup: ignore/remove site/ build output; remove local runner/type logs; ignore .basedpyright_full.txt
• Internals: defaults indirection, protocols split, typing fixes for wrappers and context cleanups

PyInj 1.1.0

PyInj 1.1.0 (stable)

Highlights:

• register_context (+ sync/async helpers) with per-scope cleanup
• immutable registrations
• clarified cleanup model; nested request scopes; async cleanup ordering
• LLM usage guide

Changelog:

• feat(container): add register_context for sync/async context-managed providers with per-scope cleanup (request/session/singleton)
• feat(container): enforce immutable registrations; re-registering the same token raises ValueError
• feat(api): add register_context_sync and register_context_async helpers; typed overloads on register_context
• refactor(cleanup): remove broad introspective resource tracking; container now tracks resources registered via context-managed registration while request/session scopes continue to clean up scoped resources
• docs: update examples to use register_context_sync/async; clarify nested request scopes and async cleanup ordering
• docs: add LLM Guide for concise usage in agent/tool contexts
• tests: refactor to context-managed registrations; fix asyncpg/httpx test indentation; update circular dependency expectation