Prepare devel/0.2.0 branch for active development#1
Merged
Conversation
…paper, archive v0.1.2 artifacts
Activates ENABLE_LEHMER_PI flag to align with docs/defaults.md section 8. Changes: - config.py:74: ENABLE_LEHMER_PI = False → True - Updated comments to reflect v0.2.0 default status Impact: - Resolves critical contract violation #1 from readiness audit - Enables Tier B guarantees for indices > 250k - Activates O(x^2/3) sublinear π(x) backend per docs/0.2.0/part_6.md Validation: - 43 π(x) tests passed (tests/test_pi.py) - 11 resolve tests passed (tests/test_resolve.py) - 15 API contract tests passed (tests/test_api_contracts.py) - No regressions detected 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Implements tiered Prime Number Theorem refinements as specified in: - docs/0.2.0/part_6.md (forecasting and approximation) - docs/0.2.0/part_2.md (Tier C contract: <0.2% error at n≥10^8) - docs/0.2.0/part_0.md (scalability invariant for n>10^12) Changes: - forecast.py: Added refinement_level parameter (1/2/3) - Level 1: Base PNT → <0.3% error (v0.1.2 compatible) - Level 2: + higher-order term → <0.2% error (v0.2.0 target) - Level 3: Reserved for future (additional terms) - Updated signature: forecast(n: int, refinement_level: int = 1) -> int - Changed rounding from truncation to nearest int for improved accuracy - Added comprehensive type hints and docstring updates Testing: - tests/test_forecast.py: 13 new tests added - Backward compatibility validation (Level 1 = default) - Accuracy bounds verification per Part 6 table - Invalid refinement_level rejection - Determinism across all levels - All 182 tests pass (169 baseline + 13 new) - No regressions in existing functionality Contract Compliance: - Resolves critical contract violation #2 from readiness audit - Tier C error guarantees now met per Part 2 specification - Maintains pure Python stdlib-only requirement Phase Alignment: - Phase 1: Contract Compliance (forecast refinement is second item) - No performance optimizations (Phase 2) - No usability features (Phase 3) - Scope strictly limited per phase plan 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Implements LRU caching for logarithmic computations per Part 1, section 2.1. Changes: - utils.py: Added @lru_cache(maxsize=2048) to log_n() and log_log_n() - Added functools import for caching support - Updated docstrings to document caching behavior Performance Impact (Part 1 targets): - 25-35% reduction in computation time for hot paths - Cache hit rate >95% in typical workloads (n up to 10^12+) - Memory overhead <1 MB (2048 entries × ~16 bytes/entry) - Benefits forecast.py, pi.py, and simulation paths Implementation Details: - Cache size: 2048 entries (tuned per Part 1 recommendation) - Pure functions: input (n: int) → output (float), no side effects - Thread-safe: stdlib lru_cache provides built-in thread safety - Deterministic: same input always yields same cached output Validation: - All 182 tests pass (no behavioral changes) - Test runtime improved: 14.78s vs ~24s baseline (~38% faster) - Cache verification: hits/misses tracking confirmed working - Numerical equivalence: cached results identical to math.log() Phase Alignment: - Phase 2 (Performance & Resource Efficiency), Task 1 - No algorithm changes, pure performance optimization - No dependency on generator mode or annealing (later tasks) - Strictly scoped: only utils.py modified 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
- Add as_generator parameter to simulate() for O(1) memory streaming - Implement _simulate_generator() internal helper function - Preserve determinism: same seed yields identical sequence in both modes - Add 12 new tests validating equivalence, determinism, and memory efficiency - Maintain backward compatibility: default behavior unchanged - Total tests: 194 passing (12 new generator tests) Refs: Phase 2 Task 2, docs/0.2.0/part_1.md §2.2, part_5.md §2.3
- Add anneal_tau parameter for optional β annealing (Part 5 §2.1) - Formula: β_eff(n) = beta * (1 - exp(-n / anneal_tau)) * (beta_decay)^n - Exponential ramp-up from β≈0 early to β≈beta_initial later - Reduces early transient variance and improves convergence stability - Backward compatible: anneal_tau=None preserves exact existing behavior - Determinism preserved: same seed + tau yields identical results - Works in both list mode and generator mode - Add 14 comprehensive tests validating: * Backward compatibility (anneal_tau=None) * Determinism (same seed, list vs generator) * Behavior changes (different tau values) * Input validation (zero, negative, inf, NaN, non-numeric) * Integration with generator and diagnostics modes - Total tests: 208 passing (14 new annealing tests) Refs: Phase 2 Task 3, docs/0.2.0/part_5.md §2.1
- Replace random.choices() with CDF + binary search for O(log k) sampling - Precompute cumulative distribution function (CDF) from gap probabilities - Use bisect_right for O(log k) lookup vs O(k) for random.choices - Maintains exact probability distribution P(g|w,β) semantics - Preserves determinism: same seed yields same sequence - Add 17 comprehensive tests in test_gaps.py validating: * Sampling determinism with fixed seed * Distribution coverage and probability correctness * CDF properties (strictly increasing, ends at 1.0) * Tilted distribution properties (normalized, positive weights) * Simulation determinism (list vs generator, with annealing) * Empirical distribution properties - Total tests: 225 passing (17 new gap tests) - No backward compatibility requirement per Part 2 (Tier C Statistical) - Determinism preserved within new implementation Performance improvement: O(k) → O(log k) per sample, ~200 gaps typical Refs: Phase 2 optimization, docs/0.2.0/part_5.md §2.2, part_1.md §2
- Mark Phase 2 (Performance) as COMPLETE in part_9.md - Add Phase tracking section with all 4 Phase 2 tasks documented - Update CHANGELOG with Phase 2 performance optimizations - Mark Phase 3 (Usability) as ACTIVE - Document metrics: 56 new tests (169→225), 20-60% faster, 75% memory reduction Phase 2 completed tasks: 1. Log caching (commit 369e8a4) 2. Generator mode (commit 2b640fd) 3. Dynamic β annealing (commit 32f36ca) 4. CDF gap sampling (commit 05342f5) Ready to begin Phase 3, Task 1: Minimal CLI interface
Implements Phase 3, Task 1: Minimal CLI using argparse (stdlib only). Commands implemented: - lulzprime resolve <n>: Compute exact nth prime p_n - lulzprime pi <x>: Compute π(x), count of primes <= x - lulzprime simulate <n_steps>: Generate pseudo-primes via OMPC - Flags: --seed, --anneal-tau, --generator Features: - Human-readable text output (one value per line) - Strict argument validation with user-friendly errors - Deterministic behavior (seed-based for simulate) - Entry point: python -m lulzprime <command> Testing: - 15 new CLI tests (subprocess-based validation) - All 240 tests passing (225 existing + 15 CLI) - Coverage: argument parsing, error handling, determinism Phase 3 (Usability) - Task 1 complete. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Implements Phase 3, Task 2: JSON export support for simulation output. New Functions: - simulation_to_json(): Returns JSON-serializable dict - simulation_to_json_string(): Returns deterministic JSON string Schema (lulzprime.simulation.v0.2): - params: All simulation parameters (n_steps, seed, anneal_tau, etc.) - sequence: List of pseudo-prime integers - diagnostics: Optional diagnostic records (or null) - meta: Library name, version, timestamp (null for determinism) Features: - Stdlib-only (uses json module) - Deterministic output (sort_keys=True, no timestamps) - Supports all simulation modes (list, generator, diagnostics, annealing) - CLI integration: --json flag for simulate command Testing: - 18 new JSON export tests in test_json_export.py - 1 new CLI test for JSON export - All 258 tests passing (240 existing + 18 new) CLI Usage: python -m lulzprime simulate 100 --seed 42 --json output.json Python Usage: seq = simulate(100, seed=42) json_data = simulation_to_json(seq, n_steps=100, seed=42) Phase 3, Task 2 complete. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
…iers) Updates user-facing documentation to reflect current v0.2.0 feature set. README.md: - Added CLI Quickstart section with all commands (resolve, pi, simulate) - Added Python API examples for new features: - forecast(n, refinement_level=2) usage and benefits - simulate() with as_generator=True and anneal_tau parameters - simulation_to_json() and simulation_to_json_string() examples - Clarified simulation Tier C guarantees: - Deterministic with seed (same seed → same sequence) - Statistical correctness (not exact primes) - No cross-implementation identity guarantee - Updated test count: 169 → 258 (225 core + 15 CLI + 18 JSON export) - Updated project structure (docs/0.2.0/ current, docs/0.1.2/ historical) - Updated status: v0.1.2 (v0.2.0 in development) docs/0.2.0/part_8.md: - Documented CLI commands with all flags and options - Documented JSON export schema (lulzprime.simulation.v0.2): - schema, params, sequence, diagnostics, meta fields - Deterministic output (sorted keys, no timestamps) - Added comprehensive CLI usage examples - Updated configuration defaults section docs/0.2.0/part_6.md: - Expanded forecast() refinement_level documentation - Explained when to use refinement_level=2 (n >= 10^8) - Added error bound comparisons (Level 1 vs Level 2) - Added usage examples with actual error rates docs/defaults.md: - Updated test count target (200+ → 258 current) - Noted CLI availability via python -m lulzprime - Noted JSON export support (Phase 3 Task 2) - Noted ENABLE_LEHMER_PI = True in v0.2.0 No code changes. All 258 tests passing. Phase 3 Task 3 (Documentation) Push #1 complete. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Add strict mypy configuration and fix all typing issues across the codebase: - Updated pyproject.toml with strict mypy settings (Python 3.10 baseline) - Fixed 17 typing errors across 5 modules (cli, lehmer, gaps, pi, lookup) - Added type annotations for function signatures and memoization caches - Used typing.cast for complex union types in CLI - Added mypy step to CI workflow after pytest All 258 tests pass. No runtime behavior changes. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
This PR establishes the
devel/0.2.0branch as the protected long-lived development branch for v0.2.0 while preservingmainas the immutable v0.1.2 release.Key structural changes:
archive/0.1.2/docs/docs/manual/→docs/0.1.2/(historical v0.1.2 development manual)docs/0.2.0/with updated development manual (Parts 0–9)docs/paper/as canonical location for OMPC paperdocs/defaults.mdanddocs/autostart.mdto reference new structure.gitignoreto allow trackingdocs/paper/while ignoring other paper foldersAll future v0.2.0 work will occur on the
devel/0.2.0branch and its feature branches.The
mainbranch remains the pristine v0.1.2 release.Repository State After Merge
mainbranch: Immutable v0.1.2 release (tagged)devel/0.2.0branch: Active development for v0.2.0docs/0.1.2/(Parts 0–9 from v0.1.2)docs/0.2.0/(Parts 0–9 for v0.2.0)docs/paper/OMPC_v1.33.7lulz.pdfarchive/0.1.2/docs/🤖 Generated with Claude Code