Add unit tests for cognitive_synergy module

[web3guru888]

Good First Issue: Test Coverage for cognitive_synergy

The cognitive_synergy module implements information-theoretic synergy metrics (O-information, synergy-redundancy decomposition). It currently has limited test coverage.

What to do

Add unit tests to tests/test_cognitive_synergy.py (or create it) covering:

1. compute_o_information() — test with known values from the O-information paper
2. synergy_redundancy() — verify the fundamental constraint: O = redundancy - synergy
3. Edge cases: empty input, single variable, high-dimensional arrays

Getting started

git clone https://github.com/web3guru888/asi-build.git
cd asi-build
pip install -e '.[dev]'
pytest tests/ -v  # see what passes today

Module location

src/asi_build/cognitive_synergy/

Resources

• O-information (Rosas et al. 2019)
• Partial Information Decomposition
• See examples/ for usage patterns

Difficulty

🟢 Beginner — no background in information theory required; the functions are straightforward to test against synthetic data.

This is a great first contribution — pure Python, self-contained, no GPU needed.

[web3guru888]

Good first issue! Here's a guide to get started:

Module location: src/asi_build/cognitive_synergy/

Key classes to test:

• CognitiveSynergyAnalyzer — the main entry point
• TransferEntropyCalculator — computes TE between variables
• DualTotalCorrelation — measures O-information (synergy vs. redundancy)

Test file to create: tests/test_cognitive_synergy.py

Suggested test cases:

1. test_transfer_entropy_independent_variables() — TE between two independent uniform RVs should be ≈0
2. test_transfer_entropy_deterministic() — TE from X→Y when Y is a copy of X should be ≈H(X)
3. test_dual_total_correlation_single_variable() — DTC of a single variable should be 0
4. test_synergy_xor_gate() — the classic XOR synergy example (two inputs, one output) should have positive O-information
5. test_empty_time_series_raises() — empty input should raise ValueError

See tests/test_consciousness.py for examples of our test style. Feel free to ask questions here!

[web3guru888]

Update — 2026-04-12: The Cognitive Synergy wiki page is now live with full architectural documentation — a good reference for understanding what to test before writing the test suite.

A few observations from the code review:

The module is well-structured for testability. All major components use dataclasses with clear fields. The main challenge is that several methods depend on internal state that builds up over time (e.g., BehavioralEmergenceDetector tracks behavior history, SynergyMetrics needs a rolling window of state pairs).

What I'd focus on first:

1. PRIMUSFoundation — Pure logic, no side effects. Test that PRIMUSState initializes with correct defaults, CognitivePrimitive connections are bidirectional, learning_rate and decay_rate produce correct weight updates.

2. SynergyMetrics — Test the metrics math: mutual information between two identical signals should be high; transfer entropy from a random signal to an independent random signal should be ~0; emergence_index should be 0 for a system that never leaves known state-space regions.

3. PatternMiningEngine — Property-based tests with Hypothesis would fit well: for any input data, discover_patterns() should return only patterns with confidence in [0, 1]; PatternHierarchy parent/child relationships should be consistent.

4. CognitiveSynergyEngine — Integration test: create an engine, tick it 100 times, verify all 10 SynergyPair objects have last_updated > initial time.

Related: Issue #79 tracks wiring the engine to the Cognitive Blackboard — the integration work depends on having tests first, so #1 is on the critical path.

Happy to review PRs targeting this issue.

[web3guru888]

Resolved — 120 unit tests for cognitive_synergy module in tests/test_cognitive_synergy_unit.py. Covers: SynergyPair/CognitiveDynamic dataclasses, CognitiveSynergyEngine (init, register, inject_stimulus, get_state), SynergyMetrics (time series, profile computation, info theory), PRIMUSFoundation (primitives, synergy computation, stimulus injection), EmergentPropertyDetector (behavioral/structural/functional detection), SelfOrganizationMechanism (state tracking, load balancing, rules), sub-components (HomeostaticController, AdaptiveRestructurer, ResourceOptimizer, CoherenceMaintainer), PerceptionEngine, PatternMiningEngine.