docs(memory): replace implementation details with functional diagrams…

[sbharatjoshi]

Description

Major documentation overhaul across the memory module — replaces internal Java implementation details with product-focused functional diagrams, architecture tables, and behavior descriptions.

Documentation (26 files)

All memory module docs have been refactored to focus on what the system does rather than how it's coded:

• Replaced Java code snippets and class references with Mermaid diagrams showing data flow and component interactions
• Replaced implementation-detail sections with functional behavior tables (inputs → processing → outputs)
• Retained mathematical equations where relevant (decay functions, scoring formulas) but removed their Java implementations
• Updated: cortex, hebbian, synapse, dopamine, habituation, inhibition, interference, prospective, focus-mode, lateral-retrieval, importance-fusion, sync, performance, wal-design, panama-design, cognitive-profiles, metamemory, hippocampus, amygdala, scoring-pipeline
• Updated ancillary docs: why-spector, roadmap, mcp-integration, sdk-usage/mcp-server
• Updated module READMEs: spector-memory, spector-node, spector-mcp

Runtime & Node changes

• SpectorRuntime: Changed default OLLAMA_CHAT_MODEL fallback from qwen3:0.6b to tag-extractor:latest — the 0.6B model cannot reliably produce structured ENTITY/RELATION output for the entity extraction prompt, causing all graph edges to default to RELATED_TO
• EventStreamEndpoint: SSE streaming improvements
• MemoryService: Entity graph and memory service updates
• DefaultSpectorMemory: Minor internal updates

Deploy

• Updated deploy scripts (deploy.ps1, deploy.sh, entrypoint.sh) with model configuration improvements

Related Issue

N/A — Ongoing documentation and deployment quality improvements

Type of Change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Performance improvement (change that improves throughput or latency)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Documentation update

Module(s) Affected

• spector-core (SIMD kernels)
• spector-storage (Panama storage)
• spector-index (HNSW / BM25)
• spector-query (query orchestration)
• spector-engine (engine facade)
• spector-node (REST API)
• spector-bench (benchmarks)

Checklist

• My code follows the code style of this project
• I have added Javadoc for all public classes/methods
• I have added tests to cover my changes
• All new and existing tests passed (mvn test)
• No hardcoded secrets or credentials are included
• JMH benchmark results included (if performance-related)

Note: 1 flaky test failure in PerformanceBenchmarkTest.p1_reverseIndexIsConstantTime — timing assertion (1284ns vs 1000ns threshold) fails on loaded systems. Pre-existing, not related to this PR.