feat: Add semantic grouping for dynamic topic creation in queries

Adds detailed optimization analysis reports for recent LOTUS runs

fix: Update locomo_path and enable helper model with configuration adjustments

feat: Enhance sem_join_with_optional_cascade to handle join statistics and return types

Updates dependencies and config paths for simulation

Adds pluggable retrieval backend config and hashing RM

Refactor code structure for improved readability and maintainability

Refactors pipeline config to YAML-driven and centralizes logic

feat: Enhance fact and entity retrieval with lateral similarity joins for improved accuracy

feat: Add optimization configuration and cascade support for Mem0, EverMemOS and Zep simulations

feat: Add community inference for Zep entities and enhance duplicate/contradiction detection in edge processing

feat: Introduce Mem0 and Zep LOTUS simulations for memory graph processing, update gitignore, and adjust a path in `evermemos_lotus.py`.

feat: Implement EverMemOS semantic queries using LOTUS, relocating the script to `lotus-simulation/` and updating `.gitignore`.

refactor: Move data flow analysis and problem documents into a dedicated `data_flows` subdirectory.

feat: Document semantic queries for Zep, Mem0, and EverMemOS, including LOTUS pseudocode examples.