How antonio-cruciani/dreg-dynamicrandomexpandergenerator works
Project type and purpose
• This is a Python-based simulation framework (an open-source tool) for generating and analyzing dynamic random graphs under different models (Edge Dynamics, Vertex Dynamics, mixed models). It is driven by configuration files and supports both single-process and multiprocessing simulations.
• Users define P2P protocols (RAES, ERAES, VRAES, EVRAES) and experiment parameters in INI files, then run CLI scripts to produce adjacency lists, spectral statistics, flooding results, and other metrics.
File-structure analysis
• Top-level scripts:
• main.py – entry point for ERAES/VRAES simulations
• mixedDynamic.py – entry point for mixed EVRAES with multiprocessing
• ERAES… / EVRAES… – convenience scripts for variants
• to_plot_eraes.py, write_properties.py – post-processing/plotting
• Configuration:
• config.ini – master INI defining model and simulation params
• configurations/EdgeDynamics, VertexDynamics – sample INIs for reproduction
• Core code under src/:
• Algorithms/ – model implementations (EdgeDynamic.py, VertexDynamic.py)
• Graphs/ – custom graph data structures (MultipleEdge, Queue)
• MathModules/ – numeric utilities (MathTools.py)
• Protocols/ – P2P protocol logic (Consensus.py, FloodOBJ.py)
• FileOperations/ – persistence (WriteOnFile.py)
• StastModules/ – statistics & analysis (GraphAnalysis, SpectralAnalysis, Snapshot)
• tail_estimation_degree_distribution – tail-estimation script
Key README insights
• Distributed protocols run at each time step to maintain bounded node degree.
• External events (edge/node disappearance) are stochastic.
• Supports “offline” mode (dumping adjacency lists) and “only spectral” mode.
• Multiprocessing for mixed dynamics via mixedDynamic.py.
a. Main components:
1. Input Layer:
– User INI files (config.ini, model-specific INIs)
2. Orchestrator:
– CLI scripts (main.py, mixedDynamic.py)
– Configuration parser (reads INI, builds parameter sets)
3. Simulation Engine:
– Model Manager (instantiates EdgeDynamic or VertexDynamic classes)
– Protocol Executor (Consensus, FloodOBJ)
– Event Generator (internal random falls/join events)
4. Core Modules:
– Algorithms (model logic per time step)
– Graph Data Structures (Graph, Node, Edge handling)
– Math Utilities (epsilon estimation, Poisson rates)
5. Analysis & Output:
– Statistics Modules (GraphAnalysis, SpectralAnalysis, Snapshot)
– FileOperations (writes adjacency, metrics, logs)
– Post-processing scripts (tail estimation, plotting)
6. Multiprocessing Layer (for EVRAES mixedDynamic.py)
b. Relationships and data flow:
– Config files → Config parser → Parameter set
– Orchestrator → Simulation Engine → Model instance
– Each simulation round:
• Model instance invokes Protocol Executor
• Graph object updated via Algorithms
• MathTools invoked for convergence checks
• Events generated by Event Generator
– After each round or at end:
• StastModules compute metrics
• FileOperations flush data to disk
– Post run:
• to_plot_eraes.py / tail-estimation.py read output and generate visuals
c. Architectural patterns / principles:
– Layered architecture: separation of CLI, core engine, data structures, math, analysis
– Modular design: each functionality in its own package under src/
– Pipeline pattern: config → simulation → analysis → output → plotting
d. Key technologies and frameworks:
– Python 3.x
– ConfigParser (for INI parsing)
– Multiprocessing (built-in)
– Custom graph and protocol classes (no external graph library)
– Standard math, random, file I/O modules
Details to capture
• The dual entry points (main.py vs mixedDynamic.py) and when each is used.
• The role of each src/ subdirectory.
• How external events (edge/node falls) plug into the main loop.
• Where and how users can plug in new protocols or analysis modules.