# baitingzbt/peda architecture

> Architecture diagram of the GitHub repository baitingzbt/peda, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2025-07-30). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/baitingzbt/peda
- Repository: https://github.com/baitingzbt/peda

## Overview

##### Project Type and Purpose
- This is an academic research codebase (“PEDA”) for offline multi-objective reinforcement learning experiments, not a classic full-stack app. It provides (a) dataset generation pipelines, (b) environment wrappers for MuJoCo tasks, (c) two model implementations (Decision Transformer “modt” and RVS “rvs”), and (d) an experiment orchestrator to run training & evaluation.

##### High-Level Components
a. Environment Layer
- Directory “environments/” and “data_generation/environments/”: XML assets + Python wrappers for MuJoCo tasks (Ant, HalfCheetah, Hopper, Swimmer, Walker2d, Humanoid).
b. Data Generation Pipeline
- “data_generation/”: scripts to collect and preprocess trajectories.
- Sub-folders “externals/baselines” and “externals/pytorch-a2c-ppo-acktr-gail”: embed open-source RL libraries to generate or replay trajectories.
c. Offline Dataset Storage
- Generated data lands in “data/” (pickled trajectories) and “data_generation/dummy”.
d. Model Implementations
- “modt/”: decision transformer code—models, trainers, evaluation, utilities.
- “rvs/”: return‐conditional VAE style model—data loading, policy definitions, training, analysis, visualization.
e. Experiment Orchestrator
- “experiment.py” + shell scripts (“all_env_uniform.sh”): parse CLI args, set up paths, launch training runs.
f. Pretrained Models & Outputs
- “lr_models/”: precomputed expert models for dataset generation.
- Logging directories under “experiment_runs/” (not listed) or specified output folders.

##### Relationships & Data Flow
- Start: Environment wrappers ← MuJoCo XML assets
↓
- Data Generation Module (scripts invoke external libraries baselines / pytorch)
↓
- Offline Dataset (“data/… .pkl”)
↓
- Two Training Pipelines in Parallel:
- modt/training → modt/models → modt/evaluation
- rvs/src/rvs/train.py → rvs/src/rvs/policies, dataset
↓
- Evaluation & Visualization modules produce metrics & plots
↓
- experiment.py orchestrates runs, collects logs, saves trained checkpoints

##### Architectural Patterns & Principles
- **Pipeline Pattern**: clear stages (data generation → storage → training → evaluation).
- **Modularity**: separate components for each model type (modt vs. rvs) and for environment wrappers vs. dataset code.
- **Extensibility**: “externals/” allows swapping in other RL libraries; one can plug new envs, new models.

## Diagram (Mermaid)

```mermaid
flowchart TD
    subgraph "Conda Environment"
    end

    subgraph "Environment Layer" 
        EnvWrappers["Environment Wrappers & Assets"]:::codeModules
    end

    PretrainedModels["Pretrained Models (lr_models/*.pkl)"]:::dataArtifacts

    subgraph "Data Generation Pipeline"
        DataGenCore["collect_env_new.py"]:::codeModules
        DataGenScripts["collect_all.sh"]:::codeModules
        ExternalsBaselines["baselines"]:::externals
        ExternalsPyTorch["pytorch-a2c-ppo-acktr-gail"]:::externals
    end

    DatasetStorage["Offline Dataset Storage (data/ & dummy)"]:::dataArtifacts

    subgraph "Decision Transformer (modt)"
        ModtTraining["modt/training/trainer.py"]:::codeModules
        ModtEval["modt/evaluation/evaluate_episodes.py"]:::codeModules
    end

    subgraph "Return-Conditional VAE (RVS)"
        RvsTrain["rvs/src/rvs/train.py"]:::codeModules
        RvsEval["rvs/src/rvs/visualize.py"]:::codeModules
    end

    subgraph "Experiment Orchestrator"
        ExperimentCLI["experiment.py"]:::codeModules
        ExperimentScript["all_env_uniform.sh"]:::codeModules
    end

    %% Data Flow
    EnvWrappers -->|"provide env & assets"| DataGenCore
    PretrainedModels -->|"expert models"| DataGenCore
    DataGenCore -->|"generate trajectories"| DatasetStorage
    DataGenScripts -->|"batch collect"| DatasetStorage
    ExternalsBaselines -->|"use baselines"| DataGenCore
    ExternalsPyTorch -->|"use PyTorch externals"| DataGenCore

    DatasetStorage -->|"load dataset"| ModtTraining
    ModtTraining -->|"train DT model"| ModtEval
    DatasetStorage -->|"load dataset"| RvsTrain
    RvsTrain -->|"train RVS model"| RvsEval

    ExperimentCLI -->|"orchestrate pipeline"| DataGenCore
    ExperimentCLI -->|"orchestrate pipeline"| ModtTraining
    ExperimentCLI -->|"orchestrate pipeline"| RvsTrain
    ExperimentScript -->|"launch uniform runs"| ExperimentCLI

    %% Click Events
    click EnvWrappers "https://github.com/baitingzbt/peda/tree/master/environments/"
    click PretrainedModels "https://github.com/baitingzbt/peda/tree/master/lr_models/"
    click DataGenCore "https://github.com/baitingzbt/peda/blob/master/data_generation/collect_env_new.py"
    click DataGenScripts "https://github.com/baitingzbt/peda/blob/master/data_generation/collect_all.sh"
    click ExternalsBaselines "https://github.com/baitingzbt/peda/tree/master/data_generation/externals/baselines/"
    click ExternalsPyTorch "https://github.com/baitingzbt/peda/tree/master/data_generation/externals/pytorch-a2c-ppo-acktr-gail/"
    click ModtTraining "https://github.com/baitingzbt/peda/blob/master/modt/training/trainer.py"
    click ModtEval "https://github.com/baitingzbt/peda/blob/master/modt/evaluation/evaluate_episodes.py"
    click RvsTrain "https://github.com/baitingzbt/peda/blob/master/rvs/src/rvs/train.py"
    click RvsEval "https://github.com/baitingzbt/peda/blob/master/rvs/src/rvs/visualize.py"
    click ExperimentCLI "https://github.com/baitingzbt/peda/blob/master/experiment.py"
    click ExperimentScript "https://github.com/baitingzbt/peda/blob/master/all_env_uniform.sh"

    %% Styles
    classDef codeModules fill:#ADD8E6,stroke:#000,stroke-width:1px
    classDef dataArtifacts fill:#90EE90,stroke:#000,stroke-width:1px
    classDef externals fill:#FFA500,stroke:#000,stroke-width:1px
```
