# anonymous-report-421/gpt-as-policy architecture

> Architecture diagram of the GitHub repository anonymous-report-421/gpt-as-policy, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-10-05). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/anonymous-report-421/gpt-as-policy
- Repository: https://github.com/anonymous-report-421/gpt-as-policy

## Overview

This repository evaluates GPT 6 Astra as an embodied policy, comparing direct GPT actions with a hybrid policy that reviews π₀.₅ proposals, and publishes a bilingual technical report. The principal runtime path is case setup, simulator observation, policy proposal and review, action execution, then recorded results. Separate RoboLab integrations and local annotation/review tools support related evaluation work. The report frontend and preview serve a distinct publication workflow. The README documents the overall capabilities; several subsystem links below are dashed where the sampled source does not establish exact wiring.

Main entry points: [`hybrid_rollout/robodojo/robodojo_server/client.py`](https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robodojo/robodojo_server/client.py) (Rollout Client), [`hybrid_rollout/robodojo/robodojo_server/server.py`](https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robodojo/robodojo_server/server.py) (Simulator Server), [`hybrid_rollout/robolab/robolab_server/server.py`](https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robolab/robolab_server/server.py) (RoboLab Server), [`hybrid_rollout/report_site/app/src/main.jsx`](https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/report_site/app/src/main.jsx) (Report Frontend)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_evaluation["RoboDojo Evaluation"]
  node_skill["Rollout Tools<br/>[run.py]"]
  node_rollout["Rollout Client<br/>[client.py]"]
  node_simserver["Simulator Server<br/>[server.py]"]
  node_session["Episode Session<br/>[session.py]"]
  node_evaluationcases["Case Evaluation<br/>[evaluation.py]"]
  node_artifacts["Run Artifacts<br/>[io.py]"]
end

subgraph group_policy["Policy Services"]
  node_pi_server["π₀.₅ Service<br/>[server.py]"]
  node_pi_client["π₀.₅ Client<br/>[client.py]"]
  node_gpt["GPT Baseline<br/>[gpt_only_client.py]"]
  node_decision["Decision Records<br/>[decision_log.py]"]
end

subgraph group_robolab["RoboLab Integration"]
  node_robolabserver["RoboLab Server<br/>[server.py]"]
  node_robolabsession["RoboLab Session<br/>[session.py]"]
  node_eef["EEF Controller<br/>[eef_controller.py]"]
  node_robolabpi["RoboLab π₀.₅<br/>[server.py]"]
end

subgraph group_report["Report Experience"]
  node_reportapp["Report Frontend<br/>[main.jsx]"]
  node_reportcontent["Report Content<br/>[ReportContent.jsx]"]
  node_charts["Chart State<br/>[ChartState.jsx]"]
  node_dataquery["Query Boundary"]
  node_preview["Local Preview<br/>[preview.py]"]
  node_reportserver["Report Server<br/>[report_server.py]"]
end

subgraph group_review["Annotation Review"]
  node_annotation["Annotation Dashboard"]
  node_annotationui["Annotation UI<br/>[app.js]"]
  node_dashboard["History Dashboard"]
end

node_operator(("Evaluator"))

node_operator -.->|"starts rollout"| node_skill
node_skill -.->|"drives"| node_rollout
node_rollout -->|"RPC calls"| node_simserver
node_simserver -->|"serves episodes"| node_session
node_simserver -->|"validates cases"| node_evaluationcases
node_rollout -.->|"requests inference"| node_pi_client
node_pi_client -.->|"WebSocket inference"| node_pi_server
node_rollout -.->|"alternative baseline"| node_gpt
node_rollout -.->|"writes run data"| node_artifacts
node_rollout -.->|"records decisions"| node_decision
node_operator -.->|"runs RoboLab"| node_robolabserver
node_robolabserver -->|"creates session"| node_robolabsession
node_robolabserver -.->|"uses controller"| node_eef
node_operator -.->|"runs policy service"| node_robolabpi
node_reportapp -.->|"renders"| node_reportcontent
node_dataquery -->|"shows loading state"| node_charts
node_reportapp -.->|"hosts query views"| node_dataquery
node_operator -->|"previews report"| node_preview
node_preview -.->|"serves report"| node_reportserver
node_operator -.->|"reviews annotations"| node_annotation
node_annotation -.->|"serves interface"| node_annotationui
node_operator -.->|"inspects history"| node_dashboard

click node_skill "https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robodojo/skill/run.py"
click node_rollout "https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robodojo/robodojo_server/client.py"
click node_simserver "https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robodojo/robodojo_server/server.py"
click node_session "https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robodojo/robodojo_server/session.py"
click node_evaluationcases "https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robodojo/evaluation.py"
click node_artifacts "https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robodojo/io.py"
click node_pi_server "https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robodojo/pi05_server/server.py"
click node_pi_client "https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robodojo/pi05_server/client.py"
click node_gpt "https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robodojo/robodojo_server/gpt_only_client.py"
click node_decision "https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robodojo/robodojo_server/decision_log.py"
click node_robolabserver "https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robolab/robolab_server/server.py"
click node_robolabsession "https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robolab/robolab_server/session.py"
click node_eef "https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robolab/robolab_server/eef_controller.py"
click node_robolabpi "https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robolab/pi05_server/server.py"
click node_reportapp "https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/report_site/app/src/main.jsx"
click node_reportcontent "https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/report_site/app/src/content/report/ReportContent.jsx"
click node_charts "https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/report_site/app/src/charting/ChartState.jsx"
click node_dataquery "https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/report_site/app/src/components/QueryDataBoundary.jsx"
click node_preview "https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/report_site/preview.py"
click node_annotation "https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/annotation_dashboard.py"
click node_annotationui "https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/annotation_ui/app.js"
click node_dashboard "https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/history_dashboard.py"
click node_reportserver "https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/report_server.py"

classDef toneNeutral fill:#f8fafc,stroke:#334155,stroke-width:1.5px,color:#0f172a
classDef toneBlue fill:#dbeafe,stroke:#2563eb,stroke-width:1.5px,color:#172554
classDef toneAmber fill:#fef3c7,stroke:#d97706,stroke-width:1.5px,color:#78350f
classDef toneMint fill:#dcfce7,stroke:#16a34a,stroke-width:1.5px,color:#14532d
classDef toneRose fill:#ffe4e6,stroke:#e11d48,stroke-width:1.5px,color:#881337
classDef toneIndigo fill:#e0e7ff,stroke:#4f46e5,stroke-width:1.5px,color:#312e81
classDef toneTeal fill:#ccfbf1,stroke:#0f766e,stroke-width:1.5px,color:#134e4a
class node_skill,node_rollout,node_simserver,node_session,node_evaluationcases,node_artifacts toneBlue
class node_pi_server,node_pi_client,node_gpt,node_decision toneAmber
class node_robolabserver,node_robolabsession,node_eef,node_robolabpi toneMint
class node_reportapp,node_reportcontent,node_charts,node_dataquery,node_preview,node_reportserver toneRose
class node_annotation,node_annotationui,node_dashboard,node_operator toneIndigo
```

## Components

### RoboDojo Evaluation

- [Rollout Tools](https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robodojo/skill/run.py): `hybrid_rollout/robodojo/skill/run.py`
- [Rollout Client](https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robodojo/robodojo_server/client.py): `hybrid_rollout/robodojo/robodojo_server/client.py`
- [Simulator Server](https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robodojo/robodojo_server/server.py): `hybrid_rollout/robodojo/robodojo_server/server.py`
- [Episode Session](https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robodojo/robodojo_server/session.py): `hybrid_rollout/robodojo/robodojo_server/session.py`
- [Case Evaluation](https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robodojo/evaluation.py): `hybrid_rollout/robodojo/evaluation.py`
- [Run Artifacts](https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robodojo/io.py): `hybrid_rollout/robodojo/io.py`

### Policy Services

- [π₀.₅ Service](https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robodojo/pi05_server/server.py): `hybrid_rollout/robodojo/pi05_server/server.py`
- [π₀.₅ Client](https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robodojo/pi05_server/client.py): `hybrid_rollout/robodojo/pi05_server/client.py`
- [GPT Baseline](https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robodojo/robodojo_server/gpt_only_client.py): `hybrid_rollout/robodojo/robodojo_server/gpt_only_client.py`
- [Decision Records](https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robodojo/robodojo_server/decision_log.py): `hybrid_rollout/robodojo/robodojo_server/decision_log.py`

### RoboLab Integration

- [RoboLab Server](https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robolab/robolab_server/server.py): `hybrid_rollout/robolab/robolab_server/server.py`
- [RoboLab Session](https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robolab/robolab_server/session.py): `hybrid_rollout/robolab/robolab_server/session.py`
- [EEF Controller](https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robolab/robolab_server/eef_controller.py): `hybrid_rollout/robolab/robolab_server/eef_controller.py`
- [RoboLab π₀.₅](https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/robolab/pi05_server/server.py): `hybrid_rollout/robolab/pi05_server/server.py`

### Report Experience

- [Report Frontend](https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/report_site/app/src/main.jsx): `hybrid_rollout/report_site/app/src/main.jsx`
- [Report Content](https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/report_site/app/src/content/report/ReportContent.jsx): `hybrid_rollout/report_site/app/src/content/report/ReportContent.jsx`
- [Chart State](https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/report_site/app/src/charting/ChartState.jsx): `hybrid_rollout/report_site/app/src/charting/ChartState.jsx`
- [Query Boundary](https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/report_site/app/src/components/QueryDataBoundary.jsx): `hybrid_rollout/report_site/app/src/components/QueryDataBoundary.jsx`
- [Local Preview](https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/report_site/preview.py): `hybrid_rollout/report_site/preview.py`
- [Report Server](https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/report_server.py): `hybrid_rollout/report_server.py`

### Annotation Review

- [Annotation Dashboard](https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/annotation_dashboard.py): `hybrid_rollout/annotation_dashboard.py`
- [Annotation UI](https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/annotation_ui/app.js): `hybrid_rollout/annotation_ui/app.js`
- [History Dashboard](https://github.com/anonymous-report-421/gpt-as-policy/blob/main/hybrid_rollout/history_dashboard.py): `hybrid_rollout/history_dashboard.py`

### Other components

- **Evaluator**

## Connections

- Evaluator → Rollout Tools: starts rollout
- Rollout Tools → Rollout Client: drives
- Rollout Client → Simulator Server: RPC calls (evidence: `hybrid_rollout/robodojo/robodojo_server/client.py`)
- Simulator Server → Episode Session: serves episodes (evidence: `hybrid_rollout/robodojo/robodojo_server/server.py`)
- Simulator Server → Case Evaluation: validates cases (evidence: `hybrid_rollout/robodojo/robodojo_server/server.py`)
- Rollout Client → π₀.₅ Client: requests inference
- π₀.₅ Client → π₀.₅ Service: WebSocket inference
- Rollout Client → GPT Baseline: alternative baseline (evidence: `hybrid_rollout/robodojo/robodojo_server/gpt_only_client.py`)
- Rollout Client → Run Artifacts: writes run data (evidence: `hybrid_rollout/robodojo/robodojo_server/client.py`)
- Rollout Client → Decision Records: records decisions (evidence: `hybrid_rollout/robodojo/robodojo_server/client.py`)
- Evaluator → RoboLab Server: runs RoboLab
- RoboLab Server → RoboLab Session: creates session (evidence: `hybrid_rollout/robolab/robolab_server/server.py`)
- RoboLab Server → EEF Controller: uses controller
- Evaluator → RoboLab π₀.₅: runs policy service
- Report Frontend → Report Content: renders
- Query Boundary → Chart State: shows loading state (evidence: `hybrid_rollout/report_site/app/src/components/QueryDataBoundary.jsx`)
- Report Frontend → Query Boundary: hosts query views
- Evaluator → Local Preview: previews report (evidence: `README.md`)
- Local Preview → Report Server: serves report
- Evaluator → Annotation Dashboard: reviews annotations
- Annotation Dashboard → Annotation UI: serves interface
- Evaluator → History Dashboard: inspects history
