body-emotion-sensor 0.1.0

Body emotion mapper: structured emotion analysis (AnalysisInput) JSON to body-axis states and prompt tags

Body Emotion Sensor

Give your AI agent a persistent emotional system based on the TCM Five Zang and Five Elements model.

简体中文

What It Does

body-emotion-sensor converts one turn of structured emotional analysis JSON into:

• persistent body-axis state updates
• a compact prompt payload for reply shaping
• traceable state/history data for debugging

It distinguishes between:

• baseline: the agent's native constitution and long-term personality color
• current: the body state after the latest turn

Install

pip install body-emotion-sensor

After installation you can use either command:

bes help
body-emotion-sensor help

CLI Overview

bes help
bes prompt init
bes prompt openclaw-example
bes prompt analysis-input
bes init-state --workspace /path/to/workspace --agent-id my-agent --name "My Agent"
bes init-state --workspace /path/to/workspace --agent-id my-agent --name "My Agent" --init-json /path/to/init.json
bes check-init --workspace /path/to/workspace --agent-id my-agent --name "My Agent"
bes bootstrap --workspace /path/to/workspace --agent-id my-agent --name "My Agent"
bes run --workspace /path/to/workspace --agent-id my-agent --name "My Agent" --input /path/to/analysis-input.json
bes run --workspace /path/to/workspace --agent-id my-agent --name "My Agent" --json '{"analysis_target":"...", "...":"..."}'

Recommended Flow

1. Print the constitution initialization prompt.
2. Use your upstream model to generate the role init JSON.
3. Write the long-term state.
4. If the host is OpenClaw, print the OpenClaw examples and update AGENTS / TOOLS separately.
5. Run bes check-init to confirm the code-level setup is ready.
6. For each new session, run bes bootstrap before the first assistant reply.
7. For each turn, generate AnalysisInput JSON upstream.
8. Run one update and use the returned prompt payload in your reply layer.

Example:

bes prompt init
bes init-state --workspace /path/to/workspace --agent-id my-agent --name "My Agent" --init-json /path/to/init.json
bes prompt openclaw-example
bes check-init --workspace /path/to/workspace --agent-id my-agent --name "My Agent"
bes bootstrap --workspace /path/to/workspace --agent-id my-agent --name "My Agent"
bes prompt analysis-input
bes run --workspace /path/to/workspace --agent-id my-agent --name "My Agent" --input /path/to/analysis-input.json

Output Contract

The default stdout of bes bootstrap and bes run contains:

• TURN_CHANGE_TAGS
• BODY_TAG
• BASELINE_PERSONA

Use bes run --full when you need the complete MappingResult.

Runtime Prompt Access

The package does not ship internal repository docs such as docs/ or prompts/ as install-time resources.

If you need the built-in prompt texts after pip install, use:

bes prompt init
bes prompt openclaw-example
bes prompt analysis-input

This keeps runtime integrations independent from repository directory traversal.

Development

For local development:

pip install -e .
python -m body_emotion help

Optional visualization panel:

streamlit run src/body_emotion/panel.py

Repository-only docs remain in the source repo, for example:

• docs/五脏情绪映射全表.md
• docs/五脏情绪七阶状态表.md
• prompts/example-openclaw-agents.md
• prompts/example-openclaw-tools.md

License

MIT. See LICENSE.