# agent-base

OpenClaw Session Archive - 多 Agent 会话解析与归档系统。

## Architecture

Three nodes (Mac Mini, Ubuntu1, Ubuntu2) run a sync script that parses local JSONL session files and pushes structured data to a central Django API. The Django service runs in Docker Compose and stores data in a remote PostgreSQL + TimescaleDB instance on NAS.

## Quick Start

```bash
# 1. Configure environment
cp .env.example .env
# Edit .env with your database credentials and Django settings

# 2. Build and start
docker compose build
docker compose run --rm web python manage.py migrate
docker compose run --rm web python manage.py createsuperuser
docker compose up -d

# 3. Access
# Django Admin: http://<host>:8000/admin/
# API: http://<host>:8000/api/sessions/bulk_upsert/
```

## Running Tests

```bash
pip install -r requirements-dev.txt
pytest -v
```

## Client Sync Script

Deploy `scripts/sync_sessions.py` on each node:

```bash
python sync_sessions.py --remote-url http://<django-host>:8000/api/sessions/bulk_upsert/
```

Set `SOURCE_NODE` environment variable on each node (`macmini`, `ubuntu1`, `ubuntu2`).

## Daily Export

In Django Admin, select sessions and choose "Export selected sessions to Markdown" action.