Install & Deploy¶

DAAS runs locally with uv and Python 3.10+. Everything reads/writes a single SQLite file (daas.db) at the repo root.

1. Clone & install¶

git clone https://github.com/chengsishi/DAAS.git
cd DAAS
uv sync            # provisions the root venv (data libs + mkdocs-material)

uv sync installs runtime dependencies (akshare, yfinance, edgar, edinet-tools, world_bank_data, ckanapi) and the dev group (which includes mkdocs-material for this docs site).

Python version

The root venv is Python 3.10+. The dartlab datasource needs 3.12 - run it ad-hoc with uv run --python 3.12 --with dartlab ... (it is not a root dep).

2. Configure `.env`¶

Copy the template and fill in only the keys you use:

cp .env.example .env   # if a template exists; otherwise create .env

The single repo-root .env holds:

Key	Purpose
`DAAS_DATABASE_URL`	`sqlite:///daas.db` (the repo-root DB).
`HTTP_PROXY`	Outbound proxy if your network needs one.
`EDGAR_IDENTITY`	Your SEC EDGAR user agent.
`EDINET_API_KEY`	Japan EDINET API key.
`ALERTS_FEISHU_WEBHOOK_URL`	Feishu webhook for alert delivery.
`LLM_` / `LEADER_MODEL`	Leader agent (CrewAI workflow) model config.
`DASHBOARD_PORT`	Port for the dashboard server.
`CKAN_PORTAL_URL`	CKAN portal base URL.

Never commit real secrets

.env is git-ignored. Examples in this site use placeholders like <YOUR_API_KEY> - never paste a real key into a doc or commit.

3. Launch the MCP server¶

The consolidated fd-daas-mcp server is the sole entry in repo-root .mcp.json. Claude Code picks it up automatically. To launch it manually (for testing or another MCP client):

fd-daas-mcp/bin/fd-daas-mcp-server        # the .mcp.json `command`

4. Self-check¶

Run the offline invariants to confirm the server is wired correctly:

fd-daas-mcp/.venv/bin/python -m fd_daas_mcp.selfcheck

5. Query `daas.db` directly¶

From the repo root, the canonical DB is daas.db (not mcp/daas.db):

sqlite3 daas.db "SELECT count(*) FROM entities;"
sqlite3 daas.db "SELECT name, op, indicator_name FROM indicator_rules LIMIT 5;"

Use PRAGMA foreign_keys=ON; when you need FK cascade behavior.

6. (Optional) Serve this docs site locally¶

uv run mkdocs serve               # http://127.0.0.1:8000

See Contributor Guide -> Deploy the Docs for GitHub Pages deployment and WiFi/tunnel sharing.

Next¶

Continue to First Fetch in 5 Minutes.