First Fetch in 5 Minutes¶
This walkthrough fetches real data and computes a real indicator, end to end,
using the skill path. You will fetch SPY daily prices and confirm the
prebuilt SPY_ma5 (5-day simple moving average) indicator.
Prerequisites¶
- You have run
uv sync(see Install & Deploy). - The repo-root
daas.dbexists (it ships with the repo).
1. Confirm the entity exists¶
DAAS already knows about thousands of entities. Check that SPY is registered:
sqlite3 daas.db "SELECT id, code, name, entity_type FROM entities WHERE code='SPY' OR ticker='SPY' LIMIT 5;"
If SPY is not present, you can search/resolve it through the
fd-daas-fetch-data skill or the daas_search_entities MCP tool.
2. Resolve the fetch shape¶
The dispatch layer maps each source prefix to its Python library. SPY comes from
yfinance. See exactly how a function is called:
uv run python .claude/skills/fd-daas-based-data-fetch/scripts/dispatch.py --resolve yfinance_ticker_history
This prints the import + call shape the skill uses - no guessing.
3. Fetch (through the skill)¶
In Claude Code, just ask:
Fetch SPY daily OHLCV history (1 year) and persist it.
The fd-daas-based-data-fetch (or fd-daas-fetch-data) skill will:
- Resolve the SPY entity + its
yfinanceidentifier. - Call
yfinancedirectly. - Persist the rows into a
scraw_<slug>table (e.g.scraw_spy_daily) viascripts/upsert.py(which backs updaas.dbfirst).
Confirm the data landed:
4. Compute a prebuilt indicator¶
SPY_ma5 is already an indicator_rules row (5-day SMA of SPY closes). Compute
it into the observations table:
Read the result:
sqlite3 daas.db "SELECT date, value FROM observations WHERE indicator='SPY_ma5' ORDER BY date DESC LIMIT 5;"
5. List available ops¶
Want a different indicator? The math ops available:
Ops include sma, ema, rsi, pct_change, log_return, diff,
rolling_std, rolling_min, rolling_max, zscore, ratio, level.
What just happened¶
- You used the skill path: skills called a Python library (
yfinance) directly and wrotedaas.dbviasqlite3. - Raw data went to
scraw_spy_daily; the computed series went toobservations. - The MCP path sees the same data - try
daas_get_series_latestor browse via the dashboard skill.
Next¶
- Browse the full indicator catalog: Browse Indicators.
- Group entities/indicators: Create Collections.
- Turn a goal into a full study: Create a Research.