metadata
title: Examples & Workflows
summary: Short examples for Python, CLI, notebooks, and the hosted assistant.
Examples & Workflows
This page collects the most common usage patterns in one place.
Python: Market Data
from TerraFin import configure
from TerraFin.data import DataFactory
configure()
data = DataFactory()
spy = data.get("S&P 500")
unrate = data.get_fred_data("UNRATE")
portfolio = data.get_portfolio_data("Warren Buffett")
Python: External-Agent Client
from TerraFin.agent import TerraFinAgentClient
client = TerraFinAgentClient()
snapshot = client.market_snapshot("AAPL")
macro = client.macro_focus("Nasdaq", depth="auto", view="weekly")
Python: Hosted Agent Session
from TerraFin.agent import TerraFinAgentClient
client = TerraFinAgentClient()
session = client.runtime_create_session("terrafin-assistant")
reply = client.runtime_message(session["sessionId"], "Compare the S&P 500 and Nasdaq.")
CLI: Data Lookup
terrafin-agent snapshot AAPL
terrafin-agent indicators AAPL --indicators rsi,macd
terrafin-agent macro-focus "S&P 500" --view weekly
CLI: Model Management
terrafin-agent models list --all
terrafin-agent models current
terrafin-agent models use google/gemini-3.1-pro-preview
terrafin-agent models auth login-github-copilot --set-default
Notebook: Explicit Bootstrap
from TerraFin import configure, load_terrafin_config
configure()
config = load_terrafin_config()
Then call the notebook helpers:
from TerraFin.interface.chart.client import display_chart_notebook
display_chart_notebook(spy)
Browser: View-Aware TerraFin Agent
When you use the hosted assistant widget inside the interface:
- chart, stock, market-insights, and DCF pages publish structured view context
- the assistant can inspect that context with
current_view_context - page state is available on demand rather than injected into every prompt
That means prompts like “retrieve information on the top 10 companies in this portfolio” can resolve against the current page instead of relying on the user to restate the visible context manually.