epsilab 0.1.3

Python SDK for Epsilab live portfolio data (signals, weights, trades).

Epsilab Python SDK

Official Python SDK for accessing Epsilab live portfolio data (signals, weights, trades, equity) using an API key.

Install

pip install epsilab

If you want HTTP/2 for faster connections, install with:

pip install "httpx[http2]"

Install from source

cd /home/matthew/Epsilab/Epsilab-API
python -m venv .venv && source .venv/bin/activate
pip install --upgrade pip
pip install .
# or for development
pip install -e .

Environment

Set these variables or pass explicitly to the client:

• EPSILAB_API_BASE (default https://www.epsilab.ai/api/ext/v1)
• EPSILAB_API_KEY (a user API key created in Account Settings)

Note:

• API access requires an active Pro subscription. Requests from accounts without Pro will be rejected.
• Live endpoints are rate limited (e.g., ~1 request/second sustained with small bursts). Excessive requests will receive HTTP 429.
• Use nextEtaMin from get_live_status() to poll intelligently. For example, if nextEtaMin=80.07, you can sleep ~80 minutes before the next refresh instead of tight-loop polling. The results will be the same.

Set API key in your shell (bash/zsh):

export EPSILAB_API_KEY="<epsk_live_...>"

Python Usage

from epsilab import Epsilab

client = Epsilab(api_base="https://www.epsilab.ai/api/ext/v1", api_key="<epsk_live_...>")

# Note: Running the portfolio via API key is disabled. Use the website to initiate runs.

# Fetch live data (typed models)
latest = client.get_live_latest(return_results_if_fresh=True)          # epsilab.models.LiveLatest
status = client.get_live_status()                                      # epsilab.models.LiveStatus
signals = client.get_portfolio_signals(limit=500)                      # List[epsilab.models.PortfolioSignal]
weights = client.get_portfolio_weights(limit=500)                      # List[epsilab.models.PortfolioWeight]
trades = client.get_live_trades(status="PENDING,EXECUTED", include_positions=True)  # List[epsilab.models.PortfolioTrade]
equity = client.get_live_equity(limit=200)                             # List[epsilab.models.EquityPoint]

print(latest)

Examples

See examples/example.py for a short polling example that prints summary data every few seconds.

Data models

• get_live_latest() → LiveLatest with fields:

  • run_id: Optional[str]
  • timeframe: Optional[str]
  • signals: List[PortfolioSignal]
  • weights: List[PortfolioWeight]
  • fresh: Optional[bool]
  • recomputed: Optional[bool]

• get_live_status() → LiveStatus with fields:

  • latest_run_id: Optional[str]
  • timeframe: Optional[str] (tf denotes the rebalancing frequency; e.g., tf=4h means the portfolio is updated on a 4-hour cadence, based on the most granular timeframe among strategies in the portfolio)
  • next_eta_minutes: Optional[float]
  • counts: Dict[str, int]

• get_portfolio_signals() → List[PortfolioSignal] with fields:

  • symbol: str, signal_type: str, strength: Optional[float], timestamp: Optional[str], date: Optional[str]

• get_portfolio_weights() → List[PortfolioWeight] with fields:

  • symbol: str, final_weight: Optional[float], raw_weight: Optional[float], timestamp: Optional[str], date: Optional[str]

• get_live_trades() → List[PortfolioTrade] with fields:

  • id: str, symbol: str, side: str, qty: Optional[float], planned_price: Optional[float], planned_at: Optional[str], status: str, reasoning: Optional[str]

• get_live_members() → List[PortfolioMember] with fields:

  • strategy_id: str, effective_weight: Optional[float], name: Optional[str], visibility: Optional[str]

• get_live_equity() → List[EquityPoint] with fields:

  • date: str, value: float

All models expose to_dict(), to_json(), and log() helpers for convenient logging.