Serialize the largest whole-record page of ``records[offset:]`` that fits. Args: records: All available records, newest first by the call

Issue a throttled Eastmoney GET and decode the body as JSON. Args: url: Fully-qualified Eastmoney endpoint URL. params: Query par

Load every available playbook, sorted by slug. A slug found in an earlier directory shadows the same slug later, so a user file overrides the

Load the persisted jobs. A missing store (never written) is the *only* clean empty result. A store that exists but fails to parse is

Load a YAML preset by name (user directory first, then bundled). Args: name: Preset name (without .yaml extension). Returns:

Move legacy code-relative state directories under the runtime root. Idempotent: once a legacy entry has been moved (or was never there), late

Single source of truth for "what does this run actually look like". Three layered transforms, applied in order to ``run``: 1. **Hydr

Register a new artifact, assigning an ID and timestamps. Returns: The assigned ``artifact_id``.

Strict-mode bench: like ``run_bench`` but with mandatory random control. Args: zoo: Zoo id (e.g. ``alpha101``, ``gtja191``, ``qlib158``,

Validate a run directory used by generated-code tools. Args: p: User/LLM-supplied run directory. `~` expansion is supported. Returns

Save or overwrite task state. Args: task: SwarmTask instance.

Start the QQ bot with auto-reconnect loop.

Drop, flag, or reject bars that violate OHLC invariants. Loaders only drop NaN rows, so structurally dirty bars — ``high < low``, a non-posit

Calculate options backtest metrics. Args: equity: Equity series. initial_cash: Initial capital. trades: List of trade rec

Shared call shape: fixture-engineered windows + pinned today.

Create the appropriate market engine based on data source and market type. Routing priority: 1. Detect market type from symbol patterns (fu

Fetch and normalize one daily OHLCV frame, routing by symbol type.

Did a tool call return a top-level error envelope? Parses the result as JSON and checks for a top-level ``status == "error"``. Also treats ``

Normalize one gamma market through the tool and return the record.

Normalize a Futu data payload to a list of row dicts. Handles a pandas DataFrame (via ``to_dict('records')``), an existing list of dicts, a s

Split text into searchable tokens (ASCII >=2 chars + non-Latin chars).

Generate a mean-zero AR(1) series, the discrete analogue of an OU process. Args: phi: Persistence coefficient; |phi| < 1 gives a stationa

Bucket a strict bench row into one of four ``StrictCategory`` values. Required keys on ``row``: - ``alpha_t_full`` (float) - ``al

Clear all issued tokens.

Close the database connection.

Return preloaded rows for requested codes.

Format an assistant tool_calls message, preserving thinking text. Args: tool_calls: List of tool call objects. conten

Probe public candles with a short timeout (honours proxy env).

Register a tool.

Render a PASS/FAIL verdict from per-point verification results. Each result with a ``fetched_value`` is judged against the reported value at

Unrealised NAV of the fund at its most recent mark. Only the latest valuation record counts. Earlier marks are stale by definition, and summi

Run one worker task, releasing the per-task LLM client on exit. Builds a fresh :class:`ChatLLM` for the task and guarantees it is closed even

Start a swarm run. Returns immediately, execution happens in background. Args: preset_name: YAML preset name to execute.

GET ``url`` after waiting out the per-host minimum interval. Args: url: Fully-qualified request URL. host_key: Throttle/session b

Bold every cell in one worksheet row (typically a header row).

Margin (collateral) required for a position. ``abs(price)`` so collateral stays positive when the entry price is negative; for the us

Helper to create a memory file with extended frontmatter.

Build a ``run_bench`` double returning canned rows for the ``only`` ids.

Run the tool against a mocked companyfacts payload and return its periods.

Return the host ``api_server`` module for shared deps and monkeypatched symbols.

Decide whether this invocation should drive the interactive loop. Interactive entry requires: * Both stdin and stdout are TTYs (no piped / r

Build a minimal Futu kline DataFrame mirroring request_history_kline output.

A market the oracle finalized, with the payout price pinned on "No".

Return a README's text. Args: name: File name relative to the repository root. Returns: The file contents.

Import an optional backend module or raise an actionable ImportError. Args: module_path: Dotted module to import, e.g. ``"statsmodels.tsa

Reject anything that is not a ``CashFlowSeries``. Args: series: Candidate value supplied by the caller. name: Parameter name to q

Resolve base_url, api_key, model from the configured LLM provider. Provider credentials use dynamically-named env vars from :func:`provider_e

Build a throttled_get_json side_effect routing by endpoint.

Seed three hypotheses spanning multiple statuses; return their ids.

Convert project symbols into yfinance symbols. Args: code: Project symbol, for example ``AAPL.US``, ``700.HK``, or ``TD.TO``.

Return the data payload from a Futu ``(ret_code, data[, page_key])`` tuple. Futu read calls return ``(ret_code, data)``; ``request_history_kline`

Record a hit for ``key`` and report whether it stays within budget.

Append an event to events.jsonl. Args: run_id: Run identifier. event: Event to append. Raises: F

Cap each name's trade at a fraction of its average daily traded value. Args: target_weights: Desired weights by symbol, shorts negative.

Build a per-worker registry that merges local + remote MCP tools. Swarm workers receive a strict whitelist (``agent_spec.tools``). This build

Return the currency a symbol settles in. Args: code: Ticker / symbol string. Returns: A currency code such as ``"CNY"``. A f

Calculate an expiry payoff curve and analytic risk summary. Premiums are pinned at entry. An explicit per-share premium wins over the Black-S

Fit a Markov regime-switching model and label the current regime. A GARCH model says how volatile tomorrow is. This says *which state the mar

Invert Black-Scholes for volatility, Newton first then bisection. Seeds with the Brenner-Subrahmanyam at-the-money approximation ``sigma_0 =

Return ``user_message`` prefixed with a formatted context block. Args: request: The parsed ``openbb_ai.models.QueryRequest`` inst

Kaplan-Schoar public market equivalent. Every contribution and distribution is grown to a single common date (``as_of``) using the *public in

Return built-in trading connector profiles.

Build the constraint list from ``config['constraints']`` (empty if unset).