Write JSON and Markdown run cards for a backtest run. Args: run_dir: Directory where run_card.json and run_card.md are written. c

Filter DataFrame by a simple expression. Supports: - "YYYY-MM to YYYY-MM" or "YYYY-MM-DD to YYYY-MM-DD" (date range) - "symbol=XX

Fail closed when the resolved host does not match the declared environment. The host is the authoritative discriminator: testnet keys cannot reac

Fail closed when the account does not match the declared environment.

实体长度（绝对值）。 Args: o: 开盘价序列。 c: 收盘价序列。 Returns: 每根K线的实体长度。

Build the Vibe-owned Codex token store. The path deliberately does not use oauth-cli-kit's default storage, because that backend imports and

Build the CLI parser with subcommands and compatibility flags.

Build data_map, signal_map, codes for testing.

Total equity = free cash + sum(margin + unrealised) per position. Uses vectorized numpy path when _calc_pnl/_calc_margin are not over

Hybrid output contract. Return a short reason string when the worker did NOT produce a substantive deliverable, else ``None``. Content-sanity

Close position, record trade, return capital.

Return the current API auth key, if configured. Resolves the ``VIBE_TRADING_API_KEY`` alias server-side via :func:`get_env_config` so both ``

Build override dict for eToro tools, including optional credential overrides.

Core orchestration loop (runs in background thread). Steps: 1. Update run status to running 2. Initialize TaskStore,

Intraday bars via stk_mins. Args: codes: Stock codes. start_date: Start date (YYYY-MM-DD). end_date: End

Detect exchange-listed ETF / LOF symbols (e.g. 510050.SH, 159915.SZ).

Return True if the tool result does not look like an error response.

Return the raw text of one lens file. Args: slug: Lens file stem, e.g. ``munger-inversion``. Returns: Raw markdown text.

Epoch-ms of a wall-clock time in an IANA zone (``fold=0`` = first occurrence).

Build a minimal OHLCV DataFrame matching tushare's column layout.

Normalize ``run_dir`` in tool args to an absolute path when possible. If the model supplies a relative ``run_dir`` (for example ``"."`` or ``

Label i resolves `horizon` bars later, so labels overlap heavily.

Return (start_date, end_date) as YYYY-MM-DD strings.

Build a tz-naive trade_date-indexed OHLCV frame like a loader returns.

Raise ValueError if `p` starts with a UNC share prefix.

Send a JSON request with spacing and 429 Retry-After handling.

Return the shared CLI console, or a caller-supplied override.

Scan all .md files (except MEMORY.md) and parse frontmatter.

Write a durable last-known runtime_state.json for ``broker`` directly.

Write a dummy .md file with basic frontmatter.

Parse a trade journal and return a JSON analysis. Args: file_path: Path to CSV/Excel file. analysis_type: "full" | "profile" | "b

Control the false discovery rate across many simultaneous tests. Bonferroni controls the probability of *any* false positive and is far too s

Link single-period Brinson results into a residual-free multi-period attribution. Every period's effects are scaled by ``k_t / k`` and summed. Be

Bootstrap the OAuth handshake for a live broker channel (desktop only). Builds the broker's MCP tool wrappers, which forces a connection and — wh

Show auth state, active mandate, and halt state for live channels. Read-only: it loads the mandate via :func:`src.live.mandate.store.load_mandate

Compute importance via Ebbinghaus-inspired decay formula.

Create and persist a new hypothesis. Args: title: Short hypothesis title. thesis: Research thesis or rationale.

Regress one date's asset returns on their exposures. The coefficients are that date's factor returns. Weighting is by square-root market cap

Rough token count estimate (~4 chars/token). Args: messages: Message list. Returns: Estimated token count.

Fit a Nelson-Siegel or Svensson curve to observed zero rates. The betas are solved exactly by least squares at every candidate decay paramete

Apply a constant basis-point slippage to a fill price. The cheapest model, and adequate whenever the order is small enough that it does not m

Render *grounding* as a markdown block ready to splice into a prompt. Returns the empty string if no symbol has any data — callers can use th

Read a session. Args: session_id: Session ID. Returns: The Session instance, or None when it does not exist.

Check if a memory with similar content was recently written. Uses a 30-second sliding window to catch rapid-fire duplicates from retr

Return whether QVeris paid routing should be visible.

Report whether a runner's heartbeat is fresh enough to call it alive. Args: runner_id: The runner to check. now_ms: Reference tim

Test whether the breach *count* matches the model's claimed level. The Kupiec proportion-of-failures test compares the realised breach rate a

Load all hypotheses from storage.

Return the root directory for all live-channel state. Returns: ``<runtime_root>/live``. The directory is NOT created here; callers

Load all persisted jobs. Returns: A dict mapping job id to job. Empty when the store has never been written.

CLI entrypoint returning a process exit code. Behaviour: * Interactive entry (no subcommand, or ``chat`` + TTY): show banner, run onbo

Build a market-session trigger for a known market key. Args: market: A key into :data:`MARKET_SPECS` (e.g. ``"us_equity"``).

Simulate price paths under geometric Brownian motion. Each step applies ``(mu - 0.5 * sigma**2) * dt + sigma * sqrt(dt) * Z`` in log space, s

Value at Risk under a fitted normal distribution. Cheap and smooth, but it sees only a mean and a variance, so it misses fat tails. The direc

Preferred return owed to LPs, accrued per contribution from its own date. Accruing each draw separately matters: a fund that called half its capi

Record a bench result. Returns the result ID.

Redact one tool result for trace persistence and event previews. The single choke point for a tool-result string: callers redact once and han

Register a tool.

Render the SignalEngine Python source for a profile. Args: profile: ShadowProfile carrying the rules to replay. Returns: Pyt

Run chart pattern detection on OHLCV data in run_dir. Args: run_dir: Path to the run directory. patterns: Comma-separated pattern

Per-row L1 normalize so sum of absolute values equals ``a``. Rows whose abs-sum is 0 (or all-NaN) become NaN — never silent zero.

Determine if a status transition should occur. Args: current_status: Current artifact status. consecutive_signals: Re

Map a markdown document to its heading structure. A section runs from its heading to the next heading of the same or shallower level, so aski

Sum the signed amounts. Valuation marks such as NAV are excluded by default: adding a mark to settled cash would overstate what was a

Convert flows in one or more currencies into ``rate_table``'s currency. Each flow is converted using the FX rate quoted on **that flow's own

Atomically update run state. Args: run: Updated SwarmRun instance. Raises: FileNotFoundError: If the run dir

Build the weighted average cost of capital. ``Re`` is CAPM plus two optional additive premiums; ``Rd`` is the pretax cost of debt shielded by

Atomically record a runner's last-tick timestamp. Args: runner_id: The runner writing its heartbeat. now_ms: Timestamp to record,