Append pending entry at end of propagate(). No LLM call.

Parse all entries from log. Returns list of dicts.

Resolve capabilities by exact ID, then pattern, then default.

Update the configuration with custom values. Dict-valued keys (e.g. ``data_vendors``) are merged one level deep so a partial update like ``{"

Map a user/broker symbol to its canonical Yahoo Finance symbol. Resolution order (first match wins): 1. Explicit alias table (metals, energ

Return the instrument context for the current run. Prefers the identity-resolved context computed once at run start and stored on the state (

Create an LLM client for the specified provider. Provider modules are imported lazily so that simply importing this factory (e.g. during test

Return a prompt instruction for the configured output language. Returns empty string when English (default), so no extra tokens are used. App

Return formatted past context string for agent prompt injection.

Validate ``value`` is safe to interpolate into a filesystem path. Tickers come from user CLI input or from LLM tool calls, both of which can

Heuristically extract a 5-tier rating from prose text. Two-pass strategy: 1. Look for an explicit "Rating: X" label (tolerant of markdown bol

Route method calls to appropriate vendor implementation with fallback support.

Helper function to make API requests and handle responses. Raises: AlphaVantageRateLimitError: When API rate limit is exceeded

Get the current configuration.

Deterministic thread ID for a ticker+date pair. ``signature`` folds in graph-shape-affecting run choices so a resume under a different graph

Execute a yfinance call with exponential backoff on rate limits. yfinance raises YFRateLimitError on HTTP 429 responses but does not retry th

Set/clear env vars then reload default_config to re-evaluate DEFAULT_CONFIG.

Store a decision then immediately resolve it via the API.

Return the env var name for `provider`'s API key, or None if not applicable. Unknown providers also return None — callers should treat that as

Replace pending tag and append REFLECTION section using atomic write. Finds the first pending entry matching (trade_date, ticker), updates

Check if model name is valid for the given provider. For ollama, openrouter, and openai_compatible - any model is accepted.

Pick the benchmark ticker for alpha calculation against ``ticker``. ``config["benchmark_ticker"]`` overrides everything when set; otherwise

Create a sentiment analyst node for the trading graph. Pre-fetches news + StockTwits + Reddit data, injects them into the prompt as structure

Reject OHLCV whose latest row is far older than curr_date. Raises NoMarketDataError (with a stale-specific detail) so the router treats it li

Extract bind() kwargs from a with_structured_output result.

Validate an ``llm_max_retries`` value to a non-negative int. Accepts an int or a numeric string (env vars arrive as strings). Rejects boolean

Fetch raw and alpha return for ticker over holding_days from trade_date. ``benchmark`` is the index used as the alpha baseline (resolved by t

Get provider-specific kwargs for LLM client creation.

OHLCV frame whose date column is named `date_col`.

Describe the exact instrument so agents preserve identity and ticker. When ``identity`` is provided (resolved deterministically via :func:`re

Return the crypto base (e.g. ``BTC``) for a known USD/USDT/USDC-quoted crypto symbol in any form the pipeline may hold — ``BTC-USD``, ``BTCUSD``,

Context manager yielding a SqliteSaver backed by a per-ticker DB.

Resolve deterministic identity metadata (company name, sector, …) for a ticker. This exists to stop the pipeline from hallucinating a *different*

Warn when the model is outside the known list for the provider.

Build a _request replacement that dispatches on the endpoint path.

Write a completed entry directly to file, bypassing the API.

Create the initial state for the agent graph. ``instrument_context`` is the deterministic ticker-identity string resolved once at run

Return entries with outcome:pending (for Phase B).

Run the structured call and render to markdown; fall back to free-text on any failure. ``prompt`` is whatever the underlying LLM accepts (a strin

Normalize LLM response content to a plain string. Multiple providers (OpenAI Responses API, Google Gemini 3) return content as a list of type

Single reflection call on the final trade decision with outcome context. Used by Phase B deferred reflection. The final_trade_decision alread

Normalise a langchain LLM input to a list of message objects. Accepts a list of messages, a ``ChatPromptValue`` (from a ChatPromptTemplate),

True when ``base_url`` is unset or points at api.openai.com. The Responses API (/v1/responses) only exists on native OpenAI. A custom base_ur

Minimal AgentState dict for portfolio_manager_node.

Return ``llm.with_structured_output(schema)`` or ``None`` if unsupported. Logs a warning when the binding fails so the user understands the agent

Classify on the canonical symbol so e.g. BTCUSD and BTC-USDT both read as crypto (#981/#982), matching what the data path will actually fetch.

Check whether a resumable checkpoint exists for ticker+date.

Resolve user input to its canonical Yahoo symbol (single source of truth). Delegates to the data layer's ``normalize_symbol`` so the symbol the C

Return one of Buy / Overweight / Hold / Underweight / Sell.

Return the default backend URL for a provider key, or None if unknown.

Render a SentimentReport to the markdown shape the rest of the system expects. The structured header (band + score + confidence) is prepended to

Return the SQLite checkpoint DB path for a ticker.

Drop annual/quarterly reports dated after curr_date to prevent look-ahead. ``_make_api_request`` returns the fundamentals payload as a JSON strin

Prompt for a required value; exit cleanly if the user cancels. ``questionary.text(...).ask()`` returns None on Ctrl-C/Esc; mirror the exit-on

Resolve pending log entries for ticker at the start of a new run. Fetches returns for each same-ticker pending entry, generates reflections,

A minimal context-manager response whose read() runs ``read_fn``.

Graph-shape inputs that must invalidate a checkpoint if changed. Keyed into the checkpoint thread ID so a resume under a different analyst

Build a MagicMock LLM whose with_structured_output binding captures the prompt and returns a real PortfolioDecision (so render_pm_decision works).

MagicMock LLM whose structured binding captures the prompt and returns a real SentimentReport so render_sentiment_report works.

Return the step number of the latest checkpoint, or None if none exists.