canswim 0.0.20260713

Developer toolkit for CANSLIM-style investors (CLI, Gradio GUI, MCP)

canswim

Developer toolkit for CANSLIM investment style practitioners

NOT FINANCIAL OR INVESTMENT ADVICE. USE AT YOUR OWN RISK.

For a brief introduction read this blog post.

Here is also a video recording of a CANSWIM presentation for the Python Austin Meetup.

Documentation

Published site: https://ivelin.github.io/canswim/ (built from main / docs/ via GitHub Actions — not the legacy website branch).

Doc	Contents
docs/cli.md	CLI tasks, recipes, env vars
docs/run_triggers.md	Get market data / run forecast (CLI · GUI · MCP)
docs/mcp.md	MCP server (tools, opt-in writes)
docs/data_store.md	Parquet (SoT) vs DuckDB (search/UI)
AGENTS.md	CI, merge rules, docs DoD for agents

Flags: python -m canswim -h is the source of truth. Local site preview: pip install mkdocs-material && mkdocs serve.

Setup

pip install canswim
# pin a release: pip install canswim==0.0.20260713

# dev checkout
pip install -e ".[dev]"

# recommended for this repo
conda activate canswim

See CHANGELOG.md for release notes. Docs: https://ivelin.github.io/canswim/.

Local-first market data

By default gatherdata does not download or upload the full Hugging Face dataset. That HF snapshot step is slow and was a common reason the CLI “hung”. Instead:

1. Use local parquet under data/data-3rd-party/ (created/updated by gather).
2. Refresh from FMP / yfinance APIs as needed.
3. Resolve ticker universes from checked-in symbol_lists/*.csv.

# full-universe local gather (no HF dataset sync)
hfhub_sync=False python -m canswim gatherdata

# scoped list via --tickers (same pipeline as Dashboard Run + MCP gather_tickers)
hfhub_sync=False python -m canswim gatherdata --tickers "AAPL, MSFT"

Env	Default	Meaning
hfhub_sync	False	Full dataset/model sync off
SYNC_SYMBOL_LISTS	False	If True, fetch only light CSVs from the HF dataset once
YFINANCE_USE_CACHE	False	Avoid multi‑GB SQLite cache hang
MCP_ALLOW_RUNS	unset	Enable MCP gather/forecast tools (CLI/GUI do not need this)

Train/forecast skip tickers without complete ground-truth OHLCV (no synthetic price fill). Details: docs/data_store.md, docs/cli.md.

Get market data & run forecasts (CLI · GUI · MCP)

Two separate steps, same backend (canswim.run_triggers). Details: docs/run_triggers.md.

	Get market data	Run a forecast	Check start date
CLI	gatherdata --tickers "…"	forecast --tickers "…" [date] [--dry_run]	resolve_start
GUI	Update market data	Run forecast	Check start date
MCP	gather_tickers*	forecast_tickers*	resolve_forecast_start

*MCP write tools need MCP_ALLOW_RUNS=1. Full MCP guide: docs/mcp.md.

Scoped get-market-data uses ~2 years of history, fundamentals (unless --no_covariates), and skips downloads when local files are already complete. Forecasts stop if data is incomplete (no invented prices).

python -m canswim gatherdata --tickers "AAPL, MSFT"
python -m canswim resolve_start --forecast_start_date 2026-03-05
python -m canswim forecast --tickers AAPL --forecast_start_date 2026-03-05 --dry_run
python -m canswim dashboard --same_data True

Full recipes: docs/cli.md.

Dashboard (GUI)

python -m canswim dashboard --same_data True

Tab	Purpose
Charts	Price history + forecast bands for a symbol
Scans	Filter forecasts by as-of date, reward, risk, confidence
Run	Update market data / Run forecast / Check start date
Advanced Queries	Read-only SQL against the search DB

Sample screenshots

(If images are missing in a fork, open the dashboard locally and refresh captures under docs/images/ — same PR as any UI change.)

Historical example chart:

Command line (quick)

python -m canswim -h

Main tasks: dashboard, gatherdata, forecast, resolve_start, mcp, train, modelsearch, downloaddata, uploaddata.

Flag	Used by	Meaning
--tickers	gatherdata, forecast	Scoped run via shared orchestration
--forecast_start_date	forecast, resolve_start	Origin date (week-aligned for scoped runs)
--dry_run	forecast --tickers	Resolve start + validate only
--no_covariates	gatherdata --tickers	Prices only (skip fundamentals)
--same_data	dashboard	Reuse DuckDB search DB
--new_model	train	Fresh model vs continue

→ docs/cli.md

MCP server (quick)

Read-only by default. Write tools need MCP_ALLOW_RUNS=1.

python -m canswim mcp
MCP_ALLOW_RUNS=1 python -m canswim mcp

→ docs/mcp.md for tools, client config, and prerequisites.