stock-kit

Local-first A-share stock research toolkit with data adapters, indicator engines, signal parsing, strategy helpers, validation gates, and reinforcement utilities.

This repository is packaged as a public-safe toolkit. It does not include personal watchlists, brokerage credentials, cookies, generated databases, private research logs, or local browser profiles.

What It Does

• Fetches market data through an adapter layer, with AkShare as an optional live provider.
• Stores local market bars, signals, predictions, rule changes, and indicators in SQLite.
• Computes technical, volume, VCP, CANSLIM, and Wyckoff-style indicator signals.
• Provides strategy helpers for macro health, playbooks, screening, reasoning, and trade-plan construction.
• Supports validation gates for auction, intraday, feasibility, freshness, and post-close review.
• Includes optional read-only X and Reddit monitors for public signal capture experiments.
• Provides reinforcement scoring and rule-tuning helpers for reviewing past predictions.

Repository Shape

src/stockkit/
  data/           market data providers and fallback chain
  db/             SQLite schema and repositories
  indicators/     technical and fundamental-style signal engines
  signals/        public signal parsing and optional read-only monitors
  strategy/       screening, playbooks, reasoning, and trade planning
  validation/     auction, intraday, freshness, and post-close checks
  reinforcement/  scorer and tuning helpers
tests/            unit and integration coverage

Quick Start

python -m pip install -e ".[dev]"
python -m pytest
stock-kit --help
stock-kit fetch daily --tickers 000001,600519
stock-kit scan

Optional public signal monitors need extra dependencies:

python -m pip install -e ".[monitors]"
stock-kit monitor reddit --subs stocks,investing
stock-kit monitor x --accounts example_research,sector_watch

The X monitor expects a local browser that you explicitly start with a DevTools port. It does not log in, post, vote, follow, or perform account-changing actions.

Public Safety

• Local SQLite databases are ignored under data/.
• .env, logs, pytest output, build artifacts, and generated output folders are ignored.
• Internal implementation handoffs and machine-local planning docs are not part of this public package.
• Examples use placeholder accounts and public subreddit names.
• The project is for research tooling and software demonstration only. It is not financial advice and does not place trades.

Development

python -m pip install -e ".[dev,monitors]"
python -m pytest
python -m stockkit --version

The test suite is designed to run without live brokerage access or private credentials.