yfinance-ta-patterns 0.1.3

CLI and helpers to scan yfinance data for TA-Lib candlestick patterns.

Forex Candlestick Scanner

Python package + CLI that downloads Yahoo Finance data via yfinance and runs TA-Lib candlestick detectors for your symbol, timeframe, and date filters.

Quick start

# create env
python -m venv .venv
.\.venv\Scripts\activate  # PowerShell; adjust for your shell

# install the package (editable for local dev)
pip install -e .

# run CLI (two entrypoints)
yfinance-ta-patterns --pattern KICKING --symbol EURUSD --timeframe 5m --period 60d
# or
yftp --all-patterns --symbol EURUSD --timeframe 5m --period 60d --date 2025-04-01

pyproject.toml targets Python >=3.12.

CLI usage

yfinance-ta-patterns [--pattern NAME | --all-patterns]
                     [--symbol EURUSD] [--period 60d]
                     [--timeframe 15m] [--date YYYY-MM-DD]
                     [--start-date YYYY-MM-DD] [--end-date YYYY-MM-DD]

• --pattern: Single candlestick name (with or without CDL prefix).
• --all-patterns: Scan every TA-Lib candlestick detector.
• --symbol: Ticker without suffix; =X is appended automatically for Forex (default EURUSD).
• --period: History window passed to yfinance (e.g., 60d, 1mo).
• --timeframe: Use M1/M5/M15/M30/H1/H4/D1 or raw yfinance intervals (1m, 5m, 1h, 4h, 1d, etc.). H4/4h is produced by resampling 1h data.
• --date: Filter signals for a single day.
• --start-date / --end-date: Inclusive range filter (cannot be combined with --date).

Examples

• All patterns for a single day:

yftp --all-patterns --symbol EURUSD --timeframe 5m --period 60d --date 2025-04-01

• All patterns across a range:

yftp --all-patterns --symbol EURUSD --timeframe 5m --period 60d --start-date 2025-04-01 --end-date 2025-04-10

• One pattern without date filter:

yftp --pattern KICKING --symbol EURUSD --timeframe 5m --period 60d

Data loader

yfinance_ta_patterns/forex_data_loader.py fetches and normalizes OHLC data. It appends =X to symbols when missing and converts timestamps to UTC before shifting to the configured timezone (Europe/Moscow by default).

H4 timeframes are fetched as 1h data and resampled to 4-hour candles.

Pattern analysis

yfinance_ta_patterns/pattern_analyzer.py wraps TA-Lib's CDL* functions, returning non-zero signals and applying optional date filters. When --all-patterns is used, it iterates over the full catalog and prints hits per pattern.

Pattern ranking helper (optional)

yfinance_ta_patterns/pattern_tester.py contains a backtesting-style ranking tool. It depends on utils.pattern_helper.PatternHelper to enumerate patterns; add that helper before running comparisons or exports.

Project layout

• yfinance_ta_patterns/cli.py: CLI entry point and argument parsing.
• yfinance_ta_patterns/forex_data_loader.py: Data download and timezone normalization.
• yfinance_ta_patterns/pattern_analyzer.py: Candlestick signal extraction.
• yfinance_ta_patterns/pattern_tester.py: Experimental ranking/backtest utilities.
• main.py: Thin wrapper to launch the CLI.

Packaging and releases

• Nightly GitHub Actions workflow builds onefile Nuitka binaries for Windows/macOS/Linux (PyPI wheels for TA-Lib on all platforms) and publishes nightly prereleases.
• PyPI publish workflow runs on tags (needs secret PYPI_API_TOKEN); tags like v0.1.0 will build sdist/wheel and upload.