pnf-chart-system

Production-ready Python bindings for the PnF (Point and Figure) engine.

Package name is pnf-chart-system; import name is pypnf.

Why This Package

pypnf is built for real analysis workflows, not only chart construction:

What Is New In 0.2.0

• Multi-symbol scanner: scan, rank_by_bullish_percent, latest_patterns, fresh_breakouts, near_support, near_resistance.
• ML features: FeatureVector, FeatureMatrix, to_numpy, to_pandas, to_polars, to_pyarrow.
• Data adapters: CSV, MetaTrader, TradingView, dataframe conversion, parquet, and optional Yahoo Finance.
• Pattern quality scoring, explicit event stream events, walk-forward optimization, and publication-quality plot exports.
• Release banner: version sync, docs coverage, Python tests, and market-data smoke validation are part of the release checklist.

Installation

pip install pnf-chart-system

Quick Start

import pypnf

cfg = pypnf.ChartConfig()
cfg.method = pypnf.ConstructionMethod.HighLow
cfg.box_size_method = pypnf.BoxSizeMethod.Traditional
cfg.box_size = 0.0
cfg.reversal = 3

chart = pypnf.Chart(cfg)

# high, low, close, timestamp
chart.add_data(5000.0, 4950.0, 4985.0, 1700000000)
chart.add_data(5040.0, 4980.0, 5030.0, 1700003600)
chart.add_data(5065.0, 5010.0, 5055.0, 1700007200)

indicators = pypnf.Indicators(pypnf.IndicatorConfig())
indicators.calculate(chart)

print(chart.to_ascii())
print(indicators.summary())

Trendline and Bias Workflow

last_price = 5055.0

print("Bullish bias:", chart.has_bullish_bias())
print("Bearish bias:", chart.has_bearish_bias())
print("Above bullish support:", chart.is_above_bullish_support(last_price))
print("Below bearish resistance:", chart.is_below_bearish_resistance(last_price))

These checks are the normal first gate before acting on breakout or breakdown patterns.

Indicators and Momentum

indicators.calculate(chart)

sma_short = indicators.sma_short()
bands = indicators.bollinger()
rsi = indicators.rsi()
obv = indicators.obv()

col = chart.column_count() - 1
if col >= 0:
    print("SMA short:", sma_short.value(col))
    print("Bollinger upper:", bands.upper(col))
    print("RSI:", rsi.value(col))
    print("OBV:", obv.value(col))

Signals and Pattern Detection

signals = indicators.signals()
patterns = indicators.patterns()

print("Current signal:", signals.current_signal())
print("Buy count:", signals.buy_count())
print("Sell count:", signals.sell_count())

print("Pattern count:", patterns.pattern_count())
print("Bullish patterns:", len(patterns.bullish_patterns()))
print("Bearish patterns:", len(patterns.bearish_patterns()))

Support, Resistance, Objectives, Congestion

sr = indicators.support_resistance()
obj = indicators.objectives()
cong = indicators.congestion()

print("Support levels:", sr.support_levels())
print("Resistance levels:", sr.resistance_levels())
print("Significant levels (>=3 touches):", sr.significant_levels(3))

print("Bullish targets:", obj.bullish_targets())
print("Bearish targets:", obj.bearish_targets())

print("Congestion zones:", cong.zones())

Real-Time Dashboard

from pypnf_dashboard import DashboardServer

server = DashboardServer(chart, indicators)
server.start("127.0.0.1", 8761)
server.publish()
print(server.url())

You can call server.publish() after each new bar/tick batch to keep the browser in sync.

End-To-End Quant Workflow

bars = pypnf.DataAdapters.load_csv("tests/fixtures/GBPUSD_PERIOD_M1.csv")

chart = pypnf.Chart(cfg)
for bar in bars:
    chart.add_ohlc(bar)

indicators.calculate(chart)

features = pypnf.extract_features(chart, indicators)
scores = pypnf.score_patterns(chart, indicators)
matrix = pypnf.feature_matrix([chart], pypnf.IndicatorConfig()).to_pandas()

series = pypnf.SymbolSeries()
series.symbol = "GBPUSD"
series.data = bars
scan_result = pypnf.scan([series], pypnf.ScanConfig())
leaders = pypnf.rank_by_bullish_percent(scan_result)

html = pypnf.PlotExporter.to_html(chart, indicators, pypnf.PlotExportConfig())

Use local fixtures or your own market exports for CI-stable validation. Optional dataframe and Yahoo helpers import their provider packages lazily.

API Map

Core:

• Chart, ChartConfig, Box, Column

Indicators:

• Indicators, IndicatorConfig
• MovingAverage, BollingerBands, RSI, OnBalanceVolume, BullishPercent
• SignalDetector, PatternRecognizer, SupportResistance, PriceObjectiveCalculator, CongestionDetector

Data:

• OHLC, Signal, Pattern, SupportResistanceLevel, PriceObjective, CongestionZone

Quant:

• DataAdapters, FeatureVector, FeatureMatrix, PatternQualityScore
• scan, scanner filters, WalkForwardConfig, walk_forward_optimize
• PlotExporter.to_svg, to_png, to_html, to_plotly_json, to_matplotlib

Enums:

• BoxType, ColumnType, ConstructionMethod, BoxSizeMethod, SignalType, PatternType

Versioning and Compatibility

• Python package version tracks the same release as the core engine.
• Keep all bindings on the same version when mixing languages in one system.
• See CHANGELOG.md for version-by-version behavior changes.

Troubleshooting

• ImportError / native load issues: rebuild and ensure the native library is discoverable.
• Empty indicator values: verify you have enough columns for the configured lookback periods.
• Unexpected chart shape: ensure HighLow mode receives real high/low values, not close-only values.

Documentation and Links

• Python API reference: docs/bindings/python.md
• Cross-language API index: docs/reference/api-symbol-index.md
• Source: https://github.com/gregorian-09/pnf-chart-system
• Issues: https://github.com/gregorian-09/pnf-chart-system/issues
• Changelog: https://github.com/gregorian-09/pnf-chart-system/blob/master/CHANGELOG.md