Python Indicators for Indian Stock Charts

What it does

Python indicators let you turn any calculation you can express in Python into a chart overlay or a separate indicator pane on TradingLead. If you already analyse Nifty or Bank Nifty data in pandas, you can bring the same logic onto a live chart.

There is no new language to learn. You write a normal function:

def calculate(bars, params):
    closes = [bar["close"] for bar in bars]
    return {
        "title": "Close",
        "overlay": True,
        "plots": [{"id": "close", "title": "Close", "color": "#2962ff", "values": closes}],
        "drawings": [],
    }

How it works

1. Input: bars is a chronological list of OHLCV dictionaries (time in Unix milliseconds, open, high, low, close, volume; missing volume is None). params is a JSON-style dictionary of your settings.
2. Output: return a dictionary with a title, overlay (on the price chart or in its own pane), a list of plots and an optional list of drawings.
3. Runtime:
   • In the browser, compatible code runs in a pinned Pyodide build inside a fresh, isolated iframe. Packages load first; then network APIs are disabled. The sandbox is destroyed after each calculation.
   • If the TradingLead Desktop app is installed, it becomes the runtime automatically. It is required for tradinglead SDK/Strategy imports, native or OS-level modules, arbitrary packages and workloads that need hard memory/process limits.
   • If your code needs Desktop, Studio tells you exactly why and links to the desktop app.

Key capabilities

• Plain Python: functions, loops, classes, math, statistics, plus NumPy and pandas in the browser. NumPy arrays and pandas Series are accepted as plot values.
• Plot styles: line, linebr, stepline, histogram, columns, area, cross, circles, plus candles and bars for OHLC output. Per-bar colours are supported.
• Drawings: text, rectangles, trend lines, horizontal and vertical lines, rays, extended lines and up/down arrows, anchored to exact candle timestamps.
• Typed settings: add # @input comments to get form controls — int, float, bool, string, enum (dropdown) and color (picker), with titles, groups and bounds.
• Multiple timeframes and symbols: declare extra datasets with # @data comments and read them from an optional third context argument. Secondary datasets contain closed candles only, from the same provider — no silent fallback to another exchange.
• Named plot settings: each plot keeps its own name, colour, width and style in a TradingLead settings panel. Apply style does not rerun Python; Apply & recalculate does.
• AI prompt panel: generates a complete, editable authoring prompt (Indicator or Backtest strategy) to paste into your own AI; your code is included only if you tick the option.
• Saved indicators: saved Python indicators appear under My saved indicators in the chart's Indicators dialog.

How to use it

Follow Write a Python indicator on TradingLead for a full walkthrough, from a simple moving average to a multi-timeframe pandas example.

When to choose Python vs Pine

Both run in Indicator Studio and both can be saved and added from the Indicators dialog.

Limitations

• Browser Python is limited to the standard library, NumPy and pandas. Everything else requires the Desktop app's isolated package environment, which is set up explicitly — scripts cannot install packages themselves.
• Return exactly one value per primary candle per plot, using None for warmup and gaps.
• Desktop runtime limits: 100,000 combined primary and secondary candles, 64 MiB serialized input/output, 20,000 visual objects, 2 million output values, 15 seconds per calculation and 512 MiB memory. Results are never silently truncated — oversize output fails with an explicit error.
• No internet access during execution.
• Vectorized strategy backtests do not verify lookahead; write causal code. Backtests are simulations, not guarantees.