tick-to-ohlcv

Fast, exchange-agnostic trade tick aggregation for market data pipelines.

tick-to-ohlcv converts raw trade CSVs into clean OHLCV candles with volume, turnover, trade counts, taker buy/sell volume, optional flat gap candles, and CSV or Parquet output. It is built for crypto, FX, equities, and any market where raw ticks need to become analysis-ready candles.

Why Use It

• Generic CSV mapping: point the tool at your timestamp, price, size, symbol, side, turnover, and trade-id columns.
• Fast Rust CLI: aggregate large folders of trade files without writing one-off scripts.
• Python compatibility: use the same core ideas from Python data pipelines.
• Flexible intervals: create 1m, 5m, 1h, daily, or custom-second candles.
• Gap filling: emit zero-volume flat candles between observed trades when you need continuous time series.
• CSV and Parquet output: write quick CSVs or columnar datasets for analytics.
• Folder discovery: include and exclude file patterns under a root instead of listing every input file.

Install

Build the Rust CLI:

cargo build --release

Install the Python package:

pip install tick-to-ohlcv

Install Python with optional Parquet support:

pip install "tick-to-ohlcv[parquet]"

Quick Start

Convert one CSV file to one-minute candles:

tick-to-ohlcv csv trades.csv \
  --symbol BTCUSDT \
  --timestamp-column timestamp_ms \
  --price-column price \
  --size-column size \
  --turnover-column quote_volume \
  --side-column side \
  --trade-id-column trade_id \
  --interval 1m \
  --output candles.csv

Discover many files, fill gaps, and write Parquet:

cargo run --release -- csv \
  --input-root /data/raw-trades \
  --include "**/*.csv" \
  --exclude "**/bad/*.csv" \
  --symbol-column symbol \
  --timestamp-column timestamp_ms \
  --price-column price \
  --size-column size \
  --turnover-column notional \
  --side-column side \
  --interval 5m \
  --fill-gaps \
  --output-format parquet \
  --output /data/candles.parquet

Use second timestamps instead of milliseconds:

tick-to-ohlcv csv fills.csv \
  --symbol ETHUSD \
  --timestamp-column time \
  --timestamp-unit s \
  --price-column px \
  --size-column qty \
  --interval 1h \
  --output eth_1h.csv

Output Columns

Candles use Unix-second interval starts and ISO UTC datetimes:

symbol, ts, datetime, open, high, low, close, volume, turnover,
trade_count, buy_volume, sell_volume, first_trade_ts_ms, last_trade_ts_ms

Notes:

• volume is base-asset volume.
• turnover is quote/notional volume when provided, otherwise price * size.
• buy_volume and sell_volume are taker-side base volumes from the mapped side column.
• Gap candles have trade_count = 0, zero volume/turnover, and flat OHLC.

Python Library Usage

from pathlib import Path

from tick_to_ohlcv import CsvTradeMapping, aggregate_csv_files, discover_files, write_candles
from tick_to_ohlcv.intervals import parse_interval_seconds

files = discover_files(
    input_root=Path("/data/raw-trades"),
    include_patterns=["**/*.csv"],
    exclude_patterns=["**/bad/*.csv"],
)

mapping = CsvTradeMapping(
    symbol_column="symbol",
    timestamp_column="timestamp_ms",
    price_column="price",
    size_column="size",
    turnover_column="notional",
    side_column="side",
    trade_id_column="trade_id",
)

candles = aggregate_csv_files(
    files,
    mapping,
    interval_seconds=parse_interval_seconds("1m"),
    fill_gaps=True,
)

write_candles(candles, output=Path("/data/candles.csv"))

Good Fits

• Rebuilding OHLCV candles from public trade archives.
• Normalizing multiple venues into one candle schema.
• Creating research datasets from raw fills.
• Backtesting and market microstructure workflows.
• Converting tick data to Parquet for warehouse or lakehouse ingestion.

Suggested GitHub Description

Fast Rust/Python CLI for converting raw trade ticks into OHLCV candles, CSV, and Parquet.

Suggested GitHub Topics

ohlcv, tick-data, market-data, crypto, trading, parquet, rust-cli, python, csv, candles