orderpulse 0.2.22

High-performance exchange feed parser and orderflow analytics engine with Rust and Python bindings

fastreader

High-performance Rust + PyO3 library for reading binary market data and generating order book snapshots directly from Python.

fastreader is designed for Python quant developers who need:

• Fast binary message parsing
• Order and trade message inspection
• Token-level filtering
• Sequential message streaming
• Order book reconstruction
• Top-of-book / depth snapshot generation

The library is implemented in Rust for speed and exposed to Python using PyO3.

---

Features

• Parse binary market feed files
• Read order and trade messages
• Filter data by token/instrument
• Sequential message iteration
• Generate order book snapshots
• Configurable market depth
• Fast memory-efficient Rust backend
• Python-friendly API

---

Installation

Build Using maturin

pip install maturin
maturin develop --release

OR build wheel:

maturin build --release

---

Python Import

import fastreader

---

Main Classes

Class	Description
BinaryDataLoader	Loads and reads binary market data
OrderBookGenerator	Generates order book snapshots

---

BinaryDataLoader

The BinaryDataLoader class loads binary feed data into memory and provides utilities to inspect messages.

---

Constructor

BinaryDataLoader(path, token=None)

Parameters

Parameter	Type	Description
path	str	Path to binary market data file
token	int or None	Optional instrument token filter

---

Example

from fastreader import BinaryDataLoader

loader = BinaryDataLoader(
    path="data/feed.bin",
    token=26000
)

---

BinaryDataLoader Functions

---

total_messages()

Returns total number of messages loaded.

Returns

int

Example

count = loader.total_messages()
print(count)

---

total_orders()

Returns total order messages.

Returns

int

Example

orders = loader.total_orders()
print(orders)

---

total_trades()

Returns total trade messages.

Returns

int

Example

trades = loader.total_trades()
print(trades)

---

summary()

Returns formatted dataset summary.

Returns

str

Example

print(loader.summary())

Sample Output

Total Messages: 250000
Total Orders: 180000
Total Trades: 70000

---

reset_cursor()

Resets internal sequential message cursor.

Useful when using get_next_message().

Example

loader.reset_cursor()

---

get_all_messages(limit=None)

Returns all messages as formatted strings.

Parameters

Parameter	Type	Description
limit	int or None	Maximum messages to return

Returns

List[str]

Example

messages = loader.get_all_messages(limit=5)

for msg in messages:
    print(msg)

Sample Output

Order Message: SeqNo1, msg_len64 ...
Trade Message: SeqNo2, msg_len48 ...

---

get_order_messages(limit=None)

Returns only order messages.

Parameters

Parameter	Type	Description
limit	int or None	Maximum messages to return

Returns

List[str]

Example

orders = loader.get_order_messages(limit=10)

for order in orders:
    print(order)

---

get_trade_messages(limit=None)

Returns only trade messages.

Parameters

Parameter	Type	Description
limit	int or None	Maximum messages to return

Returns

List[str]

Example

trades = loader.get_trade_messages(limit=10)

for trade in trades:
    print(trade)

---

get_next_message()

Returns next message sequentially.

Internal cursor automatically moves forward after each call.

Returns

str

Returns:

• Message string
• "END" when all messages are consumed

Example

while True:
    msg = loader.get_next_message()

    if msg == "END":
        break

    print(msg)

---

OrderBookGenerator

The OrderBookGenerator reconstructs order books from order/trade messages.

It produces snapshot rows with configurable market depth.

---

Constructor

OrderBookGenerator(depth=5)

Parameters

Parameter	Type	Description
depth	int	Number of bid/ask levels

---

Example

from fastreader import OrderBookGenerator

ob = OrderBookGenerator(depth=5)

---

OrderBookGenerator Functions

---

reset()

Resets internal order book state.

Example

ob.reset()

---

header()

Returns CSV-style order book header.

Returns

str

Example

print(ob.header())

Sample Output

local_ts,exch_ts,mid_price,bid_price_0,bid_qty_0,ask_price_0,ask_qty_0

---

create_orderbook(store)

Default order book generation method.

Internally calls:

create_orderbook_from_all_messages()

Parameters

Parameter	Type	Description
store	BinaryDataLoader	Loaded binary dataset

Returns

List[str]

Example

rows = ob.create_orderbook(loader)

for row in rows[:5]:
    print(row)

---

create_orderbook_from_all_messages(store)

Generates order book snapshots using:

• Order messages
• Trade messages

Parameters

Parameter	Type	Description
store	BinaryDataLoader	Loaded binary dataset

Returns

List[str]

Example

rows = ob.create_orderbook_from_all_messages(loader)

---

create_orderbook_from_trade_messages(store)

Generates snapshots using only trade messages.

Parameters

Parameter	Type	Description
store	BinaryDataLoader	Loaded binary dataset

Returns

List[str]

Example

trade_rows = ob.create_orderbook_from_trade_messages(loader)

---

create_orderbook_from_next_messages(store)

Generates order book sequentially using cursor-based streaming.

Useful for:

• Streaming simulations
• Backtesting engines
• Sequential replay systems

Parameters

Parameter	Type	Description
store	BinaryDataLoader	Loaded binary dataset

Returns

List[str]

Example

loader.reset_cursor()

rows = ob.create_orderbook_from_next_messages(loader)

---

row_count_from_all_messages(store)

Returns total generated snapshot rows.

Parameters

Parameter	Type	Description
store	BinaryDataLoader	Loaded binary dataset

Returns

int

Example

count = ob.row_count_from_all_messages(loader)
print(count)

---

Complete Python Example

from fastreader import BinaryDataLoader
from fastreader import OrderBookGenerator

# Load binary feed
loader = BinaryDataLoader(
    path="data/feed.bin",
    token=26000
)

# Dataset information
print(loader.summary())

# Print first 5 messages
messages = loader.get_all_messages(limit=5)

for msg in messages:
    print(msg)

# Create order book generator
ob = OrderBookGenerator(depth=5)

# Generate snapshots
rows = ob.create_orderbook(loader)

# Print first rows
for row in rows[:10]:
    print(row)

---

Order Book Output Format

Each snapshot row contains:

Field	Description
local_ts	Local receive timestamp
exch_ts	Exchange timestamp
mid_price	Mid market price
bid_price_n	Bid price at level n
bid_qty_n	Bid quantity at level n
ask_price_n	Ask price at level n
ask_qty_n	Ask quantity at level n

---

Performance Notes

Since parsing and processing are implemented in Rust:

• Much faster than pure Python
• Lower memory overhead
• Efficient sequential processing
• Suitable for HFT and quant research workflows

Error Handling

If binary parsing fails:

RuntimeError

Typical reasons:

• Invalid file path
• Corrupted binary file
• Unsupported binary structure

---

Suggested Project Structure

project/
│
├── data/
│   └── feed.bin
│
├── scripts/
│   └── analysis.py
│
└── fastreader/

---

Example Workflow for Quant Developers

from fastreader import BinaryDataLoader
from fastreader import OrderBookGenerator
import pandas as pd

loader = BinaryDataLoader("data/feed.bin")

ob = OrderBookGenerator(depth=5)

rows = ob.create_orderbook(loader)

header = rows[0].split(",")
data = [r.split(",") for r in rows[1:]]

# Convert to DataFrame
book_df = pd.DataFrame(data, columns=header)

print(book_df.head())

---

Future Improvements

Possible future extensions:

• Direct NumPy output
• Pandas integration
• Multi-threaded parsing
• Async streaming
• Arrow/Parquet export
• Zero-copy Python buffers
• Real-time feed support

---

License

MIT License