Deep MM Python Client

Python client library for the Deep MM API - Real-time and historical corporate bond pricing.

Installation

pip install axor-api

Quick Start

import asyncio
import json
from axor import connect, create_get_id_token

async def main():
    # Authenticate with AWS Cognito
    get_id_token = create_get_id_token(
        region="us-east-1",
        client_id="2so174j2e4fsg1m28kc9id3hgk",  # Test client ID
        username="your-username",
        password="your-password"
    )

    # Connect to the Deep MM API
    ws = await connect()

    # Send authentication token
    await ws.send(json.dumps({'token': get_id_token()}))

    # Subscribe to real-time pricing
    subscription = {
        'inference': [{
            'rfq_label': 'spread',
            'figi': 'BBG003LZRTD5',
            'quantity': 1_000_000,
            'side': 'bid',
            'subscribe': True,
        }]
    }
    await ws.send(json.dumps(subscription))

    # Receive updates
    while True:
        response = await ws.recv()
        print(response)

asyncio.run(main())

Features

• Authentication: AWS Cognito integration with automatic token refresh
• Real-time Subscriptions: WebSocket-based live pricing updates
• Historical Data: Query historical pricing distributions
• CUSIP/FIGI Mapping: Convert CUSIPs to FIGIs using OpenFIGI API
• Distribution Fitting: Fit normal and Johnson SU distributions to percentile data

API Reference

Authentication

from axor import create_get_id_token

get_id_token = create_get_id_token(
    region="us-east-1",
    client_id="your-client-id",
    username="your-username",
    password="your-password"
)

# Token is automatically refreshed as needed
token = get_id_token()

Connection

from axor import connect

# Use default server (wss://api.deepmm.com)
ws = await connect()

# Or specify custom server
ws = await connect("wss://custom-server.com")

CUSIP to FIGI Mapping

from axor import openfigi_map_cusips_to_figis

cusip_to_figi, figi_to_cusip = openfigi_map_cusips_to_figis(
    api_key="your-openfigi-api-key",
    cusip_list=["594918BJ2", "037833100"]
)

# Use the mappings
figi = cusip_to_figi["594918BJ2"]

Distribution Fitting

from axor import fit_normal_distribution, fit_johnson_su

# Fit normal distribution to percentile data
percentiles = [5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95]
values = [...]  # Your percentile values from API response

mu, sigma, error = fit_normal_distribution(percentiles, values)
print(f"Mean: {mu}, Std Dev: {sigma}")

# Fit Johnson SU distribution (better for skewed/heavy-tailed data)
gamma, delta, loc, scale, error = fit_johnson_su(percentiles, values)

# Generate plots (requires matplotlib)
from axor import plot_cdf_of_fitted_johnson_su_distribution

plot_cdf_of_fitted_johnson_su_distribution(
    'output.png',
    percentiles,
    values,
    gamma, delta, loc, scale
)

Examples

Complete working examples are available in the examples/ directory:

• subscribe_simple.py - Basic subscription to real-time pricing
• subscribe.py - Advanced subscription with multiple variations
• timestamp_simple.py - Historical pricing queries
• timestamp_normal.py - Historical data with normal distribution fitting
• timestamp_johnson_su.py - Historical data with Johnson SU fitting

Installation for Development

# Clone the repository
git clone https://github.com/deepmarketmaking/api.git
cd api/python

# Install in editable mode with dev dependencies
pip install -e ".[dev,visualization]"

# Run tests
pytest

Requirements

• Python 3.8 or higher
• Dependencies:
  • boto3 - AWS Cognito authentication
  • websockets - WebSocket communication
  • httpx - HTTP requests for OpenFIGI API
  • numpy, scipy - Distribution fitting
  • pyarrow - Efficient data handling
  • tenacity - Retry logic
  • matplotlib (optional) - For visualization

Documentation

Full API documentation and additional examples are available at:

• Main Repository
• API Documentation

Support

For questions, issues, or feature requests:

• GitHub Issues
• Email: support@deepmm.com

License

Apache License 2.0 - See LICENSE for details.