NubraCollector

nubracollector is a desktop collector and Python helper library for Nubra realtime market data.

It combines:

• a PySide6 desktop UI for login, stream selection, field filtering, and live tables
• a reusable Python API for starting Nubra websocket streams in code
• local storage outputs for CSV, Excel, SQLite, and Parquet
• first-run instrument catalog caching for searchable instrument selection

Features

• Supports all five Nubra websocket stream types:
  • index
  • option
  • orderbook
  • greeks
  • ohlcv
• Compact UI login flow with environment, phone, OTP or TOTP, and MPIN
• Dark and light themes
• Live stream preview table with selected-field filtering
• Searchable instrument picker with cached instrument master data
• Stream output storage to:
  • CSV
  • Excel
  • SQLite
  • Parquet

Installation

From a local checkout:

pip install .

After publishing:

pip install nubracollector

Launch The Desktop UI

nubracollector

Or from Python:

from nubracollector import launch_ui

launch_ui()

Python Usage

from nubracollector import NubraCollector
from nubracollector.models import AuthConfig, StorageConfig, StorageFormat, SubscriptionConfig

collector = NubraCollector(
    AuthConfig(
        environment="UAT",
        use_env_credentials=True,
        phone_number="9999999999",
        mpin="0000",
    )
)

collector.login()

collector.start(
    SubscriptionConfig(
        stream_key="index",
        values=["NIFTY", "BANKNIFTY"],
        exchange="NSE",
        selected_fields=["indexname", "timestamp", "index_value", "volume"],
    ),
    StorageConfig(
        format=StorageFormat.SQLITE,
        target_path="market_data.db",
        table_name="index_ticks",
    ),
)

Project Layout

nubracollector/
  nubracollector/
    app.py
    auth.py
    collector.py
    instrument_cache.py
    models.py
    normalizers.py
    schemas.py
    session.py
    storage.py
    ui/

Notes

• The UI stores auth tokens locally for SDK reuse between steps.
• Instrument catalog refresh requires a valid Nubra session.
• CSV output is written immediately while streaming.
• Default output filenames include a timestamp to avoid collisions between runs.

Build A Distribution

python -m build

This creates files in dist/ ready for upload to PyPI.