Metadata-Version: 2.4
Name: numin
Version: 2.0.6
Summary: numin package
Author-email: Gautam Shroff <gautam.shroff@iiitd.ac.in>
Project-URL: Homepage, https://github.com/pypa/sampleproject
Project-URL: Issues, https://github.com/pypa/sampleproject/issues
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Requires-Python: >=3.10
Description-Content-Type: text/markdown
Requires-Dist: anvil_uplink<0.6.0,>=0.5.1
Requires-Dist: import_ipynb<0.2.0,>=0.1.4
Requires-Dist: ipython<9.0.0,>=8.12.3
Requires-Dist: numpy<1.25.0,>=1.24.3
Requires-Dist: pandas<2.0.0,>=1.5.3
Requires-Dist: pandas_ta==0.3.14b0
Requires-Dist: pytz<2025.0,>=2024.1
Requires-Dist: requests<3.0.0,>=2.32.3
Requires-Dist: scikit-learn<2.0.0,>=1.6.1
Requires-Dist: yfinance<0.2.59,>=0.2.54
Requires-Dist: torch>=2.4.0

# numin Package

**numin** is a Python package designed for algorithmic trading and backtesting providing an API called **NuminAPI**.

**numin v1** is out of service as of Dec 2025

**numin v2** is under development; features available are documented below

## Features

- **Data Retrieval:** Download training, round, and validation data.
- **Prediction Submission:**  TBD
- **Real-Time Round Management:** TBD
- **Backtesting:** TBD
- **File Management:** TBD
- **Returns Summary:** TBD

## Supported Methods

- **Data Download:**
    - `get_data_for_month(self,year,month,batch_size=4,window_size=100,target_type='rank'):`
    -   Returns a torch dataloader for the given year and month of Nifty 50 returns
    -   Dimension of each day is 100,50. Returns tensor of shape batch_size,window_size,50 for features.
    -   Targets are next day returns / ranked returns of shape batch_size,50
- **Backtesting:** 
    - `compute_pnl(positions,targets):`
    -  Run backtests - given positions n,50 dimensional 1,0,-1 for each stock per day
       and targets from dataloader for test month.
    -  Returns daily_pnl, Sharpe, total pnl etc.
    
## Installation

Install numin using pip:

```bash
pip install numin

