py-pma

A Python library for computing portfolio management analytics on normalized weight mappings.

About

Assumptions

• These functions assume long-only, normalized weights.
• load_weights_csv accepts raw values or pre-scaled weights and normalizes them to sum to 1.0.
• The core metrics operate on generic labeled weight mappings. In the common case, the labels are security identifiers, but they can also be sectors, countries, industries, sleeves, or other buckets.

Usage

from pma import active_share

portfolio = {"AAPL": 0.5, "MSFT": 0.4, "XOM": 0.1}
benchmark = {"AAPL": 0.1, "NVDA": 0.5, "XOM": 0.4}

result = active_share(portfolio, benchmark)
print(result)

Grouping

You can aggregate one labeled weight mapping into broader groups and then reuse the same metrics at the grouped level.

from pma import active_share, group_weights

portfolio = {"AAPL": 0.5, "MSFT": 0.4, "XOM": 0.1}
benchmark = {"AAPL": 0.1, "NVDA": 0.5, "XOM": 0.4}
sector_map = {
    "AAPL": "Information Technology",
    "MSFT": "Information Technology",
    "NVDA": "Information Technology",
    "XOM": "Energy",
}

portfolio_sectors = group_weights(portfolio, sector_map)
benchmark_sectors = group_weights(benchmark, sector_map)

result = active_share(portfolio_sectors, benchmark_sectors)
print(result)

group_weights excludes zero-weight groups by default. Pass include_zero_weights=True if you want them preserved.

You can also load the weights and group mapping from separate CSV files.

from pma import (
    active_share,
    group_weights,
    load_group_map_csv,
    load_weights_csv,
)

portfolio = load_weights_csv("portfolio.csv")
benchmark = load_weights_csv("benchmark.csv")
group_map = load_group_map_csv("sectors.csv")

portfolio_groups = group_weights(portfolio, group_map)
benchmark_groups = group_weights(benchmark, group_map)

result = active_share(portfolio_groups, benchmark_groups)
print(result)