heare-histograms 0.1.2

Pure python histogram implementation for analysis and prediction.

Heare Histograms Library

This library provides two main classes: Histogram and Group. The Histogram class is a generic data structure for storing and sampling data points, while the Group class represents a collection of Histogram instances or other Group instances. Groups can be used to generate synthetic composite histograms.

Installation

To install the library, you can use pip:

pip install heare-histograms

Usage

Histogram

The Histogram class can be imported and used as follows:

from heare.histograms import Histogram, Shift, Scale, Max

# Create a new histogram
hist = Histogram[int](max_size=1000)

# Observe data points
hist.observe([1, 2, 3, 4, 5])

# Sample a data point
sample = hist.sample()

# Get the value at a given percentile
percentile_val = hist.percentile(0.9)

# Combine two histograms
hist2 = Histogram[int](data=[10, 20, 30])
combined = hist + hist2

# Apply transformations to the histogram
shifted_hist = Shift(hist, 10)
scaled_hist = Scale(hist, 2.0)
capped_hist = Max(hist, 100)

The Histogram class supports various operations, such as observing new data points, sampling, computing percentiles, combining histograms, and applying transformations like shifting, scaling, and capping values.

Group

The Group class can be imported and used as follows:

from heare.histograms import Histogram, Group

# Create a new group with two histograms
hist1 = Histogram[int](data=[1, 2, 3])
hist2 = Histogram[int](data=[10, 20, 30])
group = Group[int]([("hist1", hist1), ("hist2", hist2)])

# Sample a value from the group
sample = group.sample()

# Explain the sampled value
explanation = group.explain_sample()

The Group class allows you to combine multiple Histogram instances or other Group instances into a single entity. You can sample values from the group and obtain an explanation of the sampled value, breaking it down into contributions from each element.

Contributing

Contributions to the library are welcome! If you find any issues or have suggestions for improvements, please open an issue or submit a pull request on the GitHub repository.

License

This library is released under the MIT License.