Separate series ops from imperative ops in Pluggable Analysis Framework

[paddymul]

Checks

• I have checked that this enhancement has not already been requested

How would you categorize this request. You can select multiple if not sure

Summary stats, Performance, Developer Experience/CI (feature to make it easier to devlop on Buckaroo)

Enhancement Description

For PAF most of the runtime is take up by series operations (nlargest, mean, median)

Most of the business logic is in regular python code. Look at assembling histograms as an example.

Pseudo Code Implementation

Refactor PAF so it works as follows, for polars in particular

have a series_selector function that returns statements like

class NewAnalysis:
    provides_series = ["mean", "bottom", "top"]
    @static_method
    def series_selector():
        return [
                    F.all().mean().name.prefix("mean:")
                    F.col(pl.Int64, pl.UInt32).bottom_k(5).name.prefix("bottom:"),
                    F.col(pl.Int64, pl.UInt32).top_k(5).name.prefix("top:")]

Then pluggable analysis framework works to make the rest of the logic work with regular python dag stuff

This way we can implement pandas logic and polars series methods, and have the rest of the python code work the same.

Prior Art

N/A

[paddymul]

closed by #182