Remove legacy code

[dt-woods]

This is a request to remove legacy code from ElectricityLCI (i.e., process_exchange_aggregator_uncertainty.py), which is rather poorly documented, unused by any modules in eLCI, and adds the unnecessary dependency, sympy, to requirements.txt.

My suggestion is remove this module, document the uncertainty methods in their new homes (e.g., generation.py), point to the legacy file that exists in v1.0.1, and remove sympy from requirements list.

[dt-woods]

Archive of process_exchange_aggregator_uncertainty.py for posterity. Removed because it is never called and adds unnecessary dependency to list.

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
#
# process_exchange_aggregator_uncertainty.py
#
##############################################################################
# REQUIRED MODULES
##############################################################################
import math

import numpy as np
import pandas as pd
from scipy.stats import t
from sympy import var, solve


##############################################################################
# MODULE DOCUMENTATION
##############################################################################
__doc__ = """This module is designed to compile the emission and generation
data to create emission factors, along with uncertainty information for all
relevant emission factors calculated using a log normal distribution.

The emission factors use the weight-based method developed by Troy Hawkins
([email protected]).

Note that these methods do not appear to be called by other modules in eLCI.

Last updated: 2023-11-06
"""
__all__ = [
    'compilation',
    'max_min',
    'uncertainty',
]


##############################################################################
# FUNCTIONS
##############################################################################
def compilation(db, total_gen):
    # Troy Method
    # Create a copy of database by substitution the NA emissions with zero
    db1 = db.fillna(value=0)

    # Remove all rows where emissions are not reported for second dataframe,
    # keeping the unreported emissions and facilities in separate database
    db2 = db.dropna()

    # This check is to make sure that the second database is not empty
    # after dropping NAs. If empty, then we only use first database.
    if db2.empty:
        ef1 = np.sum(db1.iloc[:, 1])/total_gen
        return ef1

    ef1 = np.sum(db1.iloc[:, 1])/total_gen
    ef2 = np.sum(db2.iloc[:, 1])/total_gen

    # Weight formula.
    #   NOTE: Why is the weight equal to total_gen/total_gen?
    weight = total_gen/total_gen
    final_ef = ef2*weight + (1 - weight)*ef1

    return final_ef


def uncertainty(db, mean_gen, total_gen, total_facility_considered):
    # This is the function for calculating log normal distribution parameters.
    # Troy Method
    data_1 = db

    df2 = pd.DataFrame([[0, 0]], columns=['Electricity', 'FlowAmount'])
    for i in range(len(data_1), total_facility_considered):
        data = data_1.append(df2, ignore_index=True)
        data_1 = data

    data = data_1
    mean = np.mean(data.iloc[:, 1])
    l,b = data.shape
    # Compute standard error of the mean (but call it sd...)
    sd = np.std(data.iloc[:, 1])/np.sqrt(l)

    # Obtain the weighted emissions factor (emissions per MWh)
    ef = compilation(db, total_gen)

    # Endpoints of the range that contains alpha percent of the distribution
    pi1, pi2 = t.interval(0.90, df=l-2, loc=mean, scale=sd)

    # Converting prediction interval to emission factors
    pi2 = pi2/mean_gen
    pi1 = pi1/mean_gen
    pi3 = (pi2 - ef)/ef;
    x = var('x')

    if math.isnan(pi3) == True:
        return (None, None)

    elif math.isnan(pi3) == False:
        # This method will not work if the interval limits are more than
        # 280% of the mean.
        if pi3 < 2.8:
            # Solve roots of the quadratic, ax^2 + bx + c = 0
            a = 0.5
            b = -(1.16308*np.sqrt(2))
            c = np.log(1+pi3)
            sd1 = (-b + np.sqrt(b**2 - (4 * a * c))) / (2 * a)
            sd2 = (-b - np.sqrt(b**2 - (4 * a * c))) / (2 * a)
        else:
            # This is a wrong mathematical statement.
            # However, we have to use it if something fails.
            sd1, sd2 = solve(0.5*x*x - (1.36*np.sqrt(2))*x + (np.log(1+pi3)),x)

        # Always choose lower standard deviation from solving the square root
        # equation.
        if sd1 < sd2:
            log_mean = np.log(ef) - 0.5*(sd1**2)
            return round(log_mean, 12), round(sd1, 12)
        else:
            log_mean = np.log(ef) - 0.5*(sd2**2)
            return round(log_mean, 12), round(sd2, 12)


def max_min(db, mean_gen, total_gen, total_facility_considered):
    # Troy Method
    data = db
    maximum = (data.iloc[:, 1]/data.iloc[:, 0]).max()
    minimum = (data.iloc[:, 1]/data.iloc[:, 0]).min()

    return (minimum, maximum)