10_pine_optimal_branching

Pine - Optimal Simulation Configuration with Branching Simulation pine_weights i_53

Install the package from pip (skip this step if the package is already installed)

! pip install ligning

Import Code for Lignin Optimization

# set ligning path (optional if installed via pip)
import sys, os
project_path = os.path.abspath(os.path.join(os.getcwd(), '..\..\..'))
sys.path.insert(0, project_path)

import ligning.optimization as opt
import ligning.characterization as ch
import ligning.utils as ut

Set the destination of simulation outputs. ResultsName sets the first level of storage, library_name specifies the folder within ‘demo_results’ and the trial_index is optional, specifying the test number. This allows data from multiple simulations to be stored, rather than overwriting past results, if no trial_index is given

ResultsName='demo_results'
library_name = 'pine_branch'
trial_index=0

Set distribution of linkages This sets the expected frequency of each type of linkage between monomers. Setting to 0 will disable that linkage for the simulation. [‘4-O-5’, ‘alpha-O-4’, ‘beta-O-4’, ‘5-5’, ‘beta-5’, ‘beta-beta’, ‘beta-1’]

linkage_distribution_input = [0, 0, 66, 0.05, 18, 16, 0]

Set monomer distributions This sets the expected frequency of each type of monomers. Setting to 0 will disable that monomer for the simulation. [‘H’, ‘G’, ‘S’]

monomer_distribution_input = [0, 100, 0]

Setting for additional metrics in simulation output Verbose reports monomer additions to polymers. Additional_metrics reports population data such as branching coefficient in the simulation output.

verbose = False
additional_metrics = True

Set branching propensity of polymers Likelihood of a monomer linking to 3 or more neighboring monomers.

branching_propensity = 0.1

Set population metrics Determines the target monomer size for the simulation. [‘number average molecular weight’, ‘weight average molecular weight’]

population_metrics = None

Additional setting for expectations the size of the polymer population, including the expected average size, maximum polymer size, and distribution of sizes of polymers.

expected_size = 100
max_size = 250

# size distribution scaling factor
distribution_scaling = 0.1

# size in MW
size_in_MW = False

Metropolis Temperatures The temperature for the Metropolis Monte Carlo in monomer addition (Tmetro) and polymer addition (Tmetro_out).

Tmetro = 10
Tmetro_out = 10

Simulation settings Maximum iterations for monomer addition to each polymer, polymer addition to the population, and ring addition to the population. N_population is the number of polymers in the population. Seed_init sets the random seed for the simulation.

i_max = 10000
i_max_out = 1000
i_max_ring = 500

n_population = 100

seed_init = 3

(Optional) Set metric weights 13 metrics are available The first seven are linkage distributions: ‘4-O-5’, ‘alpha-O-4’, ‘beta-O-4’, ‘5-5’, ‘beta-5’, ‘beta-beta’, ‘beta-1’ The next three are monomer distributions: ‘H’, ‘G’, ‘S’ Then the branching coefficient Two optional, representing ‘number average molecular weight’ and ‘weight average molecular weight’, are not included for this simulation

metrics_weights = [1, 1, 1, 1, 1, 5, 1, 1, 1, 1, 1]

Create and run the simulation

sim = opt.Simulation(linkage_distribution_input=linkage_distribution_input,
                     monomer_distribution_input=monomer_distribution_input,
                     expected_size=expected_size,
                     max_size=max_size,
                     distribution_scaling=distribution_scaling,
                     Tmetro=Tmetro,
                     Tmetro_out=Tmetro_out,
                     seed_init=seed_init,
                     ResultsName=ResultsName,
                     library_name=library_name,
                     trial_index=trial_index,
                     n_population=n_population,
                     i_max=i_max,
                     i_max_out=i_max_out,
                     i_max_ring=i_max_ring,
                     additional_metrics=additional_metrics,
                     population_metrics=population_metrics,
                     size_in_MW=size_in_MW,
                     branching_propensity=branching_propensity,
                     metrics_weights = metrics_weights,
                     verbose=verbose)

sim.run()

Starting a new trial, No.0:

C:\Users\jake_\Documents\GitHub\LigninGraphs\ligning\polymer.py:662: UserWarning: Input linkage type is not supported
  warnings.warn("Input linkage type is not supported")

Runtime for creating all polymers : 110.77 minutes

Runtime for adding the rings : 6.56 minutes

C:\Users\jake_\Documents\GitHub\LigninGraphs\ligning\characterization.py:616: FutureWarning: Dropping of nuisance columns in DataFrame reductions (with 'numeric_only=None') is deprecated; in a future version this will raise TypeError.  Select only valid columns before calling the reduction.
  numerical_metrics = list(population_data.mean().index)
C:\Users\jake_\Documents\GitHub\LigninGraphs\ligning\characterization.py:617: FutureWarning: Dropping of nuisance columns in DataFrame reductions (with 'numeric_only=None') is deprecated; in a future version this will raise TypeError.  Select only valid columns before calling the reduction.
  population_mean = np.array([population_data.mean()])
C:\Users\jake_\Documents\GitHub\LigninGraphs\ligning\characterization.py:618: FutureWarning: Dropping of nuisance columns in DataFrame reductions (with 'numeric_only=None') is deprecated; in a future version this will raise TypeError.  Select only valid columns before calling the reduction.
  population_std = np.array([population_data.std()])

Target values:
{   '4-O-5': 0.0,
    '5-5': 0.0004997501249375313,
    'G': 1.0,
    'H': 0.0,
    'S': 0.0,
    'alpha-O-4': 0.0,
    'beta-1': 0.0,
    'beta-5': 0.17991004497751126,
    'beta-O-4': 0.6596701649175413,
    'beta-beta': 0.15992003998001,
    'branching_coeff': 1.0}
Optimal simulated values:
{   '4-O-5': 0.0,
    '5-5': 0.0036851803106080546,
    'G': 1.0,
    'H': 0.0,
    'MW': 13653.292960396044,
    'MW_weighted': 18036.439767927317,
    'S': 0.0,
    'alpha-O-4': 0.0,
    'beta-1': 0.0,
    'beta-5': 0.20229007633587787,
    'beta-O-4': 0.7811266122663859,
    'beta-beta': 0.012898131087128192,
    'branching_coeff': 0.005735888410898188,
    'monomer_count': 75.95049504950495}

Acceptance Rates
Monomer Acceptance: 0.026793505608387422
Polymer Acceptance: 0.9351851851851852
Ring Acceptance: 0.0005928736586233473
Runtime for analyzing the results : 0.35 minutes

Process population

P_population = sim.P_population
population = ch.Population(P_population, name=library_name, ResultsName=ResultsName, TrialIndex=str(trial_index))
population.analyze()

C:\Users\jake_\Documents\GitHub\LigninGraphs\ligning\characterization.py:616: FutureWarning: Dropping of nuisance columns in DataFrame reductions (with 'numeric_only=None') is deprecated; in a future version this will raise TypeError.  Select only valid columns before calling the reduction.
  numerical_metrics = list(population_data.mean().index)
C:\Users\jake_\Documents\GitHub\LigninGraphs\ligning\characterization.py:617: FutureWarning: Dropping of nuisance columns in DataFrame reductions (with 'numeric_only=None') is deprecated; in a future version this will raise TypeError.  Select only valid columns before calling the reduction.
  population_mean = np.array([population_data.mean()])
C:\Users\jake_\Documents\GitHub\LigninGraphs\ligning\characterization.py:618: FutureWarning: Dropping of nuisance columns in DataFrame reductions (with 'numeric_only=None') is deprecated; in a future version this will raise TypeError.  Select only valid columns before calling the reduction.
  population_std = np.array([population_data.std()])