BM Tree Fitting Gets Stuck in a Loop

[sevyharris]

I'm trying to refit the Disproportionation tree using some new training reactions I calculated. I tried using Matt's tree fitting script and it appears to be stuck in a loop. I don't know if I just haven't let it run long enough (tried 24 hours) or if there's a setting I need to change so it can break free.

I haven't figured out a good way to log the error, so here's a portion of what it keeps printing to the terminal:

3 *4 H u0 p0 c0 {1,S}
4    H u0 p0 c0 {1,S}
5    H u0 p0 c0 {1,S}
6    H u0 p0 c0 {2,S}
7    H u0 p0 c0 {2,S}

ERROR:root:C2H5-2
ERROR:root:C2H5-2
multiplicity 2
1    C u0 p0 c0 {2,S} {3,S} {4,S} {5,S}
2 *1 C u1 p0 c0 {1,S} {6,S} {7,S}
3    H u0 p0 c0 {1,S}
4    H u0 p0 c0 {1,S}
5    H u0 p0 c0 {1,S}
6    H u0 p0 c0 {2,S}
7    H u0 p0 c0 {2,S}

ERROR:root:C2H6
ERROR:root:C2H6
1 *1 C u0 p0 c0 {2,S} {3,S} {4,S} {5,S}
2    C u0 p0 c0 {1,S} {6,S} {7,S} {8,S}
3    H u0 p0 c0 {1,S}
4    H u0 p0 c0 {1,S}
5 *4 H u0 p0 c0 {1,S}
6    H u0 p0 c0 {2,S}
7    H u0 p0 c0 {2,S}
8    H u0 p0 c0 {2,S}

ERROR:root:C2H4
ERROR:root:C2H4
1 *3 C u0 p0 c0 {2,D} {3,S} {4,S}
2 *2 C u0 p0 c0 {1,D} {5,S} {6,S}
3    H u0 p0 c0 {1,S}
4    H u0 p0 c0 {1,S}
5    H u0 p0 c0 {2,S}
6    H u0 p0 c0 {2,S}

ERROR:root:Clearing Regularization Dimensions and Reattempting

Any help here is appreciated, thanks!

[sevyharris]

If anyone wants to look at the exact script I'm using to refit it, here's my version of Matt's script:

import rmgpy
import numpy as np
from rmgpy.molecule.molecule import *
from rmgpy.species import *
from rmgpy.chemkin import *
from rmgpy.data.rmg import RMGDatabase
from rmgpy.data.thermo import ThermoLibrary
from rmgpy.rmg.react import react
from rmgpy.species import Species
from rmgpy.reaction import Reaction
from rmgpy.data.rmg import get_db
from rmgpy.molecule.group import Group
from rmgpy.kinetics.arrhenius import ArrheniusBM
import time

from rmgpy import settings


database = RMGDatabase()
database.load(
    path = settings['database.directory'],
    thermo_libraries = ['Klippenstein_Glarborg2016', 'BurkeH2O2', 'thermo_DFT_CCSDTF12_BAC',
        'DFT_QCI_thermo',
        'primaryThermoLibrary', 'primaryNS', 'NitrogenCurran', 'NOx2018', 'FFCM1(-)',
        'SulfurLibrary', 'SulfurGlarborgH2S','SABIC_aromatics'],
    transport_libraries = [],
    reaction_libraries = [],
    seed_mechanisms = [],#['BurkeH2O2inN2','ERC-FoundationFuelv0.9'],
    #kinetics_families = 'all',,
    kinetics_families = ['Disproportionation'],
    kinetics_depositories = ['training'],
    #frequenciesLibraries = self.statmechLibraries,
    depository = False, # Don't bother loading the depository information, as we don't use it
)
print('loaded database')

family = database.kinetics.families["Disproportionation"]

family.clean_tree()

start = time.time()
family.generate_tree(thermo_database=database.thermo,
                     nprocs=32,
                     new_fraction_threshold_to_reopt_node=0.25,
                     max_batch_size=800,
                     extension_iter_max=2,
                     extension_iter_item_cap=100)

end = time.time()
print(end-start)

It hasn't gotten past here, so no point in adding the rest yet. Also, I originally had it at npocs=1 for the 24 hour run, then tried increasing to see if that changed anything. Also, I'm on my experimental correlated_uncertainty branch of RMG-Py, so that's another possible source of error. I'll try rebuilding because I always forget to do that.