def plot_convergence(rxn,
dbse,
proj,
loadfrompickle,
std_mtd,
std_bsse,
std_bas_list,
std_curvestyle,
mtd_list,
bsse,
bas_list,
curvestyle,
failoninc=True,
assigned_title='default',
xlabel='default',
ylabel='default',
legend=True,
markersize='default',
linewidth='default',
standalone=True,
plotpath='default',
filename='default',
filetype='default'):
"""Prepares convergence plots for all members of array *rxn* of
reactions from database *dbse* in project(s) *proj* of modelchemistry
defined by all combinations of the method list, *mtd_list*,
counterpoise scheme, *bsse*, and basis set list, *bas_list*, arrays.
Convergence curves are plotted relative to reference value, with
style defined in array *curvestyle* for each of the curves to be
created. Curve styles are three-character strings giving marker
style, curve color, and line style.
Plot title *title* will default to a standard string identifier
based on *rxn* and *dbse*, unless otherwise specified. Axis labels
*xlabel* and *ylabel* may be specified, but have default values
"Basis Set" and "Interaction Energy" otherwise. Ticks on the x axis
will be basis sets defined in *bas_list*.
Inclusion of legend may be turned on or off by the truth value
of *legend*. Marker and line size may be changed by adjusting
*markersize* and *linewidth* arguments, both accepting float values.
Default linewidth is 1, default marker size is 5.5.
Plots are opened if *standalone* is true, in .pdf format by
default. File type can be changed by *filetype*, and will be saved in
location designated by *plotpath* as *dbse*_*rxn*_convplot.*filetype*
if *filename* is left as 'default'.
"""
# load data from corresponding function arguments
asdf = qcdb.Database([dbse], loadfrompickle=loadfrompickle)
for pr in proj:
asdf.load_qcdata_byproject(pr)
# Build 2D modelchems array for novel methods
mcs = []
for mtd in mtd_list:
for b in bsse:
temp_mcs_row = []
for bas in bas_list:
temp_bas = []
temp_bas.append(mtd + '-' + b + '-' + bas)
temp_mcs_row.append(temp_bas[0])
mcs.append(temp_mcs_row)
# Build 1D standard modelchem array
std_mcs = []
for std_bas in std_bas_list:
std_mcs.append(std_mtd + '-' + std_bsse + '-' + std_bas)
# Get benchmark values for all rxns/modelchems in database
bench = asdf.benchmark # Where to look for the benchmark values
temp_ref_array = []
for lmc, lbm, orxn in asdf.get_reactions(modelchem=bench):
temp_ref_array.append(
orxn.data[bench].value) # Get all benchmark values for database
# TODO: Form rxn array based on more intuitive string argument?
# Iterate over array *rxn* and create plot for each one
for r in rxn:
# Assign plot title
if assigned_title == 'default':
working_title = dbse + '-' + str(r) + ' ConvPlot'
else:
working_title = assigned_title
# Get benchmark value for r
ref = temp_ref_array[r - 1]
# Build rank 1 *std_dbdata* array containing the data for the standard convergence curve
std_dbdata = []
for std_mc in std_mcs:
all_std_mc_data = []
if std_mc in asdf.mcs.keys():
for lmc, lbm, orxn in asdf.get_reactions(modelchem=std_mc,
failoninc=failoninc):
try:
all_std_mc_data.append(orxn.data[std_mc].value)
except KeyError, e:
if failoninc:
raise e
else:
all_std_mc_data.append(np.NaN)
else:
print "Standard model chemistry not in mcs.keys(). Try again."
std_dbdata.append(all_std_mc_data[r - 1])
std_curve_label = std_mtd + '-' + std_bsse
std_x_vals = np.arange(1, len(std_dbdata) + 1)[:]
# Build rank 2 *dbdata* array containing all data for rxn *r* to be plotted
# Build rank 1 *curve_labels* array for use in legend
# Build rank 1 *x_vals* array against which to plot rows of *dbdata*
dbdata = []
x_vals = []
curve_labels = []
for row in mcs: # Each row contains different combo of *mcs* and *bsse*
temp_dbdata_row = []
for mc in row: # Iterate over different values of *bas* for same *mcs* & *bsse*
all_dbse_mc_data = []
if mc in asdf.mcs.keys(
): # If the modelchem exists at all, regardless of some missing data
for lmc, lbm, orxn in asdf.get_reactions(
modelchem=mc,
failoninc=failoninc): # Gets rxndata dictionary
try:
all_dbse_mc_data.append(
orxn.data[mc].value
) # Appends all *dbse*/*mc* rxn datum to temporary array
except KeyError, e: # Protects against missing data
if failoninc:
raise e # Raises offending mc label
else:
all_dbse_mc_data.append(
np.NaN
) # Appends None to dbdata array, so nothing gets plotted for no data
else:
for i in range(
24
): # Need to append 24 NaN's to simulate totally empty array
all_dbse_mc_data.append(
np.NaN
) # Append None to dbdata array for an entire mc if it doesn't exist
temp_dbdata_row.append(
all_dbse_mc_data[r - 1]
) # Appends rxn *r*/*mc* datum to array for later append to *dbdata*
# Split modelchem *mc* name *mtd*-*bsse*-*bas* by '-' delimiter, save in temp array
temp_curve_label = mc.split("-")
# Concatenate *mtd* + '-' + *bsse* to form unique curve_label, add to *curve_labels* array
curve_labels.append(temp_curve_label[0] + '-' +
temp_curve_label[1])
dbdata.append(temp_dbdata_row) # Add row to *dbdata* array
dbse + ' @ ' + mtd,
exitonfail=False)
print('')
err = 'rmse'
digits = 0.02
mode = '-unCP-atz'
active_methods = [
'PBE', 'B3LYP', 'VV10', 'LCVV10', 'WB97XD', 'WB97XV', 'M062X', 'M11L',
'M11'
]
print("""\n<<< Mardirossian:2014:9904 PCCP 2014, 16, 9904 >>>\n""")
asdf = qcdb.Database('s22')
asdf.load_qcdata_byproject('dhdft')
testem(odb=asdf, ss='default', bm='S22B', ans=answers['s22'])
asdf = qcdb.Database('hbc6')
asdf.load_qcdata_byproject('dhdft')
testem(odb=asdf, ss='default', bm='HBC6A', ans=answers['hbc'])
asdf = qcdb.Database('nbc10')
asdf.load_qcdata_byproject('dhdft')
asdf.add_Subset_union('a1', ['bzbz_pd34', 'bzbz_t', 'bzbz_s'])
asdf.add_Subset_union('a2', ['meme', 'bzme'])
asdf.add_Subset_union('a3', ['pypy_s2', 'pypy_t3'])
testem(odb=asdf, ss='a1', bm='NBC10A', ans=answers['nbca1'])
testem(odb=asdf, ss='a2', bm='NBC10A', ans=answers['nbca2'])
testem(odb=asdf, ss='a3', bm='NBC10A', ans=answers['nbca3'])
from __future__ import print_function
import sys
sys.path.append('/Users/loriab/linux/qcdb')
sys.path.append('/Users/loriab/linux/qcdb/databases')
import qcdb
a24 = qcdb.Database('A24')
a24.load_dilabio()
a24.load_f12dilabio()
import matplotlib.pyplot as plt
import numpy as np
mcs1 = [
'CCSDT-CP-adz', 'CCSDT-CP-atz', 'CCSDT-CP-aqz', 'CCSDT-CP-a5z',
'CCSDT-CP-a6z'
]
mcs2 = [
'CCSDTAF12-CP-adz', 'CCSDTAF12-CP-atz', 'CCSDTAF12-CP-aqz',
'CCSDTAF12-CP-a5z'
]
mcs3 = [
'CCSDTBF12-CP-adz', 'CCSDTBF12-CP-atz', 'CCSDTBF12-CP-aqz',
'CCSDTBF12-CP-a5z'
]
mcs4 = ['CCSDTCF12-CP-adz', 'CCSDTCF12-CP-atz', 'CCSDTCF12-CP-aqz']
mcs2f = ['CCSDTAF12-CP-dzf12', 'CCSDTAF12-CP-tzf12', 'CCSDTAF12-CP-qzf12']
mcs3f = ['CCSDTBF12-CP-dzf12', 'CCSDTBF12-CP-tzf12', 'CCSDTBF12-CP-qzf12']
mcs4f = ['CCSDTCF12-CP-dzf12', 'CCSDTCF12-CP-tzf12', 'CCSDTCF12-CP-qzf12']
mcs5 = ['CCSDT-CP-adtz', 'CCSDT-CP-atqz', 'CCSDT-CP-aq5z', 'CCSDT-CP-a56z']
mcs6 = ['CCSDTAF12-CP-adtz', 'CCSDTAF12-CP-atqz', 'CCSDTAF12-CP-aq5z']
import collections
sys.path.append('C:\Users\Owner\Documents\GitHub\qcdb')
sys.path.append('C:\Users\Owner\Documents\GitHub\qcdb\databases')
import qcdb
from qcdb.psivarrosetta import useme2psivar
from qcdb.modelchems import Method, BasisSet, methods, bases
import pandas as pd
import numpy as np
pd.set_option('display.max_columns', 10)
pd.set_option('display.width', 200)
# <<< read usemefiles and convert to giant DataFrame >>>
dbse = 'A24'
dbobj = qcdb.Database(dbse)
path = r"""C:\Users\Owner\Documents\f12dilabiousemefiles\usemefiles"""
h2kc = qcdb.psi_hartree2kcalmol
rawdata = collections.defaultdict(dict)
for useme in glob.glob('%s/%s*useme*' % (path, dbse)):
spl = os.path.basename(useme).split('.')
#dbse = spl[0].split('-')[0]
#ocalc = spl[0].split('-')[1]
#optns = '_'.join(spl[0].split('-')[2:-1])
basis = spl[0].split('-')[-1]
piece = '.'.join(spl[1:])
print(useme, basis, piece)
tmp = pd.read_csv('%s' % (useme), index_col=0, sep='\s+', comment='#', na_values='None', names=['rxn', 'dimer', 'monoA-CP', 'monoB-CP'])
print(useme, basis, piece)
print(tmp.head(10))
#dbnet['NBC10ext'] = ['saptmisc', 'dfit']
#dbnet['ACHC'] = ['saptmisc', 'dfit']
#dbnet['UBQ'] = ['saptmisc', 'bfdbmm']
#dbnet['S22by7'] = ['saptmisc']
#dbnet['S66'] = ['saptmisc']
#dbnet['A24'] = ['saptmisc', 'dilabio']
#dbnet['JSCH'] = ['saptmisc']
#dbnet[''] = []
for db, lproj in dbnet.iteritems():
print '\n<<< %s >>>' % (db)
t0 = time.time()
if db == 'DB4':
asdf = qcdb.DB4(loadfrompickle=True, path=homewrite)
else:
asdf = qcdb.Database(db, loadfrompickle=True, path=homewrite)
dbse = asdf.dbse
t1 = time.time()
print '%-70s %8.1f' % ('database.py --> Database', t1 - t0)
print 'Benchmark', asdf.benchmark
Dbfilename = homewrite + '/' + db + '_Db.pickle'
with open(Dbfilename, 'wb') as handle:
pickle.dump(asdf, handle, pickle.HIGHEST_PROTOCOL)
t2 = time.time()
print '%-70s %8.1f' % ('* Database --> database_Db.pickle', t2 - t1)
for pj in lproj:
t3 = time.time()
asdf.load_qcdata_byproject(pj)
t4 = time.time()
