1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332
|
# $Id$
#
# Copyright (C) 2002-2008 greg Landrum and Rational Discovery LLC
#
# @@ All Rights Reserved @@
# This file is part of the RDKit.
# The contents are covered by the terms of the BSD license
# which is included in the file license.txt, found at the root
# of the RDKit source tree.
#
""" command line utility to report on the contributions of descriptors to
tree-based composite models
Usage: AnalyzeComposite [optional args] <models>
<models>: file name(s) of pickled composite model(s)
(this is the name of the db table if using a database)
Optional Arguments:
-n number: the number of levels of each model to consider
-d dbname: the database from which to read the models
-N Note: the note string to search for to pull models from the database
-v: be verbose whilst screening
"""
import sys
import numpy
from rdkit.Dbase.DbConnection import DbConnect
from rdkit.ML import ScreenComposite
from rdkit.ML.Data import Stats
from rdkit.ML.DecTree import TreeUtils, Tree
import pickle
__VERSION_STRING = "2.2.0"
def ProcessIt(composites, nToConsider=3, verbose=0):
composite = composites[0]
nComposites = len(composites)
ns = composite.GetDescriptorNames()
# nDesc = len(ns)-2
if len(ns) > 2:
globalRes = {}
nDone = 1
descNames = {}
for composite in composites:
if verbose > 0:
print('#------------------------------------')
print('Doing: ', nDone)
nModels = len(composite)
nDone += 1
res = {}
for i in range(len(composite)):
model = composite.GetModel(i)
if isinstance(model, Tree.TreeNode):
levels = TreeUtils.CollectLabelLevels(model, {}, 0, nToConsider)
TreeUtils.CollectDescriptorNames(model, descNames, 0, nToConsider)
for descId in levels.keys():
v = res.get(descId, numpy.zeros(nToConsider, numpy.float))
v[levels[descId]] += 1. / nModels
res[descId] = v
for k in res:
v = globalRes.get(k, numpy.zeros(nToConsider, numpy.float))
v += res[k] / nComposites
globalRes[k] = v
if verbose > 0:
for k in res.keys():
name = descNames[k]
strRes = ', '.join(['%4.2f' % x for x in res[k]])
print('%s,%s,%5.4f' % (name, strRes, sum(res[k])))
print()
if verbose >= 0:
print('# Average Descriptor Positions')
retVal = []
for k in globalRes:
name = descNames[k]
if verbose >= 0:
strRes = ', '.join(['%4.2f' % x for x in globalRes[k]])
print('%s,%s,%5.4f' % (name, strRes, sum(globalRes[k])))
tmp = [name]
tmp.extend(globalRes[k])
tmp.append(sum(globalRes[k]))
retVal.append(tmp)
if verbose >= 0:
print()
else:
retVal = []
return retVal
def ErrorStats(conn, where, enrich=1):
fields = ('overall_error,holdout_error,overall_result_matrix,' +
'holdout_result_matrix,overall_correct_conf,overall_incorrect_conf,' +
'holdout_correct_conf,holdout_incorrect_conf')
try:
data = conn.GetData(fields=fields, where=where)
except Exception:
import traceback
traceback.print_exc()
return None
nPts = len(data)
if not nPts:
sys.stderr.write('no runs found\n')
return None
overall = numpy.zeros(nPts, numpy.float)
overallEnrich = numpy.zeros(nPts, numpy.float)
oCorConf = 0.0
oInCorConf = 0.0
holdout = numpy.zeros(nPts, numpy.float)
holdoutEnrich = numpy.zeros(nPts, numpy.float)
hCorConf = 0.0
hInCorConf = 0.0
overallMatrix = None
holdoutMatrix = None
for i in range(nPts):
if data[i][0] is not None:
overall[i] = data[i][0]
oCorConf += data[i][4]
oInCorConf += data[i][5]
if data[i][1] is not None:
holdout[i] = data[i][1]
haveHoldout = 1
else:
haveHoldout = 0
tmpOverall = 1. * eval(data[i][2])
if enrich >= 0:
overallEnrich[i] = ScreenComposite.CalcEnrichment(tmpOverall, tgt=enrich)
if haveHoldout:
tmpHoldout = 1. * eval(data[i][3])
if enrich >= 0:
holdoutEnrich[i] = ScreenComposite.CalcEnrichment(tmpHoldout, tgt=enrich)
if overallMatrix is None:
if data[i][2] is not None:
overallMatrix = tmpOverall
if haveHoldout and data[i][3] is not None:
holdoutMatrix = tmpHoldout
else:
overallMatrix += tmpOverall
if haveHoldout:
holdoutMatrix += tmpHoldout
if haveHoldout:
hCorConf += data[i][6]
hInCorConf += data[i][7]
avgOverall = sum(overall) / nPts
oCorConf /= nPts
oInCorConf /= nPts
overallMatrix /= nPts
oSort = numpy.argsort(overall)
oMin = overall[oSort[0]]
overall -= avgOverall
devOverall = numpy.sqrt(sum(overall**2) / (nPts - 1))
res = {}
res['oAvg'] = 100 * avgOverall
res['oDev'] = 100 * devOverall
res['oCorrectConf'] = 100 * oCorConf
res['oIncorrectConf'] = 100 * oInCorConf
res['oResultMat'] = overallMatrix
res['oBestIdx'] = oSort[0]
res['oBestErr'] = 100 * oMin
if enrich >= 0:
mean, dev = Stats.MeanAndDev(overallEnrich)
res['oAvgEnrich'] = mean
res['oDevEnrich'] = dev
if haveHoldout:
avgHoldout = sum(holdout) / nPts
hCorConf /= nPts
hInCorConf /= nPts
holdoutMatrix /= nPts
hSort = numpy.argsort(holdout)
hMin = holdout[hSort[0]]
holdout -= avgHoldout
devHoldout = numpy.sqrt(sum(holdout**2) / (nPts - 1))
res['hAvg'] = 100 * avgHoldout
res['hDev'] = 100 * devHoldout
res['hCorrectConf'] = 100 * hCorConf
res['hIncorrectConf'] = 100 * hInCorConf
res['hResultMat'] = holdoutMatrix
res['hBestIdx'] = hSort[0]
res['hBestErr'] = 100 * hMin
if enrich >= 0:
mean, dev = Stats.MeanAndDev(holdoutEnrich)
res['hAvgEnrich'] = mean
res['hDevEnrich'] = dev
return res
def ShowStats(statD, enrich=1):
statD = statD.copy()
statD['oBestIdx'] = statD['oBestIdx'] + 1
txt = """
# Error Statistics:
\tOverall: %(oAvg)6.3f%% (%(oDev)6.3f) %(oCorrectConf)4.1f/%(oIncorrectConf)4.1f
\t\tBest: %(oBestIdx)d %(oBestErr)6.3f%%""" % (statD)
if 'hAvg' in statD:
statD['hBestIdx'] = statD['hBestIdx'] + 1
txt += """
\tHoldout: %(hAvg)6.3f%% (%(hDev)6.3f) %(hCorrectConf)4.1f/%(hIncorrectConf)4.1f
\t\tBest: %(hBestIdx)d %(hBestErr)6.3f%%
""" % (statD)
print(txt)
print()
print('# Results matrices:')
print('\tOverall:')
tmp = numpy.transpose(statD['oResultMat'])
colCounts = sum(tmp)
rowCounts = sum(tmp, 1)
for i in range(len(tmp)):
if rowCounts[i] == 0:
rowCounts[i] = 1
row = tmp[i]
print('\t\t', end='')
for j in range(len(row)):
print('% 6.2f' % row[j], end='')
print('\t| % 4.2f' % (100. * tmp[i, i] / rowCounts[i]))
print('\t\t', end='')
for i in range(len(tmp)):
print('------', end='')
print()
print('\t\t', end='')
for i in range(len(tmp)):
if colCounts[i] == 0:
colCounts[i] = 1
print('% 6.2f' % (100. * tmp[i, i] / colCounts[i]), end='')
print()
if enrich > -1 and 'oAvgEnrich' in statD:
print('\t\tEnrich(%d): %.3f (%.3f)' % (enrich, statD['oAvgEnrich'], statD['oDevEnrich']))
if 'hResultMat' in statD:
print('\tHoldout:')
tmp = numpy.transpose(statD['hResultMat'])
colCounts = sum(tmp)
rowCounts = sum(tmp, 1)
for i in range(len(tmp)):
if rowCounts[i] == 0:
rowCounts[i] = 1
row = tmp[i]
print('\t\t', end='')
for j in range(len(row)):
print('% 6.2f' % row[j], end='')
print('\t| % 4.2f' % (100. * tmp[i, i] / rowCounts[i]))
print('\t\t', end='')
for i in range(len(tmp)):
print('------', end='')
print()
print('\t\t', end='')
for i in range(len(tmp)):
if colCounts[i] == 0:
colCounts[i] = 1
print('% 6.2f' % (100. * tmp[i, i] / colCounts[i]), end='')
print()
if enrich > -1 and 'hAvgEnrich' in statD:
print('\t\tEnrich(%d): %.3f (%.3f)' % (enrich, statD['hAvgEnrich'], statD['hDevEnrich']))
return
def Usage():
print(__doc__)
sys.exit(-1)
if __name__ == "__main__":
import getopt
try:
args, extras = getopt.getopt(sys.argv[1:], 'n:d:N:vX', ('skip',
'enrich=', ))
except Exception:
Usage()
count = 3
db = None
note = ''
verbose = 0
skip = 0
enrich = 1
for arg, val in args:
if arg == '-n':
count = int(val) + 1
elif arg == '-d':
db = val
elif arg == '-N':
note = val
elif arg == '-v':
verbose = 1
elif arg == '--skip':
skip = 1
elif arg == '--enrich':
enrich = int(val)
composites = []
if db is None:
for arg in extras:
composite = pickle.load(open(arg, 'rb'))
composites.append(composite)
else:
tbl = extras[0]
conn = DbConnect(db, tbl)
if note:
where = "where note='%s'" % (note)
else:
where = ''
if not skip:
pkls = conn.GetData(fields='model', where=where)
composites = []
for pkl in pkls:
pkl = str(pkl[0])
comp = pickle.loads(pkl)
composites.append(comp)
if len(composites):
ProcessIt(composites, count, verbose=verbose)
elif not skip:
print('ERROR: no composite models found')
sys.exit(-1)
if db:
res = ErrorStats(conn, where, enrich=enrich)
if res:
ShowStats(res)
|