def main():
if len(sys.argv) != 2:
print('Python LePrEF 2015')
print('Usage: python3 %s argv[1]' % sys.argv[0])
print('argv[1]: Execution configuration file')
print('\tJson Execution configuration file must have the attributes below:')
print('\t {')
print('\t ftrainame:\t\ttrain file name,')
print('\t fvaliname:\t\tvalidation file name,')
print('\t ftestname:\t\ttest file name,')
print('\t ffiltersresults:\tfilters result file name,')
print('\t foutputname:\tAOL output file name,')
print('\t resultdir:\t\tOut directory for results,')
print('\t randomseed:\t\tRandom Seed,')
print('\t popsize:\t\tPopulation size,')
print('\t generations:\tGenerations number,')
print('\t cxprob:\t\tCrossover Probability,')
print('\t mutprob:\t\tMutation Probability,')
print('\t hoflen:\t\tHall of Fame length')
print('\t }')
return 1
#Load config file
fexecname = sys.argv[1]
exec_config = load_config_file(fexecname)
ftrainame = exec_config['ftrainame']
fvaliname = exec_config['fvaliname']
ftestname = exec_config['ftestname']
ffiltersresults = exec_config['ffiltersresults']
foutputname = exec_config['foutputname']
resultdir = exec_config['resultdir']
randomseed = int(exec_config['randomseed'])
popsize = int(exec_config['popsize'])
generations = int(exec_config['generations'])
cxprob = float(exec_config['cxprob'])
mutprob = float(exec_config['mutprob'])
hoflen = int(exec_config['hoflen'])
#end load config file
#Start up print
print('Bm25GP [%s] - %s [%s]' %
(fexecname, pwd.getpwuid(os.getuid())[4],os.getlogin()),
'\nRunning on Python',sys.version)
logger(resultdir, 'Bm25GP [%s] - %s [%s]' %
(fexecname, pwd.getpwuid(os.getuid())[4],os.getlogin()),
'\nRunning on Python',sys.version)
#Set pset and toolbox
pset = create_primitiveset()
toolbox = create_toolbox(pset)
set_gpoperator(toolbox, pset)
#set_evaluate(toolbox, evalQuery, queries, index, aolstats, results)
#results
print('Reading results from "' + ffiltersresults + '"...')
logger(resultdir, 'Reading results from "' + ffiltersresults + '"...')
filtersresults_data = load_results(ffiltersresults)['results']
print('Processing results...')
logger(resultdir, 'Processing results...')
results = create_result_dict(filtersresults_data)
print('Results ready!')
logger(resultdir, 'Results ready!')
#Get train data
print('Reading train queries from "' + ftrainame + '"...')
logger(resultdir, 'Reading train queries from "' + ftrainame + '"...')
queries = lepref_util.carregar_queries(ftrainame)[0]
#lepref_util.configurar_idcg_medio(queries, topN = MAXEVAL)
print(len(queries), 'queries read!')
logger(resultdir, len(queries), 'queries read!')
#Create Bm25 Index
print('Creating index from queries...')
logger(resultdir, 'Creating index from queries...')
index = bm25.Bm25Index()
index.generate_from_queries(queries)
print('Index created with', len(index), 'terms!')
logger(resultdir, 'Index created with', len(index), 'terms!')
#Set evaluate function and data
print('Reading AOL output file from "' + foutputname + '"...')
logger(resultdir, 'Reading AOL output file from "' + foutputname + '"...')
aoldata, featurelist = aol_parser.output_read(foutputname)
print('Precomputing AOL statistics...')
logger(resultdir, 'Precomputing AOL statistics...')
aolstats = process_aol_stats(aoldata)
print(len(aolstats['stats']), 'AOL statistics created!')
logger(resultdir, len(aolstats['stats']), 'AOL statistics created!')
set_evaluate(toolbox, evalQuery, queries, index, aolstats, results)
print('All data is ready!')
logger(resultdir, 'All data is ready!')
if os.path.exists(resultdir + os.sep + 'checkpoint.data'):
#Load Checkpoint
print('Checkpoint found! Loading...')
logger(resultdir, 'Checkpoint found! Loading...')
pop, igen, stats, hof, logbook, randomstate = load_checkpoint(resultdir)
random.setstate(randomstate)
else:
igen = 0
logbook = None
pop, hof, stats = prepare_gp(toolbox, randomseed, popsize, hoflen)
starttime = datetime.now()
if igen == 0:
print('Starting new evolution at ', starttime,'!', sep = '')
logger(resultdir, 'Starting new evolution at ', starttime,'!', sep = '')
else:
print('Resuming evolution in generation ', igen, ' at ', starttime ,'!', sep = '')
logger(resultdir, 'Resuming evolution in generation ', igen, ' at ', starttime ,'!', sep = '')
# eaCheckpoint(pop, cxprob, MutProb, igen, generations, stats, halloffame=hof,
# logbook=logbook, cpfile_location = dirresult)
eaCheckpoint(pop, toolbox, cxprob, mutprob, igen, generations, resultdir, stats, halloffame=hof,
logbook=logbook)
endtime = datetime.now()
print('Evolution training finishes at ', endtime,'!', sep = '')
print('Total time: ', endtime - starttime)
logger(resultdir, 'Evolution training finishes at ', endtime,'!', sep = '')
logger(resultdir, 'Total time: ', endtime - starttime)
#validating
#Get validation data
print('Prepair for validation and test!')
logger(resultdir, 'Prepair for validation and test!')
queries = lepref_util.carregar_queries(fvaliname)[0]
#Set validation data to toolbox.evaluate
set_evaluate(toolbox, evalQuery, queries, index, aolstats, results)
#run validation
global tobedone
tobedone = len(hof)
valiresults = [evaluation[0] for evaluation in map(toolbox.evaluate, hof)]
#output validation
print(':validation:')
logger(resultdir, ':validation:')
for vali in valiresults:
print(vali)
logger(resultdir, vali)
bestvali = max(zip(hof, valiresults), key = lambda z : z[1])
print(':bestvalidation:')
print(bestvali[1])
logger(resultdir, ':bestvalidation:')
logger(resultdir, bestvali[1])
#Testing
#Get test data
queries = lepref_util.carregar_queries(ftestname)[0]
#Set test data to toolbox.evaluate
set_evaluate(toolbox, evalQuery, queries, index, aolstats, results)
#run test
tobedone = 1
resulttest = toolbox.evaluate(bestvali[0])
#output test
print(':test:')
print(resulttest[0])
logger(resultdir, ':test:')
logger(resultdir, resulttest[0])
def main():
if len(sys.argv) != 5:
print('Python 2015')
print('Uso: python3 bm25.py argv[1]')
print('argv[1]: arquivo da base de dados')
print('argv[2]: nome do arquivo onde serão salvos os dados')
print('argv[3]: nome do diretório onde serão salvos os arquivos de resultados')
print('argv[4]: nome do arquivo de range')
# print('argv[x]:')
sys.exit(-1)
else:
dbname = sys.argv[1]
fdataname = sys.argv[2]
dirresultname = sys.argv[3]
rangefname = sys.argv[4]
try:
with open(rangefname, 'r') as rangefile:
rangestring=rangefile.read()
except Exception as error:
raise error
rangedict = json.loads(rangestring)
qiini = rangedict['qiini']
fiini = rangedict['fiini']
qiend = rangedict['qiend']
fiend = rangedict['fiend']
global EXECSTHRESHOLD
try:
EXECSTHRESHOLD = rangedict['bkplen']
except KeyError:
print('Using bkplen default =', EXECSTHRESHOLD)
global NOSTOPWORDS
try:
NOSTOPWORDS = rangedict['nostopwords']
except KeyError:
print('Using nostopwords default =', NOSTOPWORDS)
#Filters
global filters
filters = {
1 : {'name': 'default(or)', 'query': True, 'posproc': False},
2 : {'name': 'and', 'query': True, 'posproc': True},
3 : {'name': 'filter', 'query': False, 'posproc': True}
}
if os.path.exists(fdataname):
print("Carregando dados do arquivo ", fdataname, '...',sep = '')
queries, index = load_data(fdataname)
else:
#Obter dados de treino
print("Lendo queries...")
queries, lista_invertida = lepref_util.carregar_queries(sys.argv[1])
print('Queries lidas:', len(queries))
index = Bm25Index()
print("Adicionando queries ao índice")
index.generate_from_queries(queries)
save_data(queries, index, fdataname)
lepref_util.configurar_idcg_medio(queries, topN = TOPN)
###Evaluations
##Simple
# print('MeanNDCG: ', evaluate(queries, index, topN = TOPN))
##Filters permutations
doit = False
if os.path.exists(dirresultname):
print("Diretório ", dirresultname, " de resultado encontrado!", sep = '')
while (True):
inputstring = input('Deseja calcular novamente?(N,y): ')
if inputstring in ['Y','y']:
doit = True
break
elif inputstring in ['N','n', '']:
print('Saindo...')
sys.exit()
else:
print('Opção <', inputstring, '> Invalida!', sep='')
else:
doit = True
# print_idf(index)
if doit:
#print(count_execs(queries, qiini, fiini, qiend, fiend))
#results = evaluate_filters(queries, index, topN = TOPN, qiini = 1, fiini = 0, qiend = 2, fiend = 8)
evaluate_filters(queries, index, dirresultname, topN = TOPN, qiini = qiini, fiini = fiini, qiend = qiend, fiend = fiend)
#evaluate_filters(queries, index, dirresultname, topN = TOPN, qiini = 0, fiini = 0, qiend = 3, fiend = None)
#save_results(results, dirresultname)
print('Docs lidos:', len(index.doc))
print('AVGDL:', index.avgdl)
print('Vocabulary Length:', len(index))
