results, dummyresults, settings = si.getOrganizedResults()
values = []
# for first N result used as training
for result in results[:maxtrainers]:
val = result.get(fromfield, None)
if val != None:
values.append(val)
examples = [ex.strip() for ex in examples.split(",")]
if badexamples == None:
badexamples = []
else:
badexamples = [ex.strip() for ex in badexamples.split(",")]
try:
regexes, extractions = ifl.learn(values, examples, badexamples)
except Exception, e:
# log error
stack = traceback.format_exc()
logger.error("Exception encountered: %s" % e)
logger.info("Traceback: %s" % stack)
# just tell user that we couldn't extract anything
regexes = ""
if len(regexes) == 0:
si.generateErrorResults(
'Unable to learn any extractions. Provide different examples, counterexamples, or searchresults'
)
exit(0)
rex = regexes[0]
if args['edited']:
regex = args['regex']
# args['examples'] = []
args['counterexamples'] = []
else:
try:
# run through user examples, to see if the user specified multiple values per example
seenMultipleValues = False
for exampleSet in examples:
vals = splitExampleValues(exampleSet)
if len(vals) > 1:
seenMultipleValues = True
break
# if no multiple values, use standard ifx learning
if not seenMultipleValues:
regexes, extractions = ifl.learn(events, examples, args['counterexamples'])
if len(regexes) > 0:
regex = regexes[0]
else: # if multiple values, use in multivalue ifx
counterExamples = args['counterexamples']
# !! hack
counterExamples = { 'field0': counterExamples }
sourceField = "_raw"
markedEvents = {}
for i, event in enumerate(events):
markedEvent = {}
raw = event
markedEvent["_event"] = { sourceField : raw }
for exampleSet in examples:
# !! hack
pos = 0
results,dummyresults,settings = si.getOrganizedResults()
values = []
# for first N result used as training
for result in results[:maxtrainers]:
val = result.get(fromfield, None)
if val != None:
values.append(val)
examples = [ex.strip() for ex in examples.split(",")]
if badexamples == None:
badexamples = []
else:
badexamples = [ex.strip() for ex in badexamples.split(",")]
try:
regexes, extractions = ifl.learn(values, examples, badexamples)
except Exception, e:
# log error
logger.error("%s" % e)
logger.info("Traceback: %s" % stack)
# just tell user that we couldn't extract anything
regexes = ""
if len(regexes) == 0:
si.generateErrorResults('Unable to learn any extractions. Provide different examples, counterexamples, or searchresults')
exit(0)
rex = regexes[0]
rex = rex.replace("?P<FIELDNAME>", "?P<%s>" % keywords[0])
si.addInfoMessage(messages, 'Successfully learned regex. Consider using: | rex "%s"' % rex.replace('"', '\\"'))
if args['edited']:
regex = args['regex']
# args['examples'] = []
args['counterexamples'] = []
else:
try:
# run through user examples, to see if the user specified multiple values per example
seenMultipleValues = False
for exampleSet in examples:
vals = splitExampleValues(exampleSet)
if len(vals) > 1:
seenMultipleValues = True
break
# if no multiple values, use standard ifx learning
if not seenMultipleValues:
regexes, extractions = ifl.learn(events, examples, args['counterexamples'])
if len(regexes) > 0:
regex = regexes[0]
else: # if multiple values, use in multivalue ifx
counterExamples = args['counterexamples']
# !! hack
counterExamples = { 'field0': counterExamples }
sourceField = "_raw"
markedEvents = {}
for i, event in enumerate(events):
markedEvent = {}
raw = event
markedEvent["_event"] = { sourceField : raw }
for exampleSet in examples:
# !! hack
pos = 0