def generateRules(self):
learnedRules = fieldlearner.learn(self._sourceField, self._markedUpEvents, self._counterExamples)
# keep edited rules
editedRules = [r for r in self._rules if r.isEdited()]
# empty existing rules
self._rules = []
# for each learned rule
for lrule in learnedRules:
pattern = lrule.getPattern()
fieldinfo = lrule.getFieldValues()
# for each edited rule, if we have a match with the newly learned rule, don't add it
for r in editedRules:
if r._pattern() == pattern:
break
else: # add rule that doesn't match any edited rule
rule = Rule(pattern, fieldinfo)
self._rules.append(rule)
self._rules.extend(editedRules)
def generateRules(self):
learnedRules = fieldlearner.learn(self._sourceField, self._markedUpEvents, self._counterExamples)
# keep edited rules
editedRules = [r for r in self._rules if r.isEdited()]
# empty existing rules
self._rules = []
# for each learned rule
for lrule in learnedRules:
pattern = lrule.getPattern()
fieldinfo = lrule.getFieldValues()
# for each edited rule, if we have a match with the newly learned rule, don't add it
for r in editedRules:
if r._pattern() == pattern:
break
else: # add rule that doesn't match any edited rule
rule = Rule(pattern, fieldinfo)
self._rules.append(rule)
self._rules.extend(editedRules)
pos = 0
vals = splitExampleValues(exampleSet)
# PREVENT ABUSE
if len(vals) > MAX_FIELDS:
addMessage(args, _("Too many fields specified for extraction. Using first %s values.") % MAX_FIELDS, CWARN)
vals = vals[:MAX_FIELDS]
for example in vals:
pos += 1
llog("EXAMPLE: %s" % example)
if example in raw:
llog("FOUND: example %s raw %s" % (example, raw))
markedEvent["FIELDNAME%s" % pos] = example
markedEvents[i] = markedEvent
for i, me in markedEvents.items():
llog("ME: %s" % me)
rules = mfl.learn("_raw", markedEvents, counterExamples)
regexes = [rule.getPattern() for rule in rules]
llog("EXAMPLES: %s %s" % (examples, type(examples)))
llog("REGEXES: %s" % regexes)
# for id, e in markedEvents.items():
# for rule in rules:
# extractions = rule.findExtractions(e)
# regexes, extractions = mfl.learn(events, examples, args['counterexamples'])
if len(regexes) > 0:
regex = regexes[0]
except Exception, e:
llog("PROBLEM: %s" % e)
import traceback
llog(traceback.format_exc())
raw = event
markedEvent["_event"] = { sourceField : raw }
for exampleSet in examples:
# !! hack
pos = 0
vals = splitExampleValues(exampleSet)
for example in vals:
pos += 1
llog("EXAMPLE: %s" % example)
if example in raw:
llog("FOUND: example %s raw %s" % (example, raw))
markedEvent["FIELDNAME%s" % pos] = example
markedEvents[i] = markedEvent
for i, me in markedEvents.items():
llog("ME: %s" % me)
rules = mfl.learn("_raw", markedEvents, counterExamples)
regexes = [rule.getPattern() for rule in rules]
llog("EXAMPLES: %s %s" % (examples, type(examples)))
llog("REGEXES: %s" % regexes)
# for id, e in markedEvents.items():
# for rule in rules:
# extractions = rule.findExtractions(e)
# regexes, extractions = mfl.learn(events, examples, args['counterexamples'])
if len(regexes) > 0:
regex = regexes[0]
except Exception, e:
llog("PROBLEM: %s" % e)
import traceback
llog(traceback.format_exc())