class TxtClassificatorWrapper():
_instance = None
_mutex = threading.Semaphore()
def __init__(self, dirname):
if "fake_run" in os.environ:
self.__i = 0
return
self.__documentSizeClassificator = DocumentSizeClustering(os.path.join(dirname, "clusteringData.db"))
self.__newsClassificator = NewsClassificator(os.path.join(dirname, "googlenews"), doTest=False, ignoreKlass = [HEADLINES, SPOTLIGHT, NATION])
@staticmethod
def instance(dirname=None):
if TxtClassificatorWrapper._instance is not None:
return TxtClassificatorWrapper._instance
try:
TxtClassificatorWrapper._mutex.acquire()
if TxtClassificatorWrapper._instance is None:
TxtClassificatorWrapper._instance = TxtClassificatorWrapper(dirname)
return TxtClassificatorWrapper._instance
finally:
TxtClassificatorWrapper._mutex.release()
def classify(self, txt):
if "fake_run" in os.environ:
self.__i += 1
self.__i = self.__i % 3
if self.__i == 1:
return ["long", "p"]
elif self.__i == 2:
return ["long", "b"]
else:
return ["short", "p"]
try:
return self.__documentSizeClassificator.classify(txt), self.__newsClassificator.classify(txt)
except BaseException as e:
logger.exception(u"Error in \"" + unicode(txt) + u"\"")
raise e
def probDist(self, txt):
if "fake_run" in os.environ:
return None
return self.__newsClassificator.prob_classify(txt)
def __init__(self, dirname):
if "fake_run" in os.environ:
self.__i = 0
return
self.__documentSizeClassificator = DocumentSizeClustering(os.path.join(dirname, "clusteringData.db"))
self.__newsClassificator = NewsClassificator(os.path.join(dirname, "googlenews"), doTest=False, ignoreKlass = [HEADLINES, SPOTLIGHT, NATION])
