def test_train_dummy(self):
DummyMethod.clear()
# make a dataset
dirname = os.path.dirname(__file__)
dataset_dir = os.path.join(dirname,'__tmp_dataset')
posts_file = os.path.join(dirname,'posts.json.gz')
app.build_dataset([dataset_dir,posts_file,'user_id','mentions'])
PHONY_SETTINGS_FILE = os.path.join(dirname,'__tmp_settings')
PHONY_MODEL_DIR = 'foobar' # this doesn't have to exist - Dummy won't write to it
fh = open(PHONY_SETTINGS_FILE,'w')
fh.write('{}')
fh.close()
# run the trainer
logger.debug('running training...')
args = ['DummyMethod',PHONY_SETTINGS_FILE,dataset_dir,PHONY_MODEL_DIR]
logger.debug('args = %s' % str(args))
app.train(args)
# check and see if train was called
self.assertTrue(DummyMethod.train_called)
# destroy the dataset
os.system('rm -rf %s' % dataset_dir)
# delete the settings file
os.system('rm %s' % PHONY_SETTINGS_FILE)
def test_train_dummy(self):
DummyMethod.clear()
# make a dataset
dirname = os.path.dirname(__file__)
dataset_dir = os.path.join(dirname, '__tmp_dataset')
posts_file = os.path.join(dirname, 'posts.json.gz')
app.build_dataset([dataset_dir, posts_file, 'user_id', 'mentions'])
PHONY_SETTINGS_FILE = os.path.join(dirname, '__tmp_settings')
PHONY_MODEL_DIR = 'foobar' # this doesn't have to exist - Dummy won't write to it
fh = open(PHONY_SETTINGS_FILE, 'w')
fh.write('{}')
fh.close()
# run the trainer
logger.debug('running training...')
args = [
'DummyMethod', PHONY_SETTINGS_FILE, dataset_dir, PHONY_MODEL_DIR
]
logger.debug('args = %s' % str(args))
app.train(args)
# check and see if train was called
self.assertTrue(DummyMethod.train_called)
# destroy the dataset
os.system('rm -rf %s' % dataset_dir)
# delete the settings file
os.system('rm %s' % PHONY_SETTINGS_FILE)
def test_build_dataset(self):
dirname = os.path.dirname(__file__)
dataset_dir = os.path.join(dirname,'__tmp_dataset')
posts_file = os.path.join(dirname,'posts.json.gz')
app.build_dataset([dataset_dir,posts_file,'user_id','mentions'])
# check if it's there
ds = dataset.Dataset(dataset_dir)
self.assertEquals(len(list(ds.post_iter())),15)
os.system('rm -rf %s' % dataset_dir)
def test_build_dataset(self):
dirname = os.path.dirname(__file__)
dataset_dir = os.path.join(dirname, '__tmp_dataset')
posts_file = os.path.join(dirname, 'posts.json.gz')
app.build_dataset([dataset_dir, posts_file, 'user_id', 'mentions'])
# check if it's there
ds = dataset.Dataset(dataset_dir)
self.assertEquals(len(list(ds.post_iter())), 15)
os.system('rm -rf %s' % dataset_dir)
def test_infer_posts_dummy(self):
DummyMethod.clear()
DummyModel.clear()
# make a dataset
dirname = os.path.dirname(__file__)
dataset_dir = os.path.join(dirname,'__tmp_dataset')
posts_file = os.path.join(dirname,'posts.json.gz')
app.build_dataset([dataset_dir,posts_file,'user_id','mentions'])
PHONY_MODEL_DIR = 'foobar' # this doesn't have to exist - Dummy won't load it
INFER_FILE = os.path.join(dirname,'__tmp_infer')
# run the infer
logger.debug('running inference...')
args = ['-f','DummyMethod',PHONY_MODEL_DIR,dataset_dir,INFER_FILE]
logger.debug('args = %s' % str(args))
app.infer(args,False)
# check the contents of the infer file
fh = open(INFER_FILE,'r')
num_posts = 0
fhiter = iter(fh)
fhiter.next() # skip the header
for line in fhiter:
data = line.split()
num_posts += 1
self.assertEquals(-1,float(data[1]))
self.assertEquals(-1,float(data[2]))
self.assertEquals(num_posts,DummyModel.num_posts_inferred)
# destroy the dataset
os.system('rm -rf %s' % dataset_dir)
# destroy the output file
os.system('rm %s' % INFER_FILE)
def test_infer_posts_dummy(self):
DummyMethod.clear()
DummyModel.clear()
# make a dataset
dirname = os.path.dirname(__file__)
dataset_dir = os.path.join(dirname, '__tmp_dataset')
posts_file = os.path.join(dirname, 'posts.json.gz')
app.build_dataset([dataset_dir, posts_file, 'user_id', 'mentions'])
PHONY_MODEL_DIR = 'foobar' # this doesn't have to exist - Dummy won't load it
INFER_FILE = os.path.join(dirname, '__tmp_infer')
# run the infer
logger.debug('running inference...')
args = ['-f', 'DummyMethod', PHONY_MODEL_DIR, dataset_dir, INFER_FILE]
logger.debug('args = %s' % str(args))
app.infer(args, False)
# check the contents of the infer file
fh = open(INFER_FILE, 'r')
num_posts = 0
fhiter = iter(fh)
fhiter.next() # skip the header
for line in fhiter:
data = line.split()
num_posts += 1
self.assertEquals(-1, float(data[1]))
self.assertEquals(-1, float(data[2]))
self.assertEquals(num_posts, DummyModel.num_posts_inferred)
# destroy the dataset
os.system('rm -rf %s' % dataset_dir)
# destroy the output file
os.system('rm %s' % INFER_FILE)
