def generate_dataset(self):
"""Generate the test dataset and store it to disk."""
sets = {
"train": 10,
"test": 5,
}
fields = {
"strings_list":
lambda x: str_to_ascii(self.generate_string_list(x)),
"data": lambda x: np.random.randint(0, 10, (x, 10)),
"number": lambda x: np.array(range(x)),
"field_with_a_long_name_for_printing":
lambda x: np.array(range(x)),
}
lists = {
"list_dummy_data": np.array(range(10)),
"list_dummy_number": np.array(range(10), dtype=np.uint8),
}
dataset = {}
data_fields = {}
for set_name in sets:
dataset[set_name] = self.populate_set(sets[set_name], fields,
lists)
data_fields[set_name] = sorted(dataset[set_name].keys())
return dataset, data_fields
def populate_set(self, size, fields, lists):
dataset = {}
for field in fields:
dataset[field] = fields[field](size)
for field in lists:
dataset[field] = lists[field]
obj_fields = sorted(fields.keys())
dataset['object_fields'] = str_to_ascii(obj_fields)
dataset['object_ids'] = np.array([[i] * len(obj_fields) for i in range(size)])
return dataset
def test_str_to_ascii(sample, output):
res = str_to_ascii(sample)
assert(output == res.tolist())