def test_download_dataset(self):
"""Test loading a dataset from Url. Note that this test depends on the
accessed web service to be running. It will fail otherwise."""
# Skip test if DOWNLOAD_URL is None
if DOWNLOAD_URL is None:
print('Skipping download test')
return
store = FileSystemDatastore(STORE_DIR)
ds = store.download_dataset(url=DOWNLOAD_URL)
dataset_dir = os.path.join(STORE_DIR, ds.identifier)
self.assertTrue(os.path.isdir(dataset_dir))
self.assertTrue(os.path.isfile(os.path.join(dataset_dir, DATA_FILE)))
self.assertTrue(os.path.isfile(os.path.join(dataset_dir, DESCRIPTOR_FILE)))
self.assertFalse(os.path.isfile(os.path.join(dataset_dir, METADATA_FILE)))
self.validate_class_size_dataset(ds)
# Download file into a given filestore
fs = FileSystemFilestore(FSSTORE_DIR)
ds, fh = store.download_dataset(
url=DOWNLOAD_URL,
filestore=fs
)
self.validate_class_size_dataset(ds)
self.assertEqual(len(fs.list_files()), 1)
self.assertIsNotNone(fh)
self.assertIsNotNone(fs.get_file(fh.identifier))
def setUp(self):
"""Create instances of the default datastore and filestore."""
# Drop directory if it exists
if os.path.isdir(SERVER_DIR):
shutil.rmtree(SERVER_DIR)
os.makedirs(SERVER_DIR)
self.datastore = FileSystemDatastore(DATASTORE_DIR)
self.filestore = FileSystemFilestore(FILESTORE_DIR)
def setUp(self):
"""Create an instance of the default vizier processor for an empty server
directory.
"""
# Drop directory if it exists
if os.path.isdir(SERVER_DIR):
shutil.rmtree(SERVER_DIR)
os.makedirs(SERVER_DIR)
self.datastore = FileSystemDatastore(DATASTORE_DIR)
self.filestore = FileSystemFilestore(FILESTORE_DIR)
def test_deduplicate_annotations(self):
"""Test removing duplicated annotations."""
store = FileSystemDatastore(STORE_DIR)
ds = store.create_dataset(
columns=[
DatasetColumn(identifier=0, name='A'),
DatasetColumn(identifier=1, name='B')
],
rows=[DatasetRow(identifier=0, values=['a', 'b'])],
annotations=DatasetMetadata(
cells=[
DatasetAnnotation(column_id=0, row_id=0, key='X', value=1),
DatasetAnnotation(column_id=0, row_id=0, key='X', value=2),
DatasetAnnotation(column_id=1, row_id=0, key='X', value=3),
DatasetAnnotation(column_id=1, row_id=1, key='X', value=3),
DatasetAnnotation(column_id=0, row_id=0, key='Y', value=1),
DatasetAnnotation(column_id=0, row_id=0, key='X', value=1),
DatasetAnnotation(column_id=0, row_id=0, key='X', value=2),
DatasetAnnotation(column_id=1, row_id=0, key='X', value=3),
DatasetAnnotation(column_id=1, row_id=1, key='X', value=3),
],
columns=[
DatasetAnnotation(column_id=0, key='A', value='x'),
DatasetAnnotation(column_id=1, key='A', value='x'),
DatasetAnnotation(column_id=0, key='A', value='x'),
DatasetAnnotation(column_id=1, key='A', value='x'),
DatasetAnnotation(column_id=0, key='A', value='x'),
DatasetAnnotation(column_id=1, key='A', value='x'),
DatasetAnnotation(column_id=0, key='A', value='x'),
DatasetAnnotation(column_id=1, key='A', value='x')
],
rows=[
DatasetAnnotation(row_id=0, key='E', value=100),
DatasetAnnotation(row_id=0, key='E', value=100)
]
)
)
ds = store.get_dataset(ds.identifier)
self.assertEqual(len(ds.annotations.cells), 4)
self.assertEqual(len(ds.annotations.columns), 2)
self.assertEqual(len(ds.annotations.rows), 1)
annos = ds.annotations.for_cell(column_id=0, row_id=0)
self.assertEqual(len(annos), 3)
self.assertTrue(1 in [a.value for a in annos])
self.assertTrue(2 in [a.value for a in annos])
self.assertFalse(3 in [a.value for a in annos])
self.assertEqual(len(ds.annotations.find_all(values=annos, key='X')), 2)
with self.assertRaises(ValueError):
ds.annotations.find_one(values=annos, key='X')
self.assertEqual(len(ds.annotations.for_column(column_id=0)), 1)
self.assertEqual(len(ds.annotations.for_row(row_id=0)), 1)
annotations = ds.annotations.filter(columns=[1])
self.assertEqual(len(annotations.cells), 1)
self.assertEqual(len(annotations.columns), 1)
self.assertEqual(len(annotations.rows), 1)
def test_load_dataset(self):
"""Test loading a dataset from file."""
store = FileSystemDatastore(STORE_DIR)
ds = store.load_dataset(f_handle=FILE)
dataset_dir = os.path.join(STORE_DIR, ds.identifier)
self.assertTrue(os.path.isdir(dataset_dir))
self.assertTrue(os.path.isfile(os.path.join(dataset_dir, DATA_FILE)))
self.assertTrue(os.path.isfile(os.path.join(dataset_dir, DESCRIPTOR_FILE)))
self.assertFalse(os.path.isfile(os.path.join(dataset_dir, METADATA_FILE)))
self.validate_class_size_dataset(ds)
with self.assertRaises(ValueError):
store.load_dataset(f_handle=None)
def test_query_annotations(self):
"""Test retrieving annotations via the datastore."""
store = FileSystemDatastore(STORE_DIR)
ds = store.create_dataset(
columns=[
DatasetColumn(identifier=0, name='A'),
DatasetColumn(identifier=1, name='B')
],
rows=[DatasetRow(identifier=0, values=['a', 'b'])],
properties=EXAMPLE_PROPERTIES
)
properties = store.get_properties(ds.identifier)
self.assertEqual(len(properties["columns"]), 2)
def test_create_base(self):
"""Test that the datastore base directory is created if it does not
exist.
"""
self.assertTrue(os.path.isdir(STORE_DIR))
shutil.rmtree(STORE_DIR)
self.assertFalse(os.path.isdir(STORE_DIR))
store = FileSystemDatastore(STORE_DIR)
self.assertTrue(os.path.isdir(STORE_DIR))
def set_up(self, engine):
"""Create an empty file server repository."""
if os.path.isdir(SERVER_DIR):
shutil.rmtree(SERVER_DIR)
os.mkdir(SERVER_DIR)
# Setup file server
self.fs = FileSystemFilestore(FILESERVER_DIR)
# Setup the respective datastore and Vizual engine
if engine == ENGINEENV_DEFAULT:
self.datastore = FileSystemDatastore(DATASTORE_DIR)
elif engine == ENGINEENV_MIMIR:
self.datastore = MimirDatastore(DATASTORE_DIR)
def get_datastore(self, identifier):
"""Get the datastore instance for the project with the given identifier.
Paramaters
----------
identifier: string
Unique identifier for datastore
Returns
-------
vizier.datastore.base.Datastore
"""
datastore_dir = os.path.join(self.base_path, identifier)
return FileSystemDatastore(datastore_dir)
def test_query_annotations(self):
"""Test retrieving annotations via the datastore."""
store = FileSystemDatastore(STORE_DIR)
ds = store.create_dataset(
columns=[
DatasetColumn(identifier=0, name='A'),
DatasetColumn(identifier=1, name='B')
],
rows=[DatasetRow(identifier=0, values=['a', 'b'])],
annotations=DatasetMetadata(
cells=[
DatasetAnnotation(column_id=0, row_id=0, key='X', value=1),
DatasetAnnotation(column_id=0, row_id=0, key='X', value=2),
DatasetAnnotation(column_id=1, row_id=0, key='X', value=3),
DatasetAnnotation(column_id=0, row_id=0, key='Y', value=1)
],
columns=[
DatasetAnnotation(column_id=0, key='A', value='x'),
DatasetAnnotation(column_id=1, key='A', value='x')
],
rows=[
DatasetAnnotation(row_id=0, key='E', value=100)
]
)
)
annos = store.get_annotations(ds.identifier, column_id=1)
self.assertEqual(len(annos.columns), 1)
self.assertEqual(len(annos.rows), 0)
self.assertEqual(len(annos.cells), 0)
annos = store.get_annotations(ds.identifier, column_id=0)
self.assertEqual(len(annos.columns), 1)
self.assertEqual(len(annos.rows), 0)
self.assertEqual(len(annos.cells), 0)
annos = store.get_annotations(ds.identifier, row_id=0)
self.assertEqual(len(annos.columns), 0)
self.assertEqual(len(annos.rows), 1)
self.assertEqual(len(annos.cells), 0)
annos = store.get_annotations(ds.identifier, column_id=1, row_id=0)
self.assertEqual(len(annos.columns), 0)
self.assertEqual(len(annos.rows), 0)
self.assertEqual(len(annos.cells), 1)
annos = store.get_annotations(ds.identifier, column_id=0, row_id=0)
self.assertEqual(len(annos.columns), 0)
self.assertEqual(len(annos.rows), 0)
self.assertEqual(len(annos.cells), 3)
def test_properties(self):
"""Test loading a dataset from file."""
store = FileSystemDatastore(STORE_DIR)
ds = store.create_dataset(
columns=[
DatasetColumn(identifier=0, name='A'),
DatasetColumn(identifier=1, name='B')
],
rows=[DatasetRow(identifier=0, values=[1, 2])],
properties=EXAMPLE_PROPERTIES
)
ds = store.get_dataset(ds.identifier)
column_props = ds.properties['columns']
self.assertEqual(len(column_props), 2)
self.assertTrue('A' in [prop['name'] for prop in column_props])
# Reload datastore
store = FileSystemDatastore(STORE_DIR)
ds = store.get_dataset(ds.identifier)
column_props = ds.properties['columns']
self.assertEqual(len(column_props), 2)
def test_create_dataset(self):
"""Test loading a dataset from file."""
store = FileSystemDatastore(STORE_DIR)
ds = store.create_dataset(
columns=[
DatasetColumn(identifier=0, name='A'),
DatasetColumn(identifier=1, name='B')
],
rows=[DatasetRow(identifier=0, values=['a', 'b'])]
)
ds = store.get_dataset(ds.identifier)
column_ids = [col.identifier for col in ds.columns]
self.assertEqual(len(ds.columns), 2)
for id in [0, 1]:
self.assertTrue(id in column_ids)
column_names = [col.name for col in ds.columns]
for name in ['A', 'B']:
self.assertTrue(name in column_names)
rows = ds.fetch_rows()
self.assertEqual(len(rows), 1)
self.assertEqual(rows[0].values, ['a', 'b'])
self.assertEqual(len(ds.annotations.cells), 0)
self.assertEqual(len(ds.annotations.columns), 0)
self.assertEqual(len(ds.annotations.rows), 0)
# Reload the datastore
store = FileSystemDatastore(STORE_DIR)
ds = store.get_dataset(ds.identifier)
column_ids = [col.identifier for col in ds.columns]
self.assertEqual(len(ds.columns), 2)
for id in [0, 1]:
self.assertTrue(id in column_ids)
column_names = [col.name for col in ds.columns]
for name in ['A', 'B']:
self.assertTrue(name in column_names)
rows = ds.fetch_rows()
self.assertEqual(len(rows), 1)
self.assertEqual(rows[0].values, ['a', 'b'])
self.assertEqual(len(ds.annotations.cells), 0)
self.assertEqual(len(ds.annotations.columns), 0)
self.assertEqual(len(ds.annotations.rows), 0)
class TestDefaultPyCellProcessor(unittest.TestCase):
def setUp(self):
"""Create instances of the default datastore and filestore."""
# Drop directory if it exists
if os.path.isdir(SERVER_DIR):
shutil.rmtree(SERVER_DIR)
os.makedirs(SERVER_DIR)
self.datastore = FileSystemDatastore(DATASTORE_DIR)
self.filestore = FileSystemFilestore(FILESTORE_DIR)
def tearDown(self):
"""Clean-up by dropping the server directory.
"""
if os.path.isdir(SERVER_DIR):
shutil.rmtree(SERVER_DIR)
def test_create_dataset_script(self):
"""Test running a script that creates a new datasets."""
cmd = python_cell(source=CREATE_DATASET_PY, validate=True)
result = PyCellTaskProcessor().compute(command_id=cmd.command_id,
arguments=cmd.arguments,
context=TaskContext(
datastore=self.datastore,
filestore=self.filestore))
self.assertTrue(result.is_success)
self.assertIsNotNone(result.provenance.read)
self.assertIsNotNone(result.provenance.write)
self.assertEqual(len(result.provenance.read), 0)
self.assertEqual(len(result.provenance.write), 1)
self.assertTrue('people' in result.provenance.write)
self.assertIsNotNone(result.provenance.write['people'])
self.assertEqual(len(result.outputs.stdout), 1)
self.assertEqual(len(result.outputs.stderr), 0)
self.assertEqual(result.outputs.stdout[0].value, 'Alice\nBob')
def test_print_dataset_script(self):
"""Test running a script that prints rows in an existing datasets."""
fh = self.filestore.upload_file(CSV_FILE)
ds = self.datastore.load_dataset(fh)
cmd = python_cell(source=PRINT_DATASET_PY, validate=True)
result = PyCellTaskProcessor().compute(
command_id=cmd.command_id,
arguments=cmd.arguments,
context=TaskContext(datastore=self.datastore,
filestore=self.filestore,
datasets={'people': ds.identifier}))
self.assertTrue(result.is_success)
self.assertIsNotNone(result.provenance.read)
self.assertIsNotNone(result.provenance.write)
self.assertEqual(len(result.provenance.read), 1)
self.assertEqual(len(result.provenance.write), 0)
self.assertTrue('people' in result.provenance.read)
self.assertIsNotNone(result.provenance.read['people'])
self.assertEqual(len(result.outputs.stdout), 1)
self.assertEqual(len(result.outputs.stderr), 0)
self.assertEqual(result.outputs.stdout[0].value, 'Alice\nBob')
def test_simple_script(self):
"""Test running the simple python script."""
cmd = python_cell(source='print 2+2', validate=True)
result = PyCellTaskProcessor().compute(command_id=cmd.command_id,
arguments=cmd.arguments,
context=TaskContext(
datastore=self.datastore,
filestore=self.filestore,
datasets=dict()))
self.assertTrue(result.is_success)
self.assertEqual(result.outputs.stdout[0].value, '4')
def test_unknown_dataset_script(self):
"""Test running a script that accesses an unknown datasets."""
fh = self.filestore.upload_file(CSV_FILE)
ds = self.datastore.load_dataset(fh)
cmd = python_cell(source=PRINT_UNKNOWN_DATASET_PY, validate=True)
result = PyCellTaskProcessor().compute(
command_id=cmd.command_id,
arguments=cmd.arguments,
context=TaskContext(datastore=self.datastore,
filestore=self.filestore,
datasets={'people': ds.identifier}))
self.assertFalse(result.is_success)
self.assertIsNone(result.provenance.read)
self.assertIsNone(result.provenance.write)
self.assertEqual(len(result.outputs.stdout), 0)
self.assertEqual(len(result.outputs.stderr), 1)
# Running a similar script that catches the error schould be a success
# and the access to the dataset should be recorded in the resulting
# read provenance
cmd = python_cell(source=PRINT_UNKNOWN_DATASET_PY_WITH_TRY_CATCH,
validate=True)
result = PyCellTaskProcessor().compute(
command_id=cmd.command_id,
arguments=cmd.arguments,
context=TaskContext(datastore=self.datastore,
filestore=self.filestore,
datasets={'people': ds.identifier}))
self.assertTrue(result.is_success)
self.assertIsNotNone(result.provenance.read)
self.assertIsNotNone(result.provenance.write)
self.assertEqual(len(result.provenance.read), 1)
self.assertEqual(len(result.provenance.write), 0)
self.assertTrue('employees' in result.provenance.read)
self.assertIsNone(result.provenance.read['employees'])
self.assertEqual(len(result.outputs.stdout), 1)
self.assertEqual(len(result.outputs.stderr), 0)
class TestVizierClient(unittest.TestCase):
def setUp(self):
"""Create instances of the default datastore and filestore."""
# Drop directory if it exists
if os.path.isdir(SERVER_DIR):
shutil.rmtree(SERVER_DIR)
os.makedirs(SERVER_DIR)
self.datastore = FileSystemDatastore(DATASTORE_DIR)
self.filestore = FileSystemFilestore(FILESTORE_DIR)
def tearDown(self):
"""Clean-up by dropping the server directory.
"""
if os.path.isdir(SERVER_DIR):
shutil.rmtree(SERVER_DIR)
def test_create_new_dataset(self):
"""Test creating and updating a new dataset via the client."""
client = VizierDBClient(datastore=self.datastore, datasets=dict())
ds = DatasetClient()
ds.insert_column('Name')
ds.insert_column('Age')
ds.insert_row(['Alice', '23'])
ds.insert_row(['Bob', '25'])
rows = ds.rows
ds = client.create_dataset('MyDataset', ds)
# Ensure the returned dataset contains the input data
self.assertEqual([c.name for c in ds.columns], ['Name', 'Age'])
self.assertEqual([str(v) for v in ds.rows[0].values], ['Alice', '23'])
self.assertEqual([str(v) for v in ds.rows[1].values], ['Bob', '25'])
# Modify the reference to the original rows to ensure that the rows in
# the loaded datasets are not affected
self.assertEqual([str(v) for v in rows[0].values], ['Alice', '23'])
rows[0].set_value(0, 'Jane')
self.assertEqual([str(v) for v in rows[0].values], ['Jane', '23'])
self.assertEqual([str(v) for v in ds.rows[0].values], ['Alice', '23'])
# Update dataset
ds.rows[1].set_value('Age', '26')
client.update_dataset('MyDataset', ds)
ds = client.get_dataset('MyDataset')
self.assertEqual([str(v) for v in ds.rows[1].values], ['Bob', '26'])
# Value error when creating dataset with existing name
with self.assertRaises(ValueError):
client.create_dataset('MyDataset', ds)
# Value error when retrieving unknown dataset
with self.assertRaises(ValueError):
client.get_dataset('SomeDataset')
# Ensure the returned dataset contains the modified data
client.rename_dataset('MyDataset', 'SomeDataset')
ds = client.get_dataset('SomeDataset')
# Ensure that access to unknown datasets is recorded
with self.assertRaises(ValueError):
client.get_dataset('ThisIsNotADataset')
for name in ['somedataset', 'mydataset']:
self.assertTrue(name in client.read)
self.assertTrue(name in client.write)
self.assertTrue('thisisnotadataset' in client.read)
self.assertFalse('thisisnotadataset' in client.write)
def test_dataset_annotations(self):
"""Test creating and updating an existing dataset via the client."""
# Move columns around
ds = self.datastore.load_dataset(self.filestore.upload_file(CSV_FILE))
client = VizierDBClient(datastore=self.datastore,
datasets={DATASET_NAME: ds.identifier})
ds = client.get_dataset(DATASET_NAME)
annotations = ds.annotations
annotations.add(key='comment', value='Good', column_id=0, row_id=1)
annotations.add(key='comment', value='Good', column_id=1, row_id=1)
annotations.add(key='quality', value='Nice', column_id=0, row_id=1)
ds = client.update_dataset(name=DATASET_NAME, dataset=ds)
self.assertEqual(len(ds.annotations.cells), 3)
ds = client.get_dataset(DATASET_NAME)
self.assertEqual(len(ds.annotations.cells), 3)
row = ds.rows[1]
annotations = row.annotations(0)
for key in ['comment', 'quality']:
self.assertTrue(key in list(annotations.keys()))
annotations = row.annotations(1)
self.assertTrue('comment' in list(annotations.keys()))
self.assertFalse('quality' in list(annotations.keys()))
row.set_value(0, 'New Value', clear_annotations=True)
self.assertEqual(len(ds.annotations.cells), 1)
ds = client.update_dataset(name=DATASET_NAME, dataset=ds)
self.assertEqual(len(ds.annotations.cells), 1)
ds = client.get_dataset(DATASET_NAME)
self.assertEqual(len(ds.annotations.cells), 1)
def test_update_existing_dataset(self):
"""Test creating and updating an existing dataset via the client."""
# Move columns around
ds = self.datastore.load_dataset(self.filestore.upload_file(CSV_FILE))
client = VizierDBClient(datastore=self.datastore,
datasets={DATASET_NAME: ds.identifier})
ds = client.get_dataset(DATASET_NAME)
col_1 = [row.get_value(1) for row in ds.rows]
ds.insert_column('empty', 2)
ds = client.update_dataset(DATASET_NAME, ds)
col_2 = [row.get_value(2) for row in ds.rows]
ds.move_column('empty', 1)
ds = client.update_dataset(DATASET_NAME, ds)
for i in range(len(ds.rows)):
row = ds.rows[i]
self.assertEqual(row.values[1], col_2[i])
self.assertEqual(row.values[2], col_1[i])
# Rename
ds.columns[1].name = 'allnone'
ds = client.update_dataset(DATASET_NAME, ds)
for i in range(len(ds.rows)):
row = ds.rows[i]
self.assertEqual(row.get_value('allnone'), col_2[i])
self.assertEqual(row.values[2], col_1[i])
# Insert row
row = ds.insert_row()
row.set_value('Name', 'Zoe')
ds = client.create_dataset('upd', ds)
self.assertEqual(len(ds.rows), 3)
r2 = ds.rows[2]
self.assertEqual(r2.identifier, 2)
self.assertEqual(r2.values, ['Zoe', None, None, None])
# Annotations
ds = client.get_dataset(DATASET_NAME)
col = ds.get_column('Age')
row = ds.rows[0]
ds.annotations.add(column_id=col.identifier,
row_id=row.identifier,
key='user:comment',
value='My Comment')
ds = client.update_dataset(DATASET_NAME, ds)
annotations = ds.rows[0].annotations('Age').find_all('user:comment')
self.assertEqual(len(annotations), 1)
anno = annotations[0]
self.assertEqual(anno.key, 'user:comment')
self.assertEqual(anno.value, 'My Comment')
ds.annotations.add(column_id=col.identifier,
row_id=row.identifier,
key='user:comment',
value='Another Comment')
ds = client.update_dataset(DATASET_NAME, ds)
annotations = ds.rows[0].annotations('Age').find_all('user:comment')
self.assertEqual(len(annotations), 2)
self.assertEqual(list(ds.rows[0].annotations('Age').keys()),
['user:comment'])
values = [a.value for a in annotations]
for val in ['My Comment', 'Another Comment']:
self.assertTrue(val in values)
anno = ds.rows[0].annotations('Age').find_one('user:comment')
anno.key = 'user:issue'
anno.value = 'Some Issue'
ds = client.update_dataset(DATASET_NAME, ds)
annotations = ds.rows[0].annotations('Age').find_all('user:comment')
self.assertEqual(len(annotations), 1)
keys = list(ds.rows[0].annotations('Age').keys())
for key in ['user:comment', 'user:issue']:
self.assertTrue(key in keys)
values = [
a.value
for a in ds.rows[0].annotations('Age').find_all('user:issue')
]
for val in ['Some Issue']:
self.assertTrue(val in values)
ds.annotations.remove(
column_id=col.identifier,
row_id=row.identifier,
key='user:issue',
)
ds = client.update_dataset(DATASET_NAME, ds)
annotations = ds.rows[0].annotations('Age').find_all('user:issue')
self.assertEqual(len(annotations), 0)
annotations = ds.rows[0].annotations('Age').find_all('user:comment')
self.assertEqual(len(annotations), 1)
# Delete column
ds = client.get_dataset(DATASET_NAME)
ds.delete_column('Age')
client.update_dataset(DATASET_NAME, ds)
ds = client.get_dataset(DATASET_NAME)
names = [col.name.upper() for col in ds.columns]
self.assertTrue('NAME' in names)
self.assertFalse('AGE' in names)
self.assertTrue(DATASET_NAME in client.read)
self.assertTrue(DATASET_NAME in client.write)
self.assertFalse('upd' in client.read)
self.assertTrue('upd' in client.write)
class TestVizierClient(unittest.TestCase):
def setUp(self):
"""Create instances of the default datastore and filestore."""
# Drop directory if it exists
if os.path.isdir(SERVER_DIR):
shutil.rmtree(SERVER_DIR)
os.makedirs(SERVER_DIR)
self.datastore=FileSystemDatastore(DATASTORE_DIR)
self.filestore=FileSystemFilestore(FILESTORE_DIR)
def tearDown(self):
"""Clean-up by dropping the server directory.
"""
if os.path.isdir(SERVER_DIR):
shutil.rmtree(SERVER_DIR)
def test_create_new_dataset(self):
"""Test creating and updating a new dataset via the client."""
client = VizierDBClient(
datastore=self.datastore,
datasets=dict(),
dataobjects=dict(),
source="",
project_id=7
)
ds = DatasetClient()
ds.insert_column('Name')
ds.insert_column('Age')
ds.insert_row(['Alice', '23'])
ds.insert_row(['Bob', '25'])
rows = ds.rows
ds = client.create_dataset('MyDataset', ds)
# Ensure the returned dataset contains the input data
self.assertEqual([c.name for c in ds.columns], ['Name', 'Age'])
self.assertEqual([str(v) for v in ds.rows[0].values], ['Alice', '23'])
self.assertEqual([str(v) for v in ds.rows[1].values], ['Bob', '25'])
# Modify the reference to the original rows to ensure that the rows in
# the loaded datasets are not affected
self.assertEqual([str(v) for v in rows[0].values], ['Alice', '23'])
rows[0].set_value(0, 'Jane')
self.assertEqual([str(v) for v in rows[0].values], ['Jane', '23'])
self.assertEqual([str(v) for v in ds.rows[0].values], ['Alice', '23'])
# Update dataset
ds.rows[1].set_value('Age', '26')
ds.save()
ds = client.get_dataset('MyDataset')
self.assertEqual([str(v) for v in ds.rows[1].values], ['Bob', '26'])
# Value error when creating dataset with existing name
with self.assertRaises(ValueError):
client.create_dataset('MyDataset', ds)
# Value error when retrieving unknown dataset
with self.assertRaises(ValueError):
client.get_dataset('SomeDataset')
# Ensure the returned dataset contains the modified data
client.rename_dataset('MyDataset', 'SomeDataset')
ds = client.get_dataset('SomeDataset')
# Ensure that access to unknown datasets is recorded
with self.assertRaises(ValueError):
client.get_dataset('ThisIsNotADataset')
self.assertTrue('mydataset' in client.write)
self.assertTrue('somedataset' in client.write)
self.assertTrue('thisisnotadataset' in client.read)
self.assertFalse('thisisnotadataset' in client.write)
def test_update_existing_dataset(self):
"""Test creating and updating an existing dataset via the client."""
# Move columns around
ds = self.datastore.load_dataset(self.filestore.upload_file(CSV_FILE))
client = VizierDBClient(
datastore=self.datastore,
datasets={DATASET_NAME:ds},
dataobjects=dict(),
source="",
project_id=7
)
ds = client.get_dataset(DATASET_NAME)
col_1 = [row.get_value(1) for row in ds.rows]
ds.insert_column('empty', 3)
ds = client.update_dataset(DATASET_NAME, ds)
col_2 = [row.get_value(2) for row in ds.rows]
ds.move_column('empty', 1)
ds = client.update_dataset(DATASET_NAME, ds)
for i in range(len(ds.rows)):
row = ds.rows[i]
self.assertEqual(row.values[3], col_2[i])
self.assertEqual(row.values[2], col_1[i])
# Rename
ds.columns[1].name = 'allnone'
ds = client.update_dataset(DATASET_NAME, ds)
for i in range(len(ds.rows)):
row = ds.rows[i]
self.assertEqual(row.get_value('allnone'), None)
self.assertEqual(row.values[2], col_1[i])
# Insert row
row = ds.insert_row()
row.set_value('Name', 'Zoe')
ds = client.create_dataset('upd', ds)
self.assertEqual(len(ds.rows), 3)
r2 = ds.rows[2]
self.assertEqual(r2.values, ['Zoe', None, None, None])
# Delete column
ds = client.get_dataset(DATASET_NAME)
ds.delete_column('Age')
client.update_dataset(DATASET_NAME, ds)
ds = client.get_dataset(DATASET_NAME)
names = [col.name.upper() for col in ds.columns]
self.assertTrue('NAME' in names)
self.assertFalse('AGE' in names)
self.assertTrue(DATASET_NAME in client.read)
self.assertTrue(DATASET_NAME in client.write)
self.assertFalse('upd' in client.read)
self.assertTrue('upd' in client.write)
def test_get_dataset(self):
"""Test accessing dataset handle and descriptor."""
# None for non-existing dataset
store = FileSystemDatastore(STORE_DIR)
self.assertIsNone(store.get_dataset('0000'))
ds_id = store.load_dataset(f_handle=FILE).identifier
self.assertIsNotNone(store.get_dataset(ds_id))
self.assertIsNone(store.get_dataset('0000'))
# Reload store to ensure the dataset still exists
store = FileSystemDatastore(STORE_DIR)
self.assertIsNotNone(store.get_dataset(ds_id))
self.assertIsNone(store.get_dataset('0000'))
self.validate_class_size_dataset(store.get_dataset(ds_id))
# Load a second dataset
ds_id_2 = store.load_dataset(f_handle=FILE).identifier
self.assertIsNotNone(store.get_dataset(ds_id))
self.assertIsNotNone(store.get_dataset(ds_id_2))
# Reload store to ensure the dataset still exists
store = FileSystemDatastore(STORE_DIR)
self.assertIsNotNone(store.get_dataset(ds_id))
self.assertIsNotNone(store.get_dataset(ds_id_2))
def test_delete_dataset(self):
"""Test deleting datasets."""
# None for non-existing dataset
store = FileSystemDatastore(STORE_DIR)
ds_id = store.load_dataset(f_handle=FILE).identifier
ds_id_2 = store.load_dataset(f_handle=FILE).identifier
self.assertIsNotNone(store.get_dataset(ds_id))
self.assertIsNotNone(store.get_dataset(ds_id_2))
store.delete_dataset(ds_id)
self.assertIsNone(store.get_dataset(ds_id))
self.assertIsNotNone(store.get_dataset(ds_id_2))
# Reload store to ensure only one dataset still exists
store = FileSystemDatastore(STORE_DIR)
self.assertIsNone(store.get_dataset(ds_id))
self.assertIsNotNone(store.get_dataset(ds_id_2))
# Delete the second dataset
store.delete_dataset(ds_id_2)
store = FileSystemDatastore(STORE_DIR)
self.assertIsNone(store.get_dataset(ds_id))
self.assertIsNone(store.get_dataset(ds_id_2))
def test_update_annotations(self):
"""Test updating annotations via the datastore."""
store = FileSystemDatastore(STORE_DIR)
ds = store.create_dataset(
columns=[
DatasetColumn(identifier=0, name='A'),
DatasetColumn(identifier=1, name='B')
],
rows=[DatasetRow(identifier=0, values=['a', 'b'])],
annotations=DatasetMetadata(
cells=[
DatasetAnnotation(column_id=0, row_id=0, key='X', value=1),
DatasetAnnotation(column_id=0, row_id=0, key='X', value=2),
DatasetAnnotation(column_id=1, row_id=0, key='X', value=3),
DatasetAnnotation(column_id=0, row_id=0, key='Y', value=1)
],
columns=[
DatasetAnnotation(column_id=0, key='A', value='x'),
DatasetAnnotation(column_id=1, key='A', value='x')
],
rows=[
DatasetAnnotation(row_id=0, key='E', value=100)
]
)
)
# INSERT row annotatins
store.update_annotation(
ds.identifier,
key='D',
row_id=0,
new_value=200
)
annos = store.get_annotations(ds.identifier, row_id=0)
self.assertEqual(len(annos.rows), 2)
for key in ['D', 'E']:
self.assertTrue(key in [a.key for a in annos.rows])
for val in [100, 200]:
self.assertTrue(val in [a.value for a in annos.rows])
# UPDATE column annotation
store.update_annotation(
ds.identifier,
key='A',
column_id=1,
old_value='x',
new_value='y'
)
annos = store.get_annotations(ds.identifier, column_id=1)
self.assertEqual(annos.columns[0].key, 'A')
self.assertEqual(annos.columns[0].value, 'y')
# DELETE cell annotation
store.update_annotation(
ds.identifier,
key='X',
column_id=0,
row_id=0,
old_value=2,
)
annos = store.get_annotations(ds.identifier, column_id=0, row_id=0)
self.assertEqual(len(annos.cells), 2)
for a in annos.cells:
self.assertNotEqual(a.value, 2)
result = store.update_annotation(
ds.identifier,
key='X',
column_id=1,
row_id=0,
old_value=3,
)
self.assertTrue(result)
annos = store.get_annotations(ds.identifier, column_id=1, row_id=0)
self.assertEqual(len(annos.cells), 0)
class TestDefaultVizualProcessor(unittest.TestCase):
def setUp(self):
"""Create an instance of the default vizier processor for an empty server
directory.
"""
# Drop directory if it exists
if os.path.isdir(SERVER_DIR):
shutil.rmtree(SERVER_DIR)
os.makedirs(SERVER_DIR)
self.datastore = FileSystemDatastore(DATASTORE_DIR)
self.filestore = FileSystemFilestore(FILESTORE_DIR)
def tearDown(self):
"""Clean-up by dropping the server directory.
"""
if os.path.isdir(SERVER_DIR):
shutil.rmtree(SERVER_DIR)
def count_non_null_values(self, data, column_index):
"""Return the number of values in a column that are not None."""
count = 0
for row in data:
if not row[column_index] is None:
count += 1
return count
def test_query(self):
"""Test running a query for simple chart plots."""
ds = self.datastore.load_dataset(self.filestore.upload_file(LOAD_FILE))
view = ChartViewHandle(dataset_name='ABC', x_axis=2)
view.add_series(1, range_start=25, range_end=30)
view.add_series(0, range_start=25, range_end=30)
view.add_series(3, range_start=25, range_end=30)
data = ChartQuery().exec_query(dataset=ds, view=view)[0]
self.assertEqual(len(data), 6)
for row in data:
self.assertEqual(len(row), 3)
self.assertTrue(isinstance(data[0][0], int))
self.assertTrue(isinstance(data[0][1], float))
# Remove interval end for one series. This should return all rows
# starting from index 25
view = ChartViewHandle(dataset_name='ABC', x_axis=2)
view.add_series(1, range_start=25, range_end=30)
view.add_series(0, range_start=25)
view.add_series(3, range_start=25, range_end=30)
data = ChartQuery().exec_query(dataset=ds, view=view)[0]
self.assertEqual(len(data), 29)
self.assertIsNone(data[28][0])
self.assertIsNotNone(data[28][1])
self.assertIsNone(data[28][2])
for row in data:
self.assertEqual(len(row), 3)
# Remove interval start for another series. The first series will
# contain 31 values, the second 29, and the third 6
view = ChartViewHandle(dataset_name='ABC', x_axis=2)
view.add_series(1, range_end=30)
view.add_series(0, range_start=25)
view.add_series(3, range_start=25, range_end=30)
data = ChartQuery().exec_query(dataset=ds, view=view)[0]
self.assertEqual(len(data), 31)
self.assertEqual(self.count_non_null_values(data, 0), 31)
self.assertEqual(self.count_non_null_values(data, 1), 29)
self.assertEqual(self.count_non_null_values(data, 2), 6)
for row in data:
self.assertEqual(len(row), 3)
# Without any range constraints the result should contain all 54 rows
view = ChartViewHandle(dataset_name='ABC', x_axis=2)
view.add_series(1, label='A')
view.add_series(0, label='B')
view.add_series(3)
data = ChartQuery().exec_query(dataset=ds, view=view)[0]
self.assertEqual(len(data), 54)
def test_default_config(self):
"""Run workflow with default configuration."""
# Create new work trail and retrieve the HEAD workflow of the default
# branch
self.run_workflow(FileSystemDatastore(DATASTORE_DIR))
class TestDefaultPlotProcessor(unittest.TestCase):
def setUp(self):
"""Create instances of the default datastore and filestore."""
# Drop directory if it exists
if os.path.isdir(SERVER_DIR):
shutil.rmtree(SERVER_DIR)
os.makedirs(SERVER_DIR)
self.datastore=FileSystemDatastore(DATASTORE_DIR)
self.filestore=FileSystemFilestore(FILESTORE_DIR)
def tearDown(self):
"""Clean-up by dropping the server directory.
"""
if os.path.isdir(SERVER_DIR):
shutil.rmtree(SERVER_DIR)
def test_advanced_plot(self):
"""Test running the simple plot command with a more advanced chart
definition.
"""
fh = self.filestore.upload_file(TSV_FILE)
ds = self.datastore.load_dataset(fh)
cmd = create_plot(
dataset_name=DATASET_NAME,
chart_name='My Chart',
series=[{'column': 1, 'range': '25:30', 'label': 'A'}, {'column': 0, 'range': '25:30'}],
validate=True
)
result = PlotProcessor().compute(
command_id=cmd.command_id,
arguments=cmd.arguments,
context=TaskContext(
project_id=0,
datastore=self.datastore,
filestore=self.filestore,
artifacts={DATASET_NAME: ds}
)
)
chart = result.outputs.stdout[0].value
self.assertEqual(chart['data']['data'][0]['label'], 'A')
self.assertEqual(chart['data']['data'][1]['label'], 'average_class_size')
self.assertEqual(chart['result']['series'][0]['label'], 'A')
self.assertEqual(chart['result']['series'][1]['label'], 'average_class_size')
self.assertEqual(len(chart['result']['series'][0]['data']), 6)
self.assertEqual(len(chart['result']['series'][1]['data']), 6)
def test_simple_plot(self):
"""Test running the simple plot command."""
fh = self.filestore.upload_file(CSV_FILE)
ds = self.datastore.load_dataset(fh)
cmd = create_plot(
dataset_name=DATASET_NAME,
chart_name='My Chart',
series=[{'column': 1}],
validate=True
)
result = PlotProcessor().compute(
command_id=cmd.command_id,
arguments=cmd.arguments,
context=TaskContext(
project_id=0,
datastore=self.datastore,
filestore=self.filestore,
artifacts={DATASET_NAME: ds}
)
)
class TestDefaultVizualProcessor(unittest.TestCase):
def setUp(self):
"""Create an instance of the default vizier processor for an empty server
directory.
"""
# Drop directory if it exists
if os.path.isdir(SERVER_DIR):
shutil.rmtree(SERVER_DIR)
os.makedirs(SERVER_DIR)
self.processor = VizualTaskProcessor(api=DefaultVizualApi())
self.datastore = FileSystemDatastore(DATASTORE_DIR)
self.filestore = FileSystemFilestore(FILESTORE_DIR)
def tearDown(self):
"""Clean-up by dropping the server directory.
"""
if os.path.isdir(SERVER_DIR):
shutil.rmtree(SERVER_DIR)
def test_create_api_from_dictionary(self):
"""Test creating the processor instance with properties parameter
instead of api.
"""
processor = VizualTaskProcessor(
properties={
PROPERTY_API:
ClassLoader.to_dict(
module_name='vizier.engine.packages.vizual.api.fs',
class_name='DefaultVizualApi')
})
fh = self.filestore.upload_file(CSV_FILE)
cmd = vizual.load_dataset(dataset_name=DATASET_NAME,
file={pckg.FILE_ID: fh.identifier},
validate=True)
result = processor.compute(command_id=cmd.command_id,
arguments=cmd.arguments,
context=TaskContext(
project_id=5,
datastore=self.datastore,
filestore=self.filestore,
artifacts={}))
self.assertIsNotNone(result.provenance.write)
self.assertTrue(DATASET_NAME in result.provenance.write)
dataset_id = result.provenance.write[DATASET_NAME].identifier
self.assertTrue(result.provenance.read is None
or len(result.provenance.read) == 0)
self.assertIsNotNone(result.provenance.resources)
self.assertEqual(result.provenance.resources[RESOURCE_DATASET],
dataset_id)
def load_dataset(self):
"""Load a single dataset and return the resulting database state."""
fh = self.filestore.upload_file(CSV_FILE)
cmd = vizual.load_dataset(dataset_name=DATASET_NAME,
file={pckg.FILE_ID: fh.identifier},
validate=True)
result = self.processor.compute(command_id=cmd.command_id,
arguments=cmd.arguments,
context=TaskContext(
project_id=5,
datastore=self.datastore,
filestore=self.filestore,
artifacts={}))
return result.provenance.write
def test_delete_column(self):
"""Test functionality to delete a column."""
cmd = vizual.delete_column(dataset_name=DATASET_NAME,
column=1,
validate=True)
self.validate_command(cmd)
def test_delete_row(self):
"""Test functionality to delete a row."""
cmd = vizual.delete_row(dataset_name=DATASET_NAME,
row=1,
validate=True)
self.validate_command(cmd)
def test_drop_dataset(self):
"""Test functionality to drop a dataset."""
cmd = vizual.drop_dataset(dataset_name=DATASET_NAME, validate=True)
datasets = self.load_dataset()
dataset_id = datasets[DATASET_NAME]
result = self.processor.compute(
command_id=cmd.command_id,
arguments=cmd.arguments,
context=TaskContext(project_id=5,
datastore=self.datastore,
filestore=self.filestore,
artifacts={DATASET_NAME: dataset_id}))
self.assertFalse(DATASET_NAME in result.provenance.read)
self.assertTrue(DATASET_NAME in result.provenance.delete)
self.assertFalse(DATASET_NAME in result.provenance.write)
def test_filter_columns(self):
"""Test projection of a dataset."""
# Create a new dataset
cmd = vizual.projection(dataset_name=DATASET_NAME,
columns=[{
'column': 1
}, {
'column': 2,
'name': 'MyName'
}],
validate=True)
self.validate_command(cmd)
def test_insert_column(self):
"""Test functionality to insert a columns."""
cmd = vizual.insert_column(dataset_name=DATASET_NAME,
position=1,
name='My Col',
validate=True)
self.validate_command(cmd)
def test_insert_row(self):
"""Test functionality to insert a row."""
# Create a new dataset
cmd = vizual.insert_row(dataset_name=DATASET_NAME,
position=1,
validate=True)
self.validate_command(cmd)
def test_load_dataset(self):
"""Test functionality to load a dataset."""
# Create a new dataset
fh = self.filestore.upload_file(CSV_FILE)
cmd = vizual.load_dataset(dataset_name='ABC',
file={pckg.FILE_ID: fh.identifier},
validate=True)
result = self.processor.compute(command_id=cmd.command_id,
arguments=cmd.arguments,
context=TaskContext(
project_id=5,
datastore=self.datastore,
filestore=self.filestore,
artifacts={}))
self.assertIsNotNone(result.provenance.write)
self.assertTrue('abc' in result.provenance.write)
dataset_id = result.provenance.write['abc'].identifier
self.assertTrue(result.provenance.read is None
or len(result.provenance.read) == 0)
self.assertIsNotNone(result.provenance.resources)
self.assertEqual(result.provenance.resources[RESOURCE_DATASET],
dataset_id)
# Running load again will not change the dataset identifier
result = self.processor.compute(
command_id=cmd.command_id,
arguments=cmd.arguments,
context=TaskContext(project_id=5,
datastore=self.datastore,
filestore=self.filestore,
artifacts={},
resources=result.provenance.resources))
self.assertEqual(result.provenance.write['abc'].identifier, dataset_id)
self.assertEqual(result.provenance.resources[RESOURCE_DATASET],
dataset_id)
def test_move_column(self):
"""Test functionality to move a column."""
cmd = vizual.move_column(dataset_name=DATASET_NAME,
column=0,
position=1,
validate=True)
self.validate_command(cmd)
def test_move_row(self):
"""Test functionality to move a row."""
cmd = vizual.move_row(dataset_name=DATASET_NAME,
row=0,
position=1,
validate=True)
self.validate_command(cmd)
def test_rename_column(self):
"""Test functionality to rename a column."""
cmd = vizual.rename_column(dataset_name=DATASET_NAME,
column=1,
name='The col',
validate=True)
self.validate_command(cmd)
def test_rename_dataset(self):
"""Test functionality to rename a dataset."""
cmd = vizual.rename_dataset(dataset_name=DATASET_NAME,
new_name='XYZ',
validate=True)
datasets = self.load_dataset()
dataset_id = datasets[DATASET_NAME]
result = self.processor.compute(
command_id=cmd.command_id,
arguments=cmd.arguments,
context=TaskContext(project_id=5,
datastore=self.datastore,
filestore=self.filestore,
artifacts={DATASET_NAME: dataset_id}))
self.assertFalse(DATASET_NAME in result.provenance.write)
self.assertFalse(DATASET_NAME in result.provenance.read)
self.assertTrue(DATASET_NAME in result.provenance.delete)
self.assertFalse(DATASET_NAME in result.provenance.write)
self.assertTrue('xyz' in result.provenance.write)
def test_sort_dataset(self):
"""Test sorting a dataset."""
cmd = vizual.sort_dataset(dataset_name=DATASET_NAME,
columns=[{
'column': 1,
'order': 'Z-A'
}, {
'column': 2,
'order': 'A-Z'
}],
validate=True)
self.validate_command(cmd)
def test_update_cell(self):
"""Test functionality to update a dataset cell."""
# Create a new dataset
datasets = self.load_dataset()
dataset = self.datastore.get_dataset(datasets[DATASET_NAME].identifier)
row_ids = [row.identifier for row in dataset.fetch_rows()]
cmd = vizual.update_cell(dataset_name=DATASET_NAME,
column=1,
row=row_ids[0],
value=9,
validate=True)
self.validate_command(cmd, dataset=dataset)
def validate_command(self, cmd, dataset=None):
"""Validate execution of the given command."""
if dataset is None:
datasets = self.load_dataset()
dataset = datasets[DATASET_NAME]
result = self.processor.compute(command_id=cmd.command_id,
arguments=cmd.arguments,
context=TaskContext(
project_id=5,
datastore=self.datastore,
filestore=self.filestore,
artifacts={DATASET_NAME: dataset}))
self.assertNotEqual(result.provenance.write[DATASET_NAME].identifier,
dataset.identifier)
self.assertIsNotNone(result.provenance.read)
self.assertEqual(result.provenance.read[DATASET_NAME],
dataset.identifier)
self.assertIsNotNone(result.provenance.write)
with self.assertRaises(ValueError):
result = self.processor.compute(command_id=cmd.command_id,
arguments=cmd.arguments,
context=TaskContext(
project_id=5,
datastore=self.datastore,
filestore=self.filestore,
artifacts={}))
class TestDefaultVizualApi(unittest.TestCase):
def setUp(self):
"""Create an instance of the default vizier API for an empty server
directory.
"""
# Drop directory if it exists
if os.path.isdir(SERVER_DIR):
shutil.rmtree(SERVER_DIR)
os.makedirs(SERVER_DIR)
self.api = DefaultVizualApi()
self.datastore = FileSystemDatastore(DATASTORE_DIR)
self.filestore = FileSystemFilestore(FILESTORE_DIR)
def tearDown(self):
"""Clean-up by dropping the server directory.
"""
if os.path.isdir(SERVER_DIR):
shutil.rmtree(SERVER_DIR)
def test_delete_column(self):
"""Test functionality to delete a column."""
# Create a new dataset
fh = self.filestore.upload_file(CSV_FILE)
ds = self.api.load_dataset(datastore=self.datastore,
filestore=self.filestore,
file_id=fh.identifier).dataset
ds_rows = ds.fetch_rows()
# Keep track of column and row identifier
col_ids = [col.identifier for col in ds.columns]
row_ids = [row.identifier for row in ds_rows]
# Delete Age column
col_id = ds.column_by_name('AGE').identifier
result = self.api.delete_column(ds.identifier, col_id, self.datastore)
# Resulting dataset should differ from previous one
self.assertNotEqual(result.dataset.identifier, ds.identifier)
# Retrieve modified dataset and ensure that it cobtains the following
#
# Name, Salary
# ------------
# Alice, 35K
# Bob, 30K
ds = self.datastore.get_dataset(result.dataset.identifier)
ds_rows = ds.fetch_rows()
# Schema is Name, Salary
self.assertEqual(len(ds.columns), 2)
self.assertEqual(ds.columns[0].name.upper(), 'NAME')
self.assertEqual(ds.columns[1].name.upper(), 'SALARY')
# Make sure column identifier haven't changed
del col_ids[1]
for i in range(len(ds.columns)):
self.assertEqual(ds.columns[i].identifier, col_ids[i])
# Make sure that all rows only have two columns
row = ds_rows[0]
self.assertEqual(len(row.values), 2)
self.assertEqual(len(row.values), 2)
self.assertEqual(row.values[0], 'Alice')
self.assertEqual(row.values[1], '35K')
row = ds_rows[1]
self.assertEqual(len(row.values), 2)
self.assertEqual(len(row.values), 2)
self.assertEqual(row.values[0], 'Bob')
self.assertEqual(row.values[1], '30K')
# Ensure that row identifier haven't changed
for i in range(len(ds_rows)):
self.assertEqual(ds_rows[i].identifier, row_ids[i])
# Ensure exception is thrown if dataset identifier is unknown
with self.assertRaises(ValueError):
self.api.delete_column('unknown:uri', 0, self.datastore)
# Ensure exception is thrown if column identifier is unknown
with self.assertRaises(ValueError):
self.api.delete_column(ds.identifier, col_id, self.datastore)
def test_delete_row(self):
"""Test functionality to delete a row."""
# Create a new dataset
fh = self.filestore.upload_file(CSV_FILE)
ds = self.api.load_dataset(datastore=self.datastore,
filestore=self.filestore,
file_id=fh.identifier).dataset
ds_rows = ds.fetch_rows()
# Keep track of column and row identifier
col_ids = [col.identifier for col in ds.columns]
row_ids = [row.identifier for row in ds_rows]
# Delete second row
result = self.api.delete_row(ds.identifier, 1, self.datastore)
del row_ids[1]
# Resulting dataset should differ from previous one
self.assertNotEqual(result.dataset.identifier, ds.identifier)
# Retrieve modified dataset and ensure that it contains the following
# data:
#
# Name, Age, Salary
# ------------
# Alice, 23, 35K
ds = self.datastore.get_dataset(result.dataset.identifier)
ds_rows = ds.fetch_rows()
# Schema is Name, Salary
col_names = ['Name', 'Age', 'Salary']
self.assertEqual(len(ds.columns), len(col_names))
for i in range(len(ds.columns)):
self.assertEqual(ds.columns[i].name.upper(), col_names[i].upper())
# Make sure column identifier haven't changed
for i in range(len(ds.columns)):
self.assertEqual(ds.columns[i].identifier, col_ids[i])
# There should only be one row
self.assertEqual(len(ds_rows), 1)
# Ensure that row identifier haven't changed
for i in range(len(ds_rows)):
self.assertEqual(ds_rows[i].identifier, row_ids[i])
# Ensure exception is thrown if dataset is unknown
with self.assertRaises(ValueError):
self.api.delete_row('unknown:uri', 0, self.datastore)
# Ensure exception is thrown if row index is out of bounds
with self.assertRaises(ValueError):
self.api.delete_row(ds.identifier, 100, self.datastore)
def test_filter_columns(self):
"""Test projection of a dataset."""
# Create a new dataset
fh = self.filestore.upload_file(CSV_FILE)
ds = self.api.load_dataset(datastore=self.datastore,
filestore=self.filestore,
file_id=fh.identifier).dataset
result = self.api.filter_columns(ds.identifier, [2, 0], ['BD', None],
self.datastore)
ds = self.datastore.get_dataset(result.dataset.identifier)
self.assertEqual(len(ds.columns), 2)
self.assertEqual(ds.columns[0].identifier, 2)
self.assertEqual(ds.columns[0].name.upper(), 'BD')
self.assertEqual(ds.columns[1].identifier, 0)
self.assertEqual(ds.columns[1].name.upper(), 'NAME')
rows = ds.fetch_rows()
self.assertEqual(rows[0].values, ['35K', 'Alice'])
self.assertEqual(rows[1].values, ['30K', 'Bob'])
with self.assertRaises(ValueError):
self.api.filter_columns(ds.identifier, [0, 1], ['BD', None],
self.datastore)
def test_insert_column(self):
"""Test functionality to insert a columns."""
# Create a new dataset
fh = self.filestore.upload_file(CSV_FILE)
ds = self.api.load_dataset(datastore=self.datastore,
filestore=self.filestore,
file_id=fh.identifier).dataset
ds_rows = ds.fetch_rows()
# Keep track of column and row identifier
col_ids = [col.identifier for col in ds.columns]
row_ids = [row.identifier for row in ds_rows]
# Insert columns at position 1
col_ids.insert(1, ds.max_column_id() + 1)
result = self.api.insert_column(ds.identifier, 1, 'Height',
self.datastore)
# Resulting dataset should differ from previous one
self.assertNotEqual(result.dataset.identifier, ds.identifier)
# Retrieve dataset and ensure that it has the following schema:
# Name, Height, Age, Salary
ds = self.datastore.get_dataset(result.dataset.identifier)
col_names = ['Name', 'Height', 'Age', 'Salary']
# Ensure that there are four rows
self.assertEqual(len(ds.columns), len(col_names))
for i in range(len(col_names)):
col = ds.columns[i]
self.assertEqual(col.identifier, col_ids[i])
self.assertEqual(col.name.upper(), col_names[i].upper())
# Insert columns at last position
col_ids.append(ds.max_column_id() + 1)
col_names.append('Weight')
result = self.api.insert_column(ds.identifier, 4, 'Weight',
self.datastore)
# Resulting dataset should differ from previous one
self.assertNotEqual(result.dataset.identifier, ds.identifier)
# Retrieve dataset and ensure that it has the following schema:
# Name, Height, Age, Salary, Weight
ds = self.datastore.get_dataset(result.dataset.identifier)
ds_rows = ds.fetch_rows()
# Ensure that there are five rows
self.assertEqual(len(ds.columns), len(col_names))
for i in range(len(col_names)):
col = ds.columns[i]
self.assertEqual(col.identifier, col_ids[i])
self.assertEqual(col.name.upper(), col_names[i].upper())
# The cell values for new columns are None all other values are not None
for row in ds_rows:
for i in range(len(ds.columns)):
if i == 1 or i == 4:
self.assertIsNone(row.values[i])
else:
self.assertTrue(row.values[i])
# Ensure that row identifier haven't changed
for i in range(len(ds_rows)):
self.assertEqual(ds_rows[i].identifier, row_ids[i])
# Ensure exception is thrown if dataset identifier is unknown
with self.assertRaises(ValueError):
self.api.insert_column('unknown:uri', 1, 'Height', self.datastore)
# Ensure exception is thrown if column name is invalid
self.api.insert_column(ds.identifier, 1, 'Height from ground',
self.datastore)
with self.assertRaises(ValueError):
self.api.insert_column(ds.identifier, 1, 'Height from ground!@#',
self.datastore)
# Ensure exception is thrown if column position is out of bounds
with self.assertRaises(ValueError):
self.api.insert_column(ds.identifier, 100, 'Height',
self.datastore)
def test_insert_row(self):
"""Test functionality to insert a row."""
# Create a new dataset
fh = self.filestore.upload_file(CSV_FILE)
ds = self.api.load_dataset(datastore=self.datastore,
filestore=self.filestore,
file_id=fh.identifier).dataset
# Keep track of column and row identifier
ds_rows = ds.fetch_rows()
col_ids = [col.identifier for col in ds.columns]
row_ids = [row.identifier for row in ds_rows]
# Insert row at index position 1
row_ids.insert(1, ds.max_row_id() + 1)
# Result should indicate that one row was inserted. The identifier of
# the resulting dataset should differ from the identifier of the
# original dataset
result = self.api.insert_row(ds.identifier, 1, self.datastore)
# Resulting dataset should differ from previous one
self.assertNotEqual(result.dataset.identifier, ds.identifier)
# Retrieve modified dataset
ds = self.datastore.get_dataset(result.dataset.identifier)
ds_rows = ds.fetch_rows()
# Ensure that there are three rows
self.assertEqual(len(ds_rows), 3)
# The second row has empty values for each column
row = ds_rows[1]
self.assertEqual(len(row.values), len(ds.columns))
for i in range(len(ds.columns)):
self.assertIsNone(row.values[i])
# Append row at end current dataset
row_ids.append(ds.max_row_id() + 1)
result = self.api.insert_row(ds.identifier, 3, self.datastore)
# Resulting dataset should differ from previous one
self.assertNotEqual(result.dataset.identifier, ds.identifier)
ds = self.datastore.get_dataset(result.dataset.identifier)
ds_rows = ds.fetch_rows()
# Ensure that there are three rows
self.assertEqual(len(ds_rows), 4)
# The next to last row has non-empty values for each column
row = ds_rows[2]
self.assertEqual(len(row.values), len(ds.columns))
for i in range(len(ds.columns)):
self.assertIsNotNone(row.values[i])
# The last row has empty values for each column
row = ds_rows[3]
self.assertEqual(len(row.values), len(ds.columns))
for i in range(len(ds.columns)):
self.assertIsNone(row.values[i])
# Ensure that row ids haven't changed
for i in range(len(ds_rows)):
self.assertEqual(ds_rows[i].identifier, str(row_ids[i]))
# Make sure column identifier haven't changed
for i in range(len(ds.columns)):
self.assertEqual(ds.columns[i].identifier, col_ids[i])
# Ensure exception is thrown if dataset identifier is unknown
with self.assertRaises(ValueError):
self.api.insert_row('unknown:uri', 1, self.datastore)
# Ensure exception is thrown if row index is out of bounds
with self.assertRaises(ValueError):
self.api.insert_row(ds.identifier, 5, self.datastore)
# Ensure no exception is raised
self.api.insert_row(ds.identifier, 4, self.datastore)
def test_load_dataset(self):
"""Test functionality to load a dataset."""
# Create a new dataset
fh = self.filestore.upload_file(CSV_FILE)
result = self.api.load_dataset(datastore=self.datastore,
filestore=self.filestore,
file_id=fh.identifier)
ds = result.dataset
resources = result.resources
ds_rows = ds.fetch_rows()
self.assertEqual(len(ds.columns), 3)
self.assertEqual(len(ds_rows), 2)
for row in ds_rows:
self.assertTrue(isinstance(row.values[1], int))
self.assertIsNotNone(resources)
self.assertEqual(resources[RESOURCE_FILEID], fh.identifier)
self.assertEqual(resources[RESOURCE_DATASET], ds.identifier)
# Delete file handle and run load_dataset again with returned resource
# information. This should not raise an error since the file is not
# accessed but the previous dataset reused.
self.filestore.delete_file(fh.identifier)
result = self.api.load_dataset(datastore=self.datastore,
filestore=self.filestore,
file_id=fh.identifier,
resources=resources)
self.assertEqual(result.dataset.identifier, ds.identifier)
# Doing the same without the resources should raise an exception
with self.assertRaises(ValueError):
self.api.load_dataset(datastore=self.datastore,
filestore=self.filestore,
file_id=fh.identifier)
# Ensure exception is thrown if dataset identifier is unknown
with self.assertRaises(ValueError):
self.api.load_dataset(datastore=self.datastore,
filestore=self.filestore,
file_id='unknown:uri')
# Test loading file from external resource. Skip if DOWNLOAD_URL is None
if DOWNLOAD_URL is None:
print('Skipping download test')
return
result = self.api.load_dataset(datastore=self.datastore,
filestore=self.filestore,
url=DOWNLOAD_URL)
ds = result.dataset
resources = result.resources
ds_rows = ds.fetch_rows()
self.assertEqual(len(ds.columns), 4)
self.assertEqual(len(ds_rows), 54)
self.assertIsNotNone(resources)
self.assertEqual(resources[RESOURCE_URL], DOWNLOAD_URL)
self.assertEqual(resources[RESOURCE_DATASET], ds.identifier)
# Attempt to simulate re-running without downloading again. Set the
# Uri to some fake Uri that would raise an exception if an attempt was
# made to download
url = 'some fake uri'
resources[RESOURCE_URL] = url
result = self.api.load_dataset(datastore=self.datastore,
filestore=self.filestore,
url=url,
resources=resources)
prev_id = result.dataset.identifier
self.assertEqual(result.dataset.identifier, prev_id)
# If we re-run with reload flag true a new dataset should be returned
resources[RESOURCE_URL] = DOWNLOAD_URL
result = self.api.load_dataset(datastore=self.datastore,
filestore=self.filestore,
url=DOWNLOAD_URL,
resources=resources,
reload=True)
self.assertNotEqual(result.dataset.identifier, prev_id)
def test_move_column(self):
"""Test functionality to move a column."""
# Create a new dataset
fh = self.filestore.upload_file(CSV_FILE)
ds = self.api.load_dataset(datastore=self.datastore,
filestore=self.filestore,
file_id=fh.identifier).dataset
ds_rows = ds.fetch_rows()
# Keep track of column and row identifier
col_ids = [col.identifier for col in ds.columns]
row_ids = [row.identifier for row in ds_rows]
# Swap first two columns
c = col_ids[0]
del col_ids[0]
col_ids.insert(1, c)
result = self.api.move_column(ds.identifier,
ds.column_by_name('Name').identifier, 1,
self.datastore)
self.assertNotEqual(result.dataset.identifier, ds.identifier)
ds = self.datastore.get_dataset(result.dataset.identifier)
ds_rows = ds.fetch_rows()
self.assertEqual(ds.columns[0].name.upper(), 'Age'.upper())
self.assertEqual(ds.columns[1].name.upper(), 'Name'.upper())
self.assertEqual(ds.columns[2].name.upper(), 'Salary'.upper())
row = ds_rows[0]
self.assertEqual(row.values[0], 23)
self.assertEqual(row.values[1], 'Alice')
self.assertEqual(row.values[2], '35K')
row = ds_rows[1]
self.assertEqual(row.values[0], 32)
self.assertEqual(row.values[1], 'Bob')
self.assertEqual(row.values[2], '30K')
# Ensure that row ids haven't changed
for i in range(len(ds_rows)):
self.assertEqual(ds_rows[i].identifier, row_ids[i])
# Make sure column identifier haven't changed
for i in range(len(ds.columns)):
self.assertEqual(ds.columns[i].identifier, col_ids[i])
# Swap last two columns
c = col_ids[1]
del col_ids[1]
col_ids.append(c)
result = self.api.move_column(ds.identifier,
ds.column_by_name('Salary').identifier,
1, self.datastore)
ds = self.datastore.get_dataset(result.dataset.identifier)
ds_rows = ds.fetch_rows()
self.assertEqual(ds.columns[0].name.upper(), 'Age'.upper())
self.assertEqual(ds.columns[1].name.upper(), 'Salary'.upper())
self.assertEqual(ds.columns[2].name.upper(), 'Name'.upper())
row = ds_rows[0]
self.assertEqual(row.values[0], 23)
self.assertEqual(row.values[1], '35K')
self.assertEqual(row.values[2], 'Alice')
row = ds_rows[1]
self.assertEqual(row.values[0], 32)
self.assertEqual(row.values[1], '30K')
self.assertEqual(row.values[2], 'Bob')
# Ensure that row ids haven't changed
for i in range(len(ds_rows)):
self.assertEqual(ds_rows[i].identifier, row_ids[i])
# Make sure column identifier haven't changed
for i in range(len(ds.columns)):
self.assertEqual(ds.columns[i].identifier, col_ids[i])
# No changes if source and target position are the same
result = self.api.move_column(ds.identifier, ds.columns[1].identifier,
1, self.datastore)
self.assertEqual(ds.identifier, result.dataset.identifier)
# Ensure exception is thrown if dataset identifier is unknown
with self.assertRaises(ValueError):
self.api.move_column('unknown:uri', 0, 1, self.datastore)
# Raise error if source column is out of bounds
with self.assertRaises(ValueError):
self.api.move_column(ds.identifier, 40, 1, self.datastore)
# Raise error if target position is out of bounds
with self.assertRaises(ValueError):
self.api.move_column(ds.identifier,
ds.column_by_name('Name').identifier, -1,
self.datastore)
with self.assertRaises(ValueError):
self.api.move_column(ds.identifier,
ds.column_by_name('Name').identifier, 4,
self.datastore)
def test_move_row(self):
"""Test functionality to move a row."""
# Create a new dataset
fh = self.filestore.upload_file(CSV_FILE)
ds = self.api.load_dataset(datastore=self.datastore,
filestore=self.filestore,
file_id=fh.identifier).dataset
ds_rows = ds.fetch_rows()
# Keep track of column and row identifier
col_ids = [col.identifier for col in ds.columns]
row_ids = [row.identifier for row in ds_rows]
# Swap first two rows
row_ids = [row for row in reversed(row_ids)]
result = self.api.move_row(ds.identifier, 0, 1, self.datastore)
self.assertNotEqual(result.dataset.identifier, ds.identifier)
ds = self.datastore.get_dataset(result.dataset.identifier)
ds_rows = ds.fetch_rows()
self.assertEqual(ds.columns[0].name.upper(), 'Name'.upper())
self.assertEqual(ds.columns[1].name.upper(), 'Age'.upper())
self.assertEqual(ds.columns[2].name.upper(), 'Salary'.upper())
row = ds_rows[0]
self.assertEqual(row.values[0], 'Bob')
self.assertEqual(row.values[1], 32)
self.assertEqual(row.values[2], '30K')
row = ds_rows[1]
self.assertEqual(row.values[0], 'Alice')
self.assertEqual(row.values[1], 23)
self.assertEqual(row.values[2], '35K')
# Ensure that row ids haven't changed
for i in range(len(ds_rows)):
self.assertEqual(ds_rows[i].identifier, row_ids[i])
# Make sure column identifier haven't changed
for i in range(len(ds.columns)):
self.assertEqual(ds.columns[i].identifier, col_ids[i])
# Swap last two rows
row_ids = [row for row in reversed(row_ids)]
result = self.api.move_row(ds.identifier, 1, 0, self.datastore)
ds = self.datastore.get_dataset(result.dataset.identifier)
ds_rows = ds.fetch_rows()
self.assertEqual(ds.columns[0].name.upper(), 'Name'.upper())
self.assertEqual(ds.columns[1].name.upper(), 'Age'.upper())
self.assertEqual(ds.columns[2].name.upper(), 'Salary'.upper())
row = ds_rows[0]
self.assertEqual(row.values[0], 'Alice')
self.assertEqual(row.values[1], 23)
self.assertEqual(row.values[2], '35K')
row = ds_rows[1]
self.assertEqual(row.values[0], 'Bob')
self.assertEqual(row.values[1], 32)
self.assertEqual(row.values[2], '30K')
# Ensure that row ids haven't changed
for i in range(len(ds_rows)):
self.assertEqual(ds_rows[i].identifier, row_ids[i])
# Make sure column identifier haven't changed
for i in range(len(ds.columns)):
self.assertEqual(ds.columns[i].identifier, col_ids[i])
# Move first row to the end
result = self.api.move_row(ds.identifier, 0, 2, self.datastore)
row_ids = [row for row in reversed(row_ids)]
ds = self.datastore.get_dataset(result.dataset.identifier)
ds_rows = ds.fetch_rows()
row = ds_rows[0]
self.assertEqual(row.values[0], 'Bob')
self.assertEqual(row.values[1], 32)
self.assertEqual(row.values[2], '30K')
row = ds_rows[1]
self.assertEqual(row.values[0], 'Alice')
self.assertEqual(row.values[1], 23)
self.assertEqual(row.values[2], '35K')
# Ensure that row ids haven't changed
for i in range(len(ds_rows)):
self.assertEqual(ds_rows[i].identifier, row_ids[i])
# Make sure column identifier haven't changed
for i in range(len(ds.columns)):
self.assertEqual(ds.columns[i].identifier, col_ids[i])
# No changes if source and target position are the same
result = self.api.move_row(ds.identifier, 1, 1, self.datastore)
self.assertEqual(ds.identifier, result.dataset.identifier)
# Ensure exception is thrown if dataset identifier is unknown
with self.assertRaises(ValueError):
self.api.move_row('unknown:uri', 0, 1, self.datastore)
# Raise error if source row is out of bounds
with self.assertRaises(ValueError):
self.api.move_row(ds.identifier, 3, 1, self.datastore)
# Raise error if target position is out of bounds
with self.assertRaises(ValueError):
self.api.move_row(ds.identifier, 0, -1, self.datastore)
with self.assertRaises(ValueError):
self.api.move_row(ds.identifier, 1, 4, self.datastore)
def test_rename_column(self):
"""Test functionality to rename a column."""
# Create a new dataset
fh = self.filestore.upload_file(CSV_FILE)
ds = self.api.load_dataset(datastore=self.datastore,
filestore=self.filestore,
file_id=fh.identifier).dataset
ds_rows = ds.fetch_rows()
# Keep track of column and row identifier
col_ids = [col.identifier for col in ds.columns]
row_ids = [row.identifier for row in ds_rows]
# Rename first column to Firstname
result = self.api.rename_column(ds.identifier,
ds.column_by_name('Name').identifier,
'Firstname', self.datastore)
self.assertNotEqual(result.dataset.identifier, ds.identifier)
ds = self.datastore.get_dataset(result.dataset.identifier)
self.assertEqual(ds.columns[0].name.upper(), 'Firstname'.upper())
self.assertEqual(ds.columns[1].name.upper(), 'Age'.upper())
self.assertEqual(ds.columns[2].name.upper(), 'Salary'.upper())
result = self.api.rename_column(ds.identifier,
ds.column_by_name('Age').identifier,
'BDate', self.datastore)
ds = self.datastore.get_dataset(result.dataset.identifier)
ds_rows = ds.fetch_rows()
self.assertEqual(ds.columns[0].name.upper(), 'Firstname'.upper())
self.assertEqual(ds.columns[1].name, 'BDate')
self.assertEqual(ds.columns[2].name.upper(), 'Salary'.upper())
# Ensure that row ids haven't changed
for i in range(len(ds_rows)):
self.assertEqual(ds_rows[i].identifier, row_ids[i])
# Make sure column identifier haven't changed
for i in range(len(ds.columns)):
self.assertEqual(ds.columns[i].identifier, col_ids[i])
# No changes if the old and new column name are the same (with exception
# to upper and lower cases).
result = self.api.rename_column(ds.identifier,
ds.column_by_name('BDate').identifier,
'BDate', self.datastore)
self.assertEqual(ds.identifier, result.dataset.identifier)
# Ensure exception is thrown if dataset identifier is unknown
with self.assertRaises(ValueError):
self.api.rename_column('unknown:uri', 0, 'Firstname',
self.datastore)
# Ensure exception is thrown for invalid column id
with self.assertRaises(ValueError):
self.api.rename_column(ds.identifier, 500, 'BDate', self.datastore)
def test_sequence_of_steps(self):
"""Test sequence of calls that modify a dataset."""
# Create a new dataset
fh = self.filestore.upload_file(CSV_FILE)
ds = self.api.load_dataset(datastore=self.datastore,
filestore=self.filestore,
file_id=fh.identifier).dataset
ds = self.api.insert_row(ds.identifier, 1, self.datastore).dataset
ds = self.api.insert_column(ds.identifier, 3, 'HDate',
self.datastore).dataset
ds = self.api.update_cell(ds.identifier,
ds.column_by_name('HDate').identifier, 0,
'180', self.datastore).dataset
ds = self.api.update_cell(ds.identifier,
ds.column_by_name('HDate').identifier, 2,
'160', self.datastore).dataset
ds = self.api.rename_column(ds.identifier,
ds.column_by_name('HDate').identifier,
'Height', self.datastore).dataset
ds = self.api.update_cell(ds.identifier,
ds.column_by_name('Height').identifier, 1,
'170', self.datastore).dataset
ds = self.api.move_row(ds.identifier, 1, 2, self.datastore).dataset
ds = self.api.update_cell(ds.identifier,
ds.column_by_name('Name').identifier, 2,
'Carla', self.datastore).dataset
ds = self.api.update_cell(ds.identifier,
ds.column_by_name('Age').identifier, 2, '45',
self.datastore).dataset
ds = self.api.update_cell(ds.identifier,
ds.column_by_name('Salary').identifier, 2,
'56K', self.datastore).dataset
ds = self.api.move_column(ds.identifier,
ds.column_by_name('Salary').identifier, 4,
self.datastore).dataset
ds = self.api.delete_column(ds.identifier,
ds.column_by_name('Age').identifier,
self.datastore).dataset
ds = self.api.delete_row(ds.identifier, 0, self.datastore).dataset
ds = self.api.delete_row(ds.identifier, 0, self.datastore).dataset
ds = self.datastore.get_dataset(ds.identifier)
ds_rows = ds.fetch_rows()
names = ['Name', 'Height', 'Salary']
self.assertEqual(len(ds.columns), len(names))
for i in range(len(names)):
col = ds.columns[i]
self.assertEqual(col.name.upper(), names[i].upper())
self.assertEqual([col.identifier for col in ds.columns], [0, 3, 2])
self.assertEqual(len(ds_rows), 1)
self.assertEqual(ds_rows[0].values, ['Carla', '160', '56K'])
self.assertEqual(ds_rows[0].identifier, '2')
def test_sort_dataset(self):
"""Test sorting a dataset."""
# Create a new dataset
fh = self.filestore.upload_file(SORT_FILE)
ds = self.api.load_dataset(datastore=self.datastore,
filestore=self.filestore,
file_id=fh.identifier).dataset
result = self.api.sort_dataset(ds.identifier, [1, 2, 0],
[False, False, True], self.datastore)
ds = self.datastore.get_dataset(result.dataset.identifier)
rows = ds.fetch_rows()
names = ['Alice', 'Bob', 'Dave', 'Gertrud', 'Frank']
result = list()
for row in rows:
name = row.values[0]
if name in names:
result.append(name)
for i in range(len(names)):
self.assertEqual(names[i], result[i])
result = self.api.sort_dataset(ds.identifier, [2, 1, 0],
[True, False, True], self.datastore)
ds = self.datastore.get_dataset(result.dataset.identifier)
rows = ds.fetch_rows()
names = ['Gertrud', 'Frank', 'Bob', 'Alice', 'Dave']
result = list()
for row in rows:
name = row.values[0]
if name in names:
result.append(name)
for i in range(len(names)):
self.assertEqual(names[i], result[i])
# Raises error for invalid column identifier
with self.assertRaises(ValueError):
self.api.sort_dataset(ds.identifier, [2, 10, 0],
[True, False, True], self.datastore)
def test_update_cell(self):
"""Test functionality to update a dataset cell."""
# Create a new dataset
fh = self.filestore.upload_file(CSV_FILE)
ds = self.api.load_dataset(datastore=self.datastore,
filestore=self.filestore,
file_id=fh.identifier).dataset
ds_rows = ds.fetch_rows()
# Keep track of column and row identifier
col_ids = [col.identifier for col in ds.columns]
row_ids = [row.identifier for row in ds_rows]
# Update cell [0, 0]. Ensure that one row was updated and a new
# identifier is generated. Also ensure that the resulting datasets
# has the new value in cell [0, 0]
result = self.api.update_cell(ds.identifier, 0, 0, 'MyValue',
self.datastore)
self.assertNotEqual(ds.identifier, result.dataset.identifier)
ds = self.datastore.get_dataset(result.dataset.identifier)
ds_rows = ds.fetch_rows()
self.assertEqual(ds_rows[0].values[0], 'MyValue')
result = self.api.update_cell(ds.identifier,
ds.column_by_name('Name').identifier, 0,
'AValue', self.datastore)
ds = self.datastore.get_dataset(result.dataset.identifier)
ds_rows = ds.fetch_rows()
self.assertEqual(ds_rows[0].values[0], 'AValue')
self.assertEqual(ds_rows[0].values[ds.column_index('Name')], 'AValue')
# Ensure that row ids haven't changed
for i in range(len(ds_rows)):
self.assertEqual(ds_rows[i].identifier, row_ids[i])
# Make sure column identifier haven't changed
for i in range(len(ds.columns)):
self.assertEqual(ds.columns[i].identifier, col_ids[i])
# Set value to None
result = self.api.update_cell(ds.identifier,
ds.column_by_name('Name').identifier, 0,
None, self.datastore)
ds = self.datastore.get_dataset(result.dataset.identifier)
ds_rows = ds.fetch_rows()
self.assertIsNone(ds_rows[0].values[0])
self.assertIsNone(ds_rows[0].values[ds.column_index('Name')])
# Ensure exception is thrown if dataset is unknown
with self.assertRaises(ValueError):
self.api.update_cell('unknown:uri', 0, 0, 'MyValue',
self.datastore)
# Ensure exception is thrown if column is unknown
with self.assertRaises(ValueError):
self.api.update_cell(ds.identifier, 100, 0, 'MyValue',
self.datastore)
# Ensure exception is thrown if row index is out ouf bounds
with self.assertRaises(ValueError):
self.api.update_cell(ds.identifier, 0, 100, 'MyValue',
self.datastore)
