def _get_data(self, select, query, order_by, limit, distinct,
format, callback, count, index):
params = {}
if select:
if not isinstance(select, list):
raise PyBambooException(
'select must be a list of strings.')
select = dict([(sel, 1) for sel in select])
params['select'] = safe_json_dumps(
select,
PyBambooException('select is not JSON-serializable.'))
if query:
if not isinstance(query, dict):
raise PyBambooException('query must be a dict.')
params['query'] = safe_json_dumps(
query,
PyBambooException('query is not JSON-serializable.'))
if order_by:
if not isinstance(order_by, basestring):
raise PyBambooException('order_by must be a string.')
params['order_by'] = order_by
if format:
if not isinstance(format, basestring):
raise PyBambooException('format must be a string.')
if format not in self.DATA_FORMATS:
raise PyBambooException('format must be one of: %s.' %
self.DATA_FORMATS)
params['format'] = format
if distinct:
if not isinstance(distinct, basestring):
raise PyBambooException('distinct must be a string.')
params['distinct'] = distinct
if callback:
if not isinstance(callback, basestring):
raise PyBambooException('callback must be a string.')
params['callback'] = callback
if limit:
if not isinstance(limit, int):
raise PyBambooException('limit must be an int.')
params['limit'] = safe_json_dumps(
limit,
PyBambooException('limit is not JSON-serializable.'))
if count:
params['count'] = bool(count)
if index:
params['index'] = bool(index)
return self._connection.make_api_request(
'GET', '/datasets/%s' % self._id, params=params)
def merge(cls, datasets, connection=None):
"""
Create a new dataset that is a row-wise merge of those in *datasets*.
Returns the new merged dataset.
"""
if connection is None:
connection = Connection()
# TODO: allow list of dataset_ids?
checked_datasets = []
for dataset in datasets:
if not isinstance(dataset, Dataset):
raise PyBambooException(
'Datasets need to be instances of Dataset.')
checked_datasets.append(dataset.id)
data = {'dataset_ids': safe_json_dumps(
checked_datasets,
PyBambooException('datasets is not JSON-serializable.'))}
result = connection.make_api_request(
'POST', '/datasets/merge', data=data)
if 'id' in result.keys():
return Dataset(result['id'], connection=connection)
# this is never reached...
# see TestDataset.test_merge_fail()
return False
def _resample(self, date_column, interval, how, query, format):
params = {}
if not date_column or not isinstance(date_column, basestring):
raise PyBambooException('date_column must be a string.')
params['date_column'] = date_column
if not interval or not isinstance(interval, basestring):
raise PyBambooException('interval must be a string '
'representing a frequency. cf. '
'http://pytseries.sourceforge.net/'
'core.constants.html#date-frequencies')
params['interval'] = interval
if how:
if not isinstance(how, basestring):
raise PyBambooException('how must be a string.')
params['how'] = how
if format:
if not isinstance(format, basestring):
raise PyBambooException('format must be a string.')
params['format'] = format
if query:
if not isinstance(query, dict):
raise PyBambooException('query must be a dict.')
params['query'] = safe_json_dumps(
query,
PyBambooException('query is not JSON-serializable.'))
return self._connection.make_api_request(
'GET', '/datasets/%s/resample' % self._id, params=params)
def _get_summary(self, select, groups, query, order_by,
limit, callback):
params = {}
# TODO: check input params
if select != 'all':
if not isinstance(select, list):
raise PyBambooException(
'select must be a list of strings.')
select = dict([(sel, 1) for sel in select])
params['select'] = safe_json_dumps(
select,
PyBambooException('select is not JSON-serializable.'))
else:
params['select'] = select
if groups is not None:
if not isinstance(groups, list):
raise PyBambooException(
'groups must be a list of strings.')
params['group'] = ','.join(groups)
if query is not None:
if not isinstance(query, dict):
raise PyBambooException('query must be a dict.')
params['query'] = safe_json_dumps(
query,
PyBambooException('query is not JSON-serializable.'))
if order_by:
if not isinstance(order_by, basestring):
raise PyBambooException('order_by must be a string.')
params['order_by'] = order_by
if limit:
if not isinstance(limit, int):
raise PyBambooException('limit must be an int.')
params['limit'] = safe_json_dumps(
limit,
PyBambooException('limit is not JSON-serializable.'))
if callback:
if not isinstance(callback, basestring):
raise PyBambooException('callback must be a string.')
params['callback'] = callback
return self._connection.make_api_request(
'GET', '/datasets/%s/summary' % self._id, params=params)
def row(self, action=None, index=None, payload=None):
data = None
if not action in ('show', 'delete', 'edit'):
raise PyBambooException('Row action must be show|edit|delete.')
if not isinstance(index, int):
raise PyBambooException('index must be an int.')
if action == 'edit':
data = {"data": safe_json_dumps(payload, PyBambooException(
'payload is not JSON-serializable'))}
http_action = {'show': 'GET',
'delete': 'DELETE',
'edit': 'PUT'}.get(action)
return self._connection.make_api_request(
http_action, '/datasets/%s/row/%d' % (self._id, index), data=data)
def _add_calculations(self, path, content, json):
files = {}
if path is None and content is None and json is None:
raise PyBambooException('JSON fomulae must be provided as '
'JSON, path or content.')
if json is not None:
json_data = safe_json_dumps(json,
PyBambooException('formulae are '
'not JSON-serial'
'izable'))
content = StringIO.StringIO(json_data)
data = content if content is not None else open(path)
files.update({'json_file': ('data.json', data)})
response = self._connection.make_api_request(
'POST', '/datasets/%s/calculations' % self._id, files=files)
return 'error' not in response.keys()
def update_data(self, rows):
"""
Updates this dataset with the rows given in {column: value} format.
Any unspecified columns will result in n/a values.
"""
if not isinstance(rows, list):
raise PyBambooException(
'rows must be a list of dictionaries')
if len(rows) == 0:
raise PyBambooException(
'rows must contain at least one row dictionary')
for row in rows:
if not isinstance(row, dict):
raise PyBambooException(
'rows must be a list of dictionaries')
data = {
'update': safe_json_dumps(rows, PyBambooException(
'rows is not JSON-serializable'))
}
response = self._connection.make_api_request(
'PUT', '/datasets/%s' % self._id, data=data)
return 'id' in response.keys()
def __init__(self, dataset_id=None, url=None,
path=None, content=None, data_format='csv',
schema_path=None, schema_content=None,
na_values=None, connection=None, reset=False):
"""
Create a new pybamboo.Dataset from one of the following:
* dataset_id - the id of an existing bamboo.Dataset
* url - url to a .csv file
* path - path to a local .csv or .json file
* content - a CSV or JSON string
* data_format - whether path or content is csv | json
* schema_path - path to a JSON SDF schema
* schema_content - a JSON SDF string
One can also pass in a pybamboo.Connection object. If this is not
supplied one will be created automatically with the default options.
"""
if dataset_id is None and url is None \
and path is None and content is None \
and schema_path is None and schema_content is None:
raise PyBambooException(
'Must supply dataset_id, url, content, schema or file path.')
if data_format not in self.DATA_FORMATS:
raise PyBambooException('Illegal data_format: %s. data_format'
' must be one of %s' %
(data_format, self.DATA_FORMATS))
req_data = {}
if reset:
req_data.update({'dataset_id': self._id})
if na_values is not None:
if not isinstance(na_values, (list, tuple, set)):
raise PyBambooException('N/A values must be a list.')
self.NA_VALUES = na_values
req_data.update({'na_values':
safe_json_dumps(na_values,
PyBambooException('na_values '
'are not JSON-serializable'))})
if connection is None:
self._connection = Connection()
else:
self._connection = connection
if dataset_id is not None:
# TODO: check if this dataset exists?
self._id = dataset_id
return
if url is not None:
# TODO: check valid url?
req_data.update({'url': url})
self._id = self._connection.make_api_request(
'POST', '/datasets', req_data).get('id')
return
# files might be overloaded by schema or path/content
files = {}
if schema_path is not None or schema_content is not None:
# TODO: check for bad file stuff?
schema_data = schema_content if schema_content is not None \
else open(schema_path)
files.update({'schema': ('data.schema.json', schema_data)})
if path is not None or content is not None:
# TODO: check for bad file stuff?
data = content if content is not None else open(path)
files.update({'%s_file' % data_format:
('data.%s' % data_format, data)})
self._id = self._connection.make_api_request('POST', '/datasets',
files=files,
data=req_data).get('id')