def get_metadata(self, dataset_identifier, content_type="json"):
"""
Retrieve the metadata for a particular dataset.
"""
resource = utils.format_old_api_request(dataid=dataset_identifier,
content_type=content_type)
return self._perform_request("get", resource)
def create(self, name, **kwargs):
"""
Create a dataset, including the field types. Optionally, specify args such as:
description : description of the dataset
columns : list of columns (see docs/tests for list structure)
category : must exist in /admin/metadata
tags : list of tag strings
row_identifier : field name of primary key
new_backend : whether to create the dataset in the new backend
WARNING: This api endpoint might be deprecated.
"""
new_backend = kwargs.pop("new_backend", False)
resource = utils.format_old_api_request(content_type="json")
if new_backend:
resource += "?nbe=true"
payload = {"name": name}
if "row_identifier" in kwargs:
payload["metadata"] = {
"rowIdentifier": kwargs.pop("row_identifier", None)
}
payload.update(kwargs)
payload = utils.clear_empty_values(payload)
return self._perform_update("post", resource, payload)
def publish(self, dataset_identifier, content_type="json"):
"""
The create() method creates a dataset in a "working copy" state.
This method publishes it.
"""
base = utils.format_old_api_request(dataid=dataset_identifier)
resource = "{0}/publication.{1}".format(base, content_type)
return self._perform_request("post", resource)
def update_metadata(self,
dataset_identifier,
update_fields,
content_type="json"):
"""
Update the metadata for a particular dataset.
update_fields is a dictionary containing [metadata key:new value] pairs.
This method performs a full replace for the key:value pairs listed in `update_fields`, and
returns all of the metadata with the updates applied.
"""
resource = utils.format_old_api_request(dataid=dataset_identifier,
content_type=content_type)
return self._perform_update("put", resource, update_fields)
def delete(self, dataset_identifier, row_id=None, content_type="json"):
"""
Delete the entire dataset, e.g.
client.delete("nimj-3ivp")
or a single row, e.g.
client.delete("nimj-3ivp", row_id=4)
"""
if row_id:
resource = utils.format_new_api_request(dataid=dataset_identifier,
row_id=row_id,
content_type=content_type)
else:
resource = utils.format_old_api_request(dataid=dataset_identifier,
content_type=content_type)
return self._perform_request("delete", resource)
def set_permission(self,
dataset_identifier,
permission="private",
content_type="json"):
"""
Set a dataset's permissions to private or public
Options are private, public
"""
resource = utils.format_old_api_request(dataid=dataset_identifier,
content_type=content_type)
params = {
"method": "setPermission",
"value": "public.read" if permission == "public" else permission,
}
return self._perform_request("put", resource, params=params)
def replace_non_data_file(self, dataset_identifier, params, file_data):
"""
Same as create_non_data_file, but replaces a file that already exists in a
file-based dataset.
WARNING: a table-based dataset cannot be replaced by a file-based dataset.
Use create_non_data_file in order to replace.
"""
resource = utils.format_old_api_request(dataid=dataset_identifier,
content_type="txt")
if not params.get("method", None):
params["method"] = "replaceBlob"
params["id"] = dataset_identifier
return self._perform_request("post",
resource,
params=params,
files=file_data)
def download_attachments(self,
dataset_identifier,
content_type="json",
download_dir="~/sodapy_downloads"):
"""
Download all of the attachments associated with a dataset. Return the paths of downloaded
files.
"""
metadata = self.get_metadata(dataset_identifier,
content_type=content_type)
files = []
attachments = metadata["metadata"].get("attachments")
if not attachments:
logging.info("No attachments were found or downloaded.")
return files
download_dir = os.path.join(os.path.expanduser(download_dir),
dataset_identifier)
if not os.path.exists(download_dir):
os.makedirs(download_dir)
for attachment in attachments:
file_path = os.path.join(download_dir, attachment["filename"])
has_assetid = attachment.get("assetId", False)
if has_assetid:
base = utils.format_old_api_request(dataid=dataset_identifier)
assetid = attachment["assetId"]
resource = "{0}/files/{1}?download=true&filename={2}".format(
base, assetid, attachment["filename"])
else:
base = "/api/assets"
assetid = attachment["blobId"]
resource = "{0}/{1}?download=true".format(base, assetid)
uri = "{0}{1}{2}".format(self.uri_prefix, self.domain, resource)
utils.download_file(uri, file_path)
files.append(file_path)
logging.info("The following files were downloaded:\n\t{0}".format(
"\n\t".join(files)))
return files
def test_format_old_api_request(dataid, content_type, path):
assert (
utils.format_old_api_request(dataid=dataid, content_type=content_type) == path
)
def test_format_old_api_request_exception():
with pytest.raises(Exception):
utils.format_old_api_request()